{ "meta": { "title": "FDE 现场学习图谱", "asOf": "2026-09-08", "sourceWebsite": "https://fde100.datawhale.cn/cases", "sourceChina": "https://fdechina.ai/cases", "sourceBook": "Datawhale FDE案例100(15).pdf", "bookPageCount": 247, "bookCount": 24, "websiteCount": 24, "overlapCount": 23, "originalCount": 25, "appendedCount": 124, "total": 149, "diagramCount": 31, "indexOnlyCount": 118, "appendPolicy": "保留原有25条；FDE中国社区124条按网站顺序直接追加，不做跨来源或站内项目去重。", "unit": "学习条目，含案例、方案与文章，不是已交付项目数、客户数或市场样本", "classification": "行业 / 主题为单标签；方案和痛点只对已研读资料编码，目录文章不猜填。", "diagramPolicy": "原文明示的组件与学习补全分开；全部图形分层与连线为学习重构，不代表作者生产部署图；仅目录收录的文章不伪造架构。", "scope": "原有Datawhale册子与官网25条保持不变，按用户要求追加FDE中国社区124条，不去重。" }, "cases": [ { "id": "001", "title": "诉讼律师案件智能工作台", "person": "程子洋", "book": 3, "pages": [ 25, 33 ], "evidence": "PDF P27–30，Q3", "industry": "法律服务", "originalIndustry": "法律服务", "business": "知识库/问答", "types": [ "知识检索", "文档视觉理解", "协作工作台", "内容生成" ], "pains": [ "数据分散", "重复劳动", "交付慢", "质量与合规风险" ], "summary": "把零散案卷组织为持续保存上下文的案件工作空间，辅助律师整理证据、检索法条和形成文书。", "insight": "入口从聊天框改成案件工作台：让系统记住案件，而不是让律师每次重复介绍背景。", "boundary": "AI辅助整理与初步推理；诉讼策略、专业结论与对外交付由律师承担责任。", "missing": "未披露具体模型、图数据库、法律数据库供应商及检索评测集。", "exercise": "给一个合同纠纷设计案件上下文：哪些信息要长期保存，哪些引用必须回到原始证据？", "g": [ [ "案件材料|案卷、邮件、沟通记录", "律师需求|证据清单、策略、报告" ], [ "案件工作空间|案件上下文与知识图谱", "法律资料|结构化法条、判例" ], [ "大模型|理解、摘要、初步推理" ], [ "专业检索|法条检索、判例匹配", "文书与关系整理|证据关系、清单、报告" ], [ "律师核验|策略与专业责任", "可交付成果|客户或法院所需材料", "~案件权限隔离|学习补全：按案件授权" ] ], "links": [ "0a>1a", "0b>2a", "1a>2a", "1b>3a", "2a>3a", "2a>3b", "3a>3b", "3b>4a", "4a>4b", "4c>1a" ], "delivery": { "summary": "法索 AI · 案件工作台", "platforms": [ { "name": "法索 AI", "role": "律师使用的案件工作台" } ], "details": "围绕案件持续保存上下文，承载证据清单、案件关系、办案分析与法律文书等成果，由律师继续判断和使用。", "evidence": "PDF P28–30，Q3–Q4", "caveat": "交付形态是案件工作台，不是一次性聊天答复；律师保留专业判断与责任。" }, "acceptance": { "status": "实际使用反馈", "summary": "真实案件使用反馈；正式验收未披露", "details": "原文以复杂案件中的实际使用和处理效率说明价值；案件材料及生成成果仍需律师判断，没有公开完整验收测试集。", "evidence": "PDF P30–31，Q4", "caveat": "原文未披露正式项目验收清单、完整达标阈值或签字记录；业务复核与使用反馈不等于正式验收通过。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-001", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "案件材料", "detail": "案卷、邮件、沟通记录", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "律师需求", "detail": "证据清单、策略、报告", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "案件工作空间", "detail": "案件上下文与知识图谱", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "法律资料", "detail": "结构化法条、判例", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "大模型", "detail": "理解、摘要、初步推理", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "专业检索", "detail": "法条检索、判例匹配", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "文书与关系整理", "detail": "证据关系、清单、报告", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "律师核验", "detail": "策略与专业责任", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "可交付成果", "detail": "客户或法院所需材料", "inferred": false }, { "id": "n4c", "col": 4, "row": 2, "label": "案件权限隔离", "detail": "学习补全：按案件授权", "inferred": true } ], "svg": "architectures/case-001.svg", "mermaid": "architectures/case-001.mmd", "note": "notes/case-001.md" }, { "id": "002", "title": "建筑消防施工图智能生成", "person": "闫佳宜", "book": 21, "pages": [ 210, 218 ], "evidence": "PDF P213–214，Q3；P216–218", "industry": "建筑设计", "originalIndustry": "建筑设计", "business": "运营提效", "types": [ "文档视觉理解", "规则计算/优化", "内容生成" ], "pains": [ "重复劳动", "交付慢", "质量与合规风险" ], "summary": "识别CAD空间结构，把消防规范变成可执行规则，生成可供设计人员调整的专业初稿。", "insight": "“能生成”不等于“可交付”：专业人员需要调参数、审结果，输出还要符合设计院格式。", "boundary": "设计人员决定特殊空间参数和最终方案；约30分钟指初步生成，不等于完成审签。", "missing": "未披露专用CAD模型实现、规则引擎产品及工程验收测试数据。", "exercise": "把标准办公室与复杂设备间分开：哪些参数可默认，哪些必须让设计师明确确认？", "g": [ [ "CAD建筑底图|墙、门窗、空间与标注", "消防规范|空间类型、危险等级" ], [ "专用图形解析模型|提取建筑元素与空间", "规范规则化|K值、间距、覆盖要求" ], [ "空间与规范匹配|为不同空间确定约束" ], [ "消防布置生成|喷淋、消火栓等初稿", "~格式与图层检查|学习补全：交付格式校验" ], [ "设计人员审核|调整参数、专业判断", "施工图初稿|审核后进入设计交付" ] ], "links": [ "0a>1a", "0b>1b", "1a>2a", "1b>2a", "2a>3a", "3a>4a", "4a>3a", "4a>3b", "3b>4b" ], "delivery": { "summary": "消防设计应用 · CAD 布置图初稿", "platforms": [], "details": "提供图纸上传、空间解析、消防设施自动布置及专业参数调整，产出由设计人员继续审核修改的施工图初稿。", "evidence": "PDF P213–217，Q3–Q5", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "业务质量检查", "summary": "规范核对 + 设计师确认；正式验收未披露", "details": "基于行业规则生成，核对基础输入，关键节点由设计人员确认，并在真实工作中检查结果质量与可用性。", "evidence": "PDF P214、216–217", "caveat": "图纸的专业审核或消防工程验收，不等于该 AI 软件项目已完成交付验收；原文没有披露后者的正式记录。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-002", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "CAD建筑底图", "detail": "墙、门窗、空间与标注", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "消防规范", "detail": "空间类型、危险等级", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "专用图形解析模型", "detail": "提取建筑元素与空间", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "规范规则化", "detail": "K值、间距、覆盖要求", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "空间与规范匹配", "detail": "为不同空间确定约束", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "消防布置生成", "detail": "喷淋、消火栓等初稿", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "格式与图层检查", "detail": "学习补全：交付格式校验", "inferred": true }, { "id": "n4a", "col": 4, "row": 0, "label": "设计人员审核", "detail": "调整参数、专业判断", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "施工图初稿", "detail": "审核后进入设计交付", "inferred": false } ], "svg": "architectures/case-002.svg", "mermaid": "architectures/case-002.mmd", "note": "notes/case-002.md" }, { "id": "003", "title": "本地生活短视频脚本生产", "person": "十页", "book": 15, "pages": [ 148, 159 ], "evidence": "PDF P151–154，Q3", "industry": "营销服务", "originalIndustry": "互联网/软件", "business": "营销增长", "types": [ "内容生成", "Agent编排", "流程自动化" ], "pains": [ "重复劳动", "知识难传承", "交付慢", "判断口径不一致" ], "summary": "把餐饮代运营人员的经验变成商户字段、写作参考和可迭代工作流，规模化生成可拍摄脚本。", "insight": "单条Prompt不等于内容生产线；输入字段、校验、审稿和平台反馈一起构成稳定交付。", "boundary": "AI做标准化生成，人做语感、审美与可拍性判断；不稳定节点回退人工。", "missing": "未公开具体模型、自动校验标准和各商户转化效果。", "exercise": "为一家餐厅设计最少输入字段，并列出脚本交给拍摄前必须检查的三件事。", "g": [ [ "商户资料|菜品、活动、卖点、客群", "优秀运营经验|写作技巧与参考案例" ], [ "结构化业务参数|商户差异显式输入", "提示词与Skills|写法、规则、生产步骤" ], [ "内容生产工作流|组合生成与校验循环" ], [ "脚本草案|节奏、声画配合", "平台反馈|留存、SKU和客户反馈" ], [ "编辑审核|不稳定节点回退人工", "可拍摄脚本|交给实际内容生产", "~版本对照集|学习补全：固定回归样本" ] ], "links": [ "0a>1a", "0b>1b", "1a>2a", "1b>2a", "2a>3a", "3a>4a", "4a>4b", "3b>1b", "4c>2a" ], "delivery": { "summary": "脚本生产工作流 · 可拍摄脚本", "platforms": [], "details": "将商户资料、提示词、Skill 与审核步骤串联，输出符合商户场景的短视频脚本框架；不是完整的自动视频制作平台。", "evidence": "PDF P151–154，Q3–Q4", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "业务质量检查", "summary": "脚本人工审核与反馈迭代；正式验收未披露", "details": "生成内容要经业务人员检查，按商家风格、实际可拍摄性和后续内容反馈调整；不稳定的环节保留人工处理。", "evidence": "PDF P152–155，Q3–Q5", "caveat": "原文未披露正式项目验收清单、完整达标阈值或签字记录；业务复核与使用反馈不等于正式验收通过。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-003", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "商户资料", "detail": "菜品、活动、卖点、客群", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "优秀运营经验", "detail": "写作技巧与参考案例", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "结构化业务参数", "detail": "商户差异显式输入", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "提示词与Skills", "detail": "写法、规则、生产步骤", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "内容生产工作流", "detail": "组合生成与校验循环", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "脚本草案", "detail": "节奏、声画配合", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "平台反馈", "detail": "留存、SKU和客户反馈", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "编辑审核", "detail": "不稳定节点回退人工", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "可拍摄脚本", "detail": "交给实际内容生产", "inferred": false }, { "id": "n4c", "col": 4, "row": 2, "label": "版本对照集", "detail": "学习补全：固定回归样本", "inferred": true } ], "svg": "architectures/case-003.svg", "mermaid": "architectures/case-003.mmd", "note": "notes/case-003.md" }, { "id": "004", "title": "跨境电商 AI 工作流矩阵", "person": "Wendy周", "book": null, "pages": null, "evidence": "官网 CASE-004，业务背景、核心痛点与解决方案，读取于2026-09-08", "industry": "零售/电商", "originalIndustry": "零售/电商", "business": "运营提效 / 营销增长", "types": [ "Agent编排", "流程自动化", "规则计算/优化", "内容生成" ], "pains": [ "数据分散", "重复劳动", "判断口径不一致", "质量与合规风险" ], "summary": "面向多家卖家的实践集合：将选品、Listing、广告与内容工作拆成可修改的本地Skills，而非同一客户的一套全量部署。", "insight": "自动化的对象是稳定、重复的步骤；选品最终判断与会变动的经营口径仍交给业务人员。", "boundary": "材料称本地部署和人工可审核；不据此推定模型API、办公平台和所有数据传输都完全离线。", "missing": "具体模型、各卖家实际启用模块、云模型数据路径及长期业务收益待核验。", "exercise": "把选品拆成取数、筛选、判断：哪一步可全自动，哪一步必须出依据交给运营？", "g": [ [ "飞书/钉钉入口|任务发起", "业务数据与SOP|亚马逊、1688等平台" ], [ "卖家本地运行环境|OpenClaw / WorkBuddy", "可修改规则与样本|精品、精铺等经营口径" ], [ "Skills调度|按任务组合可复用步骤" ], [ "选品与复筛|浏览器取数、筛选、打分", "Listing与广告|文案、规则检查、批量表" ], [ "运营最终审核|策略与例外判断", "表格与工作成果|选品表、Listing、投放结构", "~对外动作确认|学习补全：投放前审批" ] ], "links": [ "0a>1a", "0b>1b", "1a>2a", "1b>2a", "2a>3a", "2a>3b", "3a>4a", "3b>4a", "4a>4b", "4c>4b" ], "delivery": { "summary": "WorkBuddy / OpenClaw + 飞书 / 钉钉入口", "platforms": [ { "name": "WorkBuddy", "role": "本地智能体运行底座" }, { "name": "OpenClaw", "role": "本地智能体运行底座" }, { "name": "飞书", "role": "办公任务入口" }, { "name": "钉钉", "role": "办公任务入口" } ], "details": "交付可复用、可人工复核的 Skills 与业务工作流，包含选品表、复筛表、产品开发表、Listing、自动化脚本、操作说明及演示录屏等。", "evidence": "官网 CASE-004，「解决方案」", "caveat": "这是多个卖家的实践合集；运行底座和办公入口是原文列出的组合，不代表每个客户都同时部署了四个平台。" }, "acceptance": { "status": "工作流质量检查", "summary": "按卖家 SOP 复核输出；正式验收未披露", "details": "以卖家自己的 SOP、正反例和规则检查工作流结果，例如 Listing 长度、关键词及合规检查；最终业务选择由人作出。", "evidence": "官网 CASE-004，「解决方案」", "caveat": "原文披露了输出复核与业务流程，没有披露所有卖家统一采用的正式验收清单或签字记录。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-004", "source": "仅官网", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "飞书/钉钉入口", "detail": "任务发起", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "业务数据与SOP", "detail": "亚马逊、1688等平台", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "卖家本地运行环境", "detail": "OpenClaw / WorkBuddy", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "可修改规则与样本", "detail": "精品、精铺等经营口径", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "Skills调度", "detail": "按任务组合可复用步骤", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "选品与复筛", "detail": "浏览器取数、筛选、打分", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "Listing与广告", "detail": "文案、规则检查、批量表", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "运营最终审核", "detail": "策略与例外判断", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "表格与工作成果", "detail": "选品表、Listing、投放结构", "inferred": false }, { "id": "n4c", "col": 4, "row": 2, "label": "对外动作确认", "detail": "学习补全：投放前审批", "inferred": true } ], "svg": "architectures/case-004.svg", "mermaid": "architectures/case-004.mmd", "note": "notes/case-004.md" }, { "id": "005", "title": "电信网络数据分析 Agent", "person": "Simon", "book": 16, "pages": [ 160, 168 ], "evidence": "PDF P163–164，Q3", "industry": "电信通信", "originalIndustry": "互联网/软件", "business": "数据分析", "types": [ "数据分析/预警", "Agent编排" ], "pains": [ "重复劳动", "交付慢", "判断口径不一致", "采纳与协同困难" ], "summary": "把电信网络分析拆成可组合的输入、处理、分析和结构化输出组件，交付可重复运行的业务系统。", "insight": "验收重点包括同一问题反复执行的稳定性，而不仅是一次演示能答对。", "boundary": "访谈明确最终沿用聊天入口；云端统一治理、记忆与语义层是扩展考量，不视为全部已上线。", "missing": "未披露数据连接器、分析算法、模型版本及云端落地范围。", "exercise": "设计一个网络异常分析的固定输入和输出，让两次执行能被客观比较。", "g": [ [ "原有聊天入口|分析师提交网络问题", "网络业务数据|对应分析任务的数据" ], [ "问题与数据准备|校验输入、明确输出结构" ], [ "分析Agent|理解问题、编排分析组件" ], [ "可组合分析组件|数据处理、分析、结果检查", "~统一运行与监督|扩展考量：云端、语义层" ], [ "结构化分析结果|符合业务要求的格式", "~分析师复核|学习补全：高影响结论确认" ] ], "links": [ "0a>1a", "0b>1a", "1a>2a", "2a>3a", "3a>4a", "3b>2a", "4a>4b" ], "delivery": { "summary": "Chat 入口 · 网络分析报告", "platforms": [], "details": "通过对话接收网络分析需求，调用分析能力并返回结构化结果，供电信业务人员判断；最终入口品牌未确认。", "evidence": "PDF P163–165，Q3–Q4", "caveat": "Web、Teams、Slack 在访谈中作为可能入口举例，不能据此认定本项目交付到了这些平台。" }, "acceptance": { "status": "业务质量检查", "summary": "输入、格式与重复运行稳定性检查", "details": "原文强调输入数据准确、分析可靠、输出格式稳定，以及重复执行时结果质量可控；没有披露最终验收阈值。", "evidence": "PDF P163–166，Q3–Q5", "caveat": "原文未披露正式项目验收清单、完整达标阈值或签字记录；业务复核与使用反馈不等于正式验收通过。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-005", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "原有聊天入口", "detail": "分析师提交网络问题", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "网络业务数据", "detail": "对应分析任务的数据", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "问题与数据准备", "detail": "校验输入、明确输出结构", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "分析Agent", "detail": "理解问题、编排分析组件", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "可组合分析组件", "detail": "数据处理、分析、结果检查", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "统一运行与监督", "detail": "扩展考量：云端、语义层", "inferred": true }, { "id": "n4a", "col": 4, "row": 0, "label": "结构化分析结果", "detail": "符合业务要求的格式", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "分析师复核", "detail": "学习补全：高影响结论确认", "inferred": true } ], "svg": "architectures/case-005.svg", "mermaid": "architectures/case-005.mmd", "note": "notes/case-005.md" }, { "id": "006", "title": "工程租赁获客与长尾报价", "person": "赵云淞", "book": 23, "pages": [ 229, 236 ], "evidence": "PDF P231–233，Q3", "industry": "工程租赁", "originalIndustry": "企业服务", "business": "销售提效", "types": [ "Agent编排", "规则计算/优化", "流程自动化" ], "pains": [ "供需匹配低效", "重复劳动", "交付慢", "知识难传承" ], "summary": "围绕获客和询价入口，把标准化小单与复杂大客户分流，利用业务成本因素形成可沟通的报价范围。", "insight": "先分订单，再做自动化；大客户价格不透明，不适合交给通用报价机器人。", "boundary": "大客户及复杂订单由老板处理；小单几分钟成交及自动调度部分含方案设想，未当作已测结果。", "missing": "未披露实际渠道接入清单、价格模型、成交率和完整调度实现。", "exercise": "定义三条必须转人工的条件，避免AI为了成交突破报价底线。", "g": [ [ "获客与轻量入口|短视频、小程序等规划", "客户询价|车型、距离、作业需求" ], [ "订单分层|标准小单 / 复杂大客户", "成本与报价因素|操作员、油费、作业类型" ], [ "小单需求沟通|AI补齐信息与沟通价格" ], [ "报价区间计算|依据企业业务因素", "~成交衔接调度|材料设想：快速进入调度" ], [ "大单/异常转人工|老板判断与关系维护", "长尾订单结果|接受、拒绝或转人工" ] ], "links": [ "0a>1a", "0b>1a", "1a>2a", "1a>4a", "1b>3a", "2a>3a", "3a>2a", "2a>4b", "4b>3b" ], "delivery": { "summary": "小程序 / 轻量报价入口（方案描述）", "platforms": [], "details": "围绕工程租赁询价、标准报价和大额订单转人工，描述面向一线的轻量入口；品牌与完整落地交付件未充分披露。", "evidence": "PDF P232–234，Q3–Q5", "caveat": "原文部分内容属于方案设想，不写成已完成的小程序部署；也不把其他客户的微信机器人举例移入本案。" }, "acceptance": { "status": "方案中的验证路径", "summary": "先小 POC 看效果；最终验收未披露", "details": "访谈提出先用小范围 POC 让客户看到效果，再安排实施路线；没有提供本案例已完成的正式验收记录。", "evidence": "PDF P232–234，Q3–Q5", "caveat": "这里是访谈描述的推进方法，不代表本案已经完成 POC 或最终验收。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-006", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "获客与轻量入口", "detail": "短视频、小程序等规划", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "客户询价", "detail": "车型、距离、作业需求", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "订单分层", "detail": "标准小单 / 复杂大客户", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "成本与报价因素", "detail": "操作员、油费、作业类型", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "小单需求沟通", "detail": "AI补齐信息与沟通价格", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "报价区间计算", "detail": "依据企业业务因素", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "成交衔接调度", "detail": "材料设想：快速进入调度", "inferred": true }, { "id": "n4a", "col": 4, "row": 0, "label": "大单/异常转人工", "detail": "老板判断与关系维护", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "长尾订单结果", "detail": "接受、拒绝或转人工", "inferred": false } ], "svg": "architectures/case-006.svg", "mermaid": "architectures/case-006.mmd", "note": "notes/case-006.md" }, { "id": "007", "title": "车管所网办材料自动审核", "person": "陈景驭", "book": 2, "pages": [ 15, 24 ], "evidence": "PDF P18–20，Q3；P21–23", "industry": "政务/公共服务", "originalIndustry": "政务/公共服务", "business": "运营提效", "types": [ "文档视觉理解", "流程自动化", "规则计算/优化" ], "pains": [ "重复劳动", "交付慢", "质量与合规风险", "数据分散" ], "summary": "将材料看、读、审、填和跨系统搬运串起来，正常业务自动处理，异常材料转给工作人员。", "insight": "识别准确率只是中间指标；识别之后能否继续办件，才决定业务价值。", "boundary": "审核尺度由业务方决定；不完整图片、非标准合同等异常需人工核验。", "missing": "未披露具体识别模型、规则配置、RPA产品与各材料类别测试分布。", "exercise": "给模糊合同、缺页材料和字段冲突设计三种回退路径，不让识别置信度直接等于审核通过。", "g": [ [ "网办材料|合同、发票、证照等图像", "真实异常样本|补拍、缺页、非标准格式" ], [ "视觉与OCR|材料分类、文字识别", "大模型信息提取|复杂合同关键信息" ], [ "业务审核规则|检查材料与字段要求" ], [ "RPA执行|填单、跨Web系统流转", "样本迭代|标注、正负样本、再优化" ], [ "原业务系统|接续统一办件流程", "人工异常核验|结果报表、风险材料" ] ], "links": [ "0a>1a", "0a>1b", "1a>2a", "1b>2a", "2a>3a", "3a>4a", "2a>4b", "0b>3b", "3b>1a" ], "delivery": { "summary": "既有业务 Web 系统 + RPA 自动办理", "platforms": [], "details": "将材料识别、审核和信息填写嵌入现有办理流程，由 RPA 操作既有系统，衔接公安部统一业务系统。", "evidence": "PDF P18–20，Q3–Q4", "caveat": "交付落在现有业务系统和自动化流程中；原文未披露通用智能体平台品牌。" }, "acceptance": { "status": "原文称验收通过", "summary": "已通过项目验收；具体验收方法未披露", "details": "受访者明确表示项目最终通过验收，之后继续承接新业务并共同设立 AI 创新实验室；正式测试清单与达标条件未公开。", "evidence": "PDF P21，Q4", "caveat": "识别准确率等过程指标不能直接当作合同验收阈值；这里只记录原文的通过表述，不代表独立核验。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-007", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "网办材料", "detail": "合同、发票、证照等图像", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "真实异常样本", "detail": "补拍、缺页、非标准格式", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "视觉与OCR", "detail": "材料分类、文字识别", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "大模型信息提取", "detail": "复杂合同关键信息", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "业务审核规则", "detail": "检查材料与字段要求", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "RPA执行", "detail": "填单、跨Web系统流转", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "样本迭代", "detail": "标注、正负样本、再优化", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "原业务系统", "detail": "接续统一办件流程", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "人工异常核验", "detail": "结果报表、风险材料", "inferred": false } ], "svg": "architectures/case-007.svg", "mermaid": "architectures/case-007.mmd", "note": "notes/case-007.md" }, { "id": "008", "title": "城市规划研判与报告生成", "person": "杜天宇", "book": 4, "pages": [ 34, 46 ], "evidence": "PDF P38–44，Q3–Q5", "industry": "政务/公共服务", "originalIndustry": "政务/公共服务", "business": "数据分析/运营提效", "types": [ "知识检索", "规则计算/优化", "Agent编排", "内容生成", "协作工作台" ], "pains": [ "数据分散", "交付慢", "质量与合规风险", "采纳与协同困难" ], "summary": "围绕小范围试点治理空间数据、规则与报告模板，让AI调度GIS等专业工具形成可追溯的规划初稿。", "insight": "先缩到具体单元和两条业务流；高频规则任务用工作流，开放任务再用Agent。", "boundary": "事实与精确计算交给真实数据和专业工具，规划师最终判断；多Agent是讨论过的取舍，不是部署事实。", "missing": "本册未完整披露部署拓扑、全部工具实现和多阶段指标口径；不混入旧库更细技术栈。", "exercise": "只拿到两个部门的数据时，怎样保住一文、一图、一表的最小交付？", "g": [ [ "地图工作台|规划师提出研判需求", "试点数据|空间、业务、规则与模板" ], [ "治理后的数据底座|仅先处理试点必要数据", "规划知识检索|法规、专业文档、历史规则" ], [ "核心Agent / 工作流|理解任务与调度工具" ], [ "GIS等专业工具|空间计算、规则判断", "报告结构化生成|基于事实与模板成稿" ], [ "地图与报告初稿|一文、一图、一表", "规划师判断|核验依据、作最终决策" ] ], "links": [ "0a>2a", "0b>1a", "0b>1b", "1a>3a", "1b>2a", "2a>3a", "3a>3b", "1b>3b", "3b>4a", "4a>4b" ], "delivery": { "summary": "地图工作台 + Copilot · 文 / 图 / 表", "platforms": [], "details": "以地图为中心承载专业分析工具、对话助手与报告生成，交付规划业务所需的文字、图件和表格。", "evidence": "PDF P39–43，Q3–Q4", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "试点使用验证", "summary": "从具体流程试点，看报告可用与实际采纳", "details": "先从具体单位和工作流程验证，专业人员检查数据、规则与报告成果，并在跨部门真实使用中反馈；原文披露使用人数与耗时自述。", "evidence": "PDF P40–43，Q3–Q4", "caveat": "原文未披露正式项目验收清单、完整达标阈值或签字记录；业务复核与使用反馈不等于正式验收通过。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-008", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "地图工作台", "detail": "规划师提出研判需求", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "试点数据", "detail": "空间、业务、规则与模板", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "治理后的数据底座", "detail": "仅先处理试点必要数据", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "规划知识检索", "detail": "法规、专业文档、历史规则", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "核心Agent / 工作流", "detail": "理解任务与调度工具", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "GIS等专业工具", "detail": "空间计算、规则判断", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "报告结构化生成", "detail": "基于事实与模板成稿", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "地图与报告初稿", "detail": "一文、一图、一表", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "规划师判断", "detail": "核验依据、作最终决策", "inferred": false } ], "svg": "architectures/case-008.svg", "mermaid": "architectures/case-008.mmd", "note": "notes/case-008.md" }, { "id": "009", "title": "制造业经验传承与新人培训", "person": "蔡洁如", "book": 1, "pages": [ 7, 14 ], "evidence": "PDF P9–11，Q3", "industry": "制造业", "originalIndustry": "制造业", "business": "知识库/问答", "types": [ "知识检索", "协作工作台", "组织能力建设" ], "pains": [ "知识难传承", "重复劳动", "采纳与协同困难" ], "summary": "把老师傅隐性经验和已有文档整理成知识地图及RAG能力，嵌入原有课程学习与进度验收。", "insight": "调研得到的高频问题同时成为测试集；知识库验收不是“能聊几句”。", "boundary": "带教师傅检查进度与学习结果，生产专业判断不由问答模型替代。", "missing": "未披露知识库产品、嵌入模型和课程平台；访谈给的解析与召回阈值为该案测试目标。", "exercise": "从一次新人跟岗中收集十个高频问题，如何把它们改成知识库验收用例？", "g": [ [ "老师傅经验|跟岗观察、隐性知识", "企业已有材料|培训资料与生产文档" ], [ "知识地图|高频问题、易错知识", "文档解析与知识加工|可检索的经验材料" ], [ "RAG检索问答|回答新人的真实问题" ], [ "原培训平台|课程、打卡、进度与问答", "问题集验收|用调研问题检查召回" ], [ "新人学习|按课程完成任务", "师傅跟踪|查看进度、验证学习结果" ] ], "links": [ "0a>1a", "0b>1b", "1a>1b", "1b>2a", "2a>3a", "1a>3b", "3b>2a", "3a>4a", "3a>4b" ], "delivery": { "summary": "既有培训平台 · 知识问答与课程打卡", "platforms": [], "details": "把知识检索接入企业实际使用的培训平台，增加课程问答、学习进度和师傅查看入口，并持续陪跑调整。", "evidence": "PDF P10–11，Q3", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "原文有阶段验收标准", "summary": "高频题测试集 + 每课问答验收", "details": "以调研得到的高频问题作为测试集；最小范围内约定文档解析约 95%、问题召回片段准确率 80% 以上，技术路径跑通后再通过课程问答和使用跟踪验证。", "evidence": "PDF P10–11，Q3", "caveat": "这些数字是访谈披露的小范围阶段验证标准，不是全量生产表现或最终合同验收报告；课程问答验收针对学习结果。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-009", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "老师傅经验", "detail": "跟岗观察、隐性知识", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "企业已有材料", "detail": "培训资料与生产文档", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "知识地图", "detail": "高频问题、易错知识", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "文档解析与知识加工", "detail": "可检索的经验材料", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "RAG检索问答", "detail": "回答新人的真实问题", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "原培训平台", "detail": "课程、打卡、进度与问答", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "问题集验收", "detail": "用调研问题检查召回", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "新人学习", "detail": "按课程完成任务", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "师傅跟踪", "detail": "查看进度、验证学习结果", "inferred": false } ], "svg": "architectures/case-009.svg", "mermaid": "architectures/case-009.mmd", "note": "notes/case-009.md" }, { "id": "010", "title": "跨境物流智能装车规划", "person": "刘春生", "book": 6, "pages": [ 57, 68 ], "evidence": "PDF P61–63，Q3", "industry": "跨境物流", "originalIndustry": "跨境物流", "business": "运营提效", "types": [ "Agent编排", "规则计算/优化" ], "pains": [ "知识难传承", "交付慢", "供需匹配低效" ], "summary": "把车辆和货物约束交给专业三维装箱算法，由大模型理解任务、调度工具并解释装车方案。", "insight": "最终交付从复杂三维交互改成一张装车指导图：适合现场执行比炫酷更重要。", "boundary": "大模型不做确定性空间计算；工人按图执行并反馈现场约束。", "missing": "未披露全部载荷、安全约束、算法参数与多车调度的验证范围。", "exercise": "为一辆车列出必须交给确定性算法的约束，再设计工人能看懂的输出图。", "g": [ [ "车辆与货物信息|真实尺寸、货物参数", "装车需求|多车安排与执行要求" ], [ "业务约束整理|装载条件、现场经验" ], [ "大模型调度|理解需求、调用与解释" ], [ "Python装箱算法|高度图、贪心、遗传策略", "装车指导图生成|编号、方向、截面、顺序" ], [ "工人现场执行|按指导图装载并反馈", "~安全约束复核|学习补全：重量与稳定性" ] ], "links": [ "0a>1a", "0b>2a", "1a>3a", "2a>3a", "3a>3b", "3b>4a", "4a>1a", "4b>3a" ], "delivery": { "summary": "装车算法原型 · 装车指导图", "platforms": [], "details": "将货号、摆放方向、截面和装载顺序输出为可供现场参考的指导图，并提供原型交互页面。", "evidence": "PDF P62–64，Q3–Q4", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "原型验证", "summary": "按车型与货物规格模拟装载；尚属原型", "details": "原文称已完成原型验证，使用真实车辆尺寸和货物规格进行模拟装载，检查空间安排及装载指导结果。", "evidence": "PDF P63–64，Q4", "caveat": "现场耗时从 5 小时降到 2 小时是预计效果，不能写成已测得的最终验收成果。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-010", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "车辆与货物信息", "detail": "真实尺寸、货物参数", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "装车需求", "detail": "多车安排与执行要求", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "业务约束整理", "detail": "装载条件、现场经验", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "大模型调度", "detail": "理解需求、调用与解释", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "Python装箱算法", "detail": "高度图、贪心、遗传策略", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "装车指导图生成", "detail": "编号、方向、截面、顺序", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "工人现场执行", "detail": "按指导图装载并反馈", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "安全约束复核", "detail": "学习补全：重量与稳定性", "inferred": true } ], "svg": "architectures/case-010.svg", "mermaid": "architectures/case-010.mmd", "note": "notes/case-010.md" }, { "id": "011", "title": "消防维保流程数字化", "person": "尚启峰", "book": 13, "pages": [ 131, 137 ], "evidence": "PDF P133–134，Q3", "industry": "消防维保", "originalIndustry": "消防维保", "business": "运营提效", "types": [ "流程自动化", "内容生成", "组织能力建设" ], "pains": [ "重复劳动", "数据分散", "采纳与协同困难" ], "summary": "先把纸表和Excel流程数字化，再在月检描述和报告等重复环节加入AI，避免再造第二套重系统。", "insight": "已有监管软件不开接口，是设计边界；增加完整新平台会造成双份录入。", "boundary": "专业系统记录流程，AI辅助提取生成，消防专业判断由人完成；两阶段方案不等于全部AI已上线。", "missing": "未披露具体数字化平台、与监管软件衔接方式及AI上线覆盖率。", "exercise": "监管系统不能接API时，如何减少重复填表，又不制造两套事实来源？", "g": [ [ "月检与历史记录|纸质表单、Excel、异常描述", "既有监管流程|强制使用、接口不可开放" ], [ "流程数字化|企业自己的记录与数据基础" ], [ "AI辅助处理|历史内容调用、描述生成" ], [ "轻量业务流程|记录、整理、形成报告", "~人工衔接监管端|学习补全：接口受限替代" ], [ "维保人员确认|消防专业判断与数据核对", "维保记录与报告|过程可追踪、减少重复" ] ], "links": [ "0a>1a", "1a>2a", "2a>3a", "3a>4a", "4a>4b", "0b>3b", "4b>3b" ], "delivery": { "summary": "飞书等基础工具 · 消防维保小模块", "platforms": [ { "name": "飞书", "role": "沿用企业既有基础工具，逐步做小模块" } ], "details": "先数字化工作记录和流程，再叠加小功能，沿用企业熟悉的基础工具逐步迭代，避免额外重复填报。", "evidence": "PDF P134–136，Q3–Q5", "caveat": "原文明确写了选择飞书等基础工具的路径，但没有披露具体飞书模块清单，也不表示所有 AI 功能已完成交付。" }, "acceptance": { "status": "持续迭代", "summary": "看一线是否实际使用；最终验收未披露", "details": "依据一线员工填报负担、流程接受度与实际反馈迭代；原文称项目仍在推进，没有提供大规模结果或最终验收单。", "evidence": "PDF P134–136，Q3–Q5", "caveat": "原文未披露正式项目验收清单、完整达标阈值或签字记录；业务复核与使用反馈不等于正式验收通过。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-011", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "月检与历史记录", "detail": "纸质表单、Excel、异常描述", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "既有监管流程", "detail": "强制使用、接口不可开放", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "流程数字化", "detail": "企业自己的记录与数据基础", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "AI辅助处理", "detail": "历史内容调用、描述生成", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "轻量业务流程", "detail": "记录、整理、形成报告", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "人工衔接监管端", "detail": "学习补全：接口受限替代", "inferred": true }, { "id": "n4a", "col": 4, "row": 0, "label": "维保人员确认", "detail": "消防专业判断与数据核对", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "维保记录与报告", "detail": "过程可追踪、减少重复", "inferred": false } ], "svg": "architectures/case-011.svg", "mermaid": "architectures/case-011.mmd", "note": "notes/case-011.md" }, { "id": "012", "title": "跨境电商库存、补货与广告联动", "person": "张新卓", "book": 22, "pages": [ 219, 228 ], "evidence": "PDF P222–224，Q3与图1", "industry": "零售/电商", "originalIndustry": "跨境电商", "business": "运营提效", "types": [ "文档视觉理解", "流程自动化", "规则计算/优化", "Agent编排", "数据分析/预警" ], "pains": [ "数据分散", "重复劳动", "经营不可见", "供需匹配低效" ], "summary": "以Supabase中转聚合SaaS数据到飞书，连接物流照片识别、补货规则与库存驱动广告调整。", "insight": "并非一开始就规划完整平台，而是从库存可见逐步发现物流、补货和广告的业务连接。", "boundary": "照片回填与补货单保留人工确认；广告自动变更的审批和权限细节未披露。", "missing": "未公开数据同步频率、冲突处理、广告接口及权限审批实现。", "exercise": "库存延迟同步时，如何避免误停广告或重复补货？定义一个安全回退条件。", "g": [ [ "聚合SaaS数据|库存、订单、物流", "物流照片|工厂上传到多维表" ], [ "Supabase中转库|克服接口与共享限制", "飞书业务数据|同步库存、原有补货规则" ], [ "Agent识图与回填|物流照片结构化", "补货与广告联动|规则算量、库存约束投放" ], [ "人工确认与制单|补货审核后生成PDF订单", "广告调整|降额度或暂停缺货SKU" ], [ "供应商订单|审核后发送的补货需求", "~广告变更审批|学习补全：权限与回滚" ] ], "links": [ "0a>1a", "1a>1b", "0b>2a", "2a>1b", "1b>2b", "2b>3a", "2b>3b", "3a>4a", "4b>3b" ], "delivery": { "summary": "飞书业务工作流 · 补货 PDF 订单", "platforms": [ { "name": "飞书", "role": "员工业务入口与表格工作流" } ], "details": "将业务数据接入熟悉的飞书工作环境，照片识别结果回填表格；补货建议经人确认后生成 PDF 订单给供应商。", "evidence": "PDF P222–224，Q3–Q4", "caveat": "Supabase 是数据中转与存储层，不是面向用户的最终交付入口；补货和投放并非完全无人决策。" }, "acceptance": { "status": "实际使用反馈", "summary": "识图回填复核 + 补货人工确认", "details": "以实际物流照片处理和补货业务检验效果，员工复核识别内容及补货建议；访谈给出处理耗时对比，但未披露正式验收流程。", "evidence": "PDF P223–225，Q3–Q5", "caveat": "交付后客户收回访问权限，受访者无法持续核算完整 ROI；人工确认订单是业务控制，不等于项目正式验收。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-012", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "聚合SaaS数据", "detail": "库存、订单、物流", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "物流照片", "detail": "工厂上传到多维表", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "Supabase中转库", "detail": "克服接口与共享限制", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "飞书业务数据", "detail": "同步库存、原有补货规则", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "Agent识图与回填", "detail": "物流照片结构化", "inferred": false }, { "id": "n2b", "col": 2, "row": 1, "label": "补货与广告联动", "detail": "规则算量、库存约束投放", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "人工确认与制单", "detail": "补货审核后生成PDF订单", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "广告调整", "detail": "降额度或暂停缺货SKU", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "供应商订单", "detail": "审核后发送的补货需求", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "广告变更审批", "detail": "学习补全：权限与回滚", "inferred": true } ], "svg": "architectures/case-012.svg", "mermaid": "architectures/case-012.mmd", "note": "notes/case-012.md" }, { "id": "013", "title": "TikTok 达人建联与履约协同", "person": "KevinSun孙斌", "book": 5, "pages": [ 47, 56 ], "evidence": "PDF P50–52，Q3", "industry": "零售/电商", "originalIndustry": "零售/电商", "business": "营销增长", "types": [ "内容生成", "数据分析/预警", "流程自动化" ], "pains": [ "供需匹配低效", "重复劳动", "经营不可见" ], "summary": "连接达人画像、商品卖点、脚本辅助和履约状态，让系统先筛匹配、主动提醒，运营处理关键关系。", "insight": "帮助达人想清楚怎么拍，也是提高履约的手段；不是只给TAP做一套催办工具。", "boundary": "目标达人沟通与关系建立保留人工，提醒无效再由运营介入；AI视频属于未来探索。", "missing": "未披露匹配模型、消息通道、调度实现及对照实验。", "exercise": "设计“已收样但未出视频”的状态流转，何时提醒，何时升级人工？", "g": [ [ "达人与商品信息|内容属性、卖点、受众", "履约业务状态|收样、创作、期限、GMV" ], [ "统一业务视角|达人、商品、任务与成本" ], [ "匹配与脚本辅助|筛人选品、给拍摄方向" ], [ "状态判断与提醒|临期、超期、未履约" ], [ "BD与运营确认|建联、关系维护、异常催办", "履约与复盘|视频交付及业务结果" ] ], "links": [ "0a>1a", "0b>1a", "1a>2a", "1a>3a", "2a>4a", "3a>4a", "4a>4b", "4b>1a" ], "delivery": { "summary": "达人协同工作流 · 匹配表 / 话术 / 提醒", "platforms": [], "details": "将达人筛选、合作推进、脚本建议和履约提醒串成统一业务视图，保留 BD 与达人的实际沟通。", "evidence": "PDF P50–52，Q3–Q4", "caveat": "飞书和 Excel 是原先的协作工具，原文没有明确新方案最终部署在哪个具名平台，不能直接填成飞书交付。" }, "acceptance": { "status": "试运行验证", "summary": "试运行跟踪履约周期与业务反馈", "details": "通过实际达人合作试运行，观察匹配与履约流程；原文以履约周期变化说明效果，但未披露正式验收测试集。", "evidence": "PDF P52–53，Q4–Q5", "caveat": "原文未披露正式项目验收清单、完整达标阈值或签字记录；业务复核与使用反馈不等于正式验收通过。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-013", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "达人与商品信息", "detail": "内容属性、卖点、受众", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "履约业务状态", "detail": "收样、创作、期限、GMV", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "统一业务视角", "detail": "达人、商品、任务与成本", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "匹配与脚本辅助", "detail": "筛人选品、给拍摄方向", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "状态判断与提醒", "detail": "临期、超期、未履约", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "BD与运营确认", "detail": "建联、关系维护、异常催办", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "履约与复盘", "detail": "视频交付及业务结果", "inferred": false } ], "svg": "architectures/case-013.svg", "mermaid": "architectures/case-013.mmd", "note": "notes/case-013.md" }, { "id": "014", "title": "国企电商 AI 工作流落地", "person": "沈凯", "book": 14, "pages": [ 138, 147 ], "evidence": "PDF P140–143，Q3", "industry": "零售/电商", "originalIndustry": "零售/电商", "business": "运营提效", "types": [ "知识检索", "Agent编排", "组织能力建设", "流程自动化" ], "pains": [ "知识难传承", "采纳与协同困难", "重复劳动", "判断口径不一致" ], "summary": "通过知识萃取和业务培训，将岗位隐性经验变成企业自己的规则、RAG和工作流，让员工实际用起来。", "insight": "技术底座做好不等于组织采纳；员工要知道如何贡献知识，也要保留确认与责任。", "boundary": "法务、财务等业务最终由人确认签字；不以人员替代作为已验证因果结果。", "missing": "未披露各岗位Agent清单、接入接口和具体模型部署方案。", "exercise": "给“合同初审助手”写岗位说明：输入、输出、例外、必须转人的条件分别是什么？", "g": [ [ "业务人员经验|原流程、例外与判断", "AI通识培训|帮助员工参与知识萃取" ], [ "知识加工与管理|企业文档和岗位知识", "企业SOP|明确步骤、规则与例外" ], [ "RAG与岗位Agent|依企业流程执行辅助工作" ], [ "既有信息化系统|在实际工作中承载结果" ], [ "员工确认签字|法务、财务等责任边界", "持续知识沉淀|业务反馈与流程迭代" ] ], "links": [ "0a>1a", "0b>0a", "0a>1b", "1a>2a", "1b>2a", "2a>3a", "3a>4a", "4a>4b", "4b>1a" ], "delivery": { "summary": "岗位 Agent / 知识库 + SOP 与培训", "platforms": [], "details": "将员工经验加工为岗位知识、企业 SOP 与工作流，结合既有信息化系统和员工培训，让业务人员参与使用与迭代。", "evidence": "PDF P141–144，Q3–Q4", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "实际使用反馈", "summary": "岗位使用与知识反馈；正式验收未披露", "details": "访谈重视员工是否参与、实际工作能否用起来以及知识能否持续更新；法务财务等事项仍由责任人确认，没有公开最终验收门槛。", "evidence": "PDF P142–146，Q3–Q5", "caveat": "原文未披露正式项目验收清单、完整达标阈值或签字记录；业务复核与使用反馈不等于正式验收通过。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-014", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "业务人员经验", "detail": "原流程、例外与判断", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "AI通识培训", "detail": "帮助员工参与知识萃取", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "知识加工与管理", "detail": "企业文档和岗位知识", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "企业SOP", "detail": "明确步骤、规则与例外", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "RAG与岗位Agent", "detail": "依企业流程执行辅助工作", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "既有信息化系统", "detail": "在实际工作中承载结果", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "员工确认签字", "detail": "法务、财务等责任边界", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "持续知识沉淀", "detail": "业务反馈与流程迭代", "inferred": false } ], "svg": "architectures/case-014.svg", "mermaid": "architectures/case-014.mmd", "note": "notes/case-014.md" }, { "id": "015", "title": "线下零售门店智能对账", "person": "Michael", "book": 10, "pages": [ 97, 107 ], "evidence": "PDF P100–102，Q3", "industry": "零售/电商", "originalIndustry": "零售/电商", "business": "运营提效", "types": [ "文档视觉理解", "流程自动化", "数据分析/预警" ], "pains": [ "重复劳动", "交付慢", "质量与合规风险" ], "summary": "周期性识别和比对门店与商场单据，先自动筛出确定结果，员工只判断异常差异。", "insight": "少花多少工时、少漏多少款，能较快验证；试点轻量但要知道扩店后的冗余需求。", "boundary": "实际落地为上海个位数门店；PDF P102明确Cherry Studio提供Agent能力，并另建用户前端与链接层，不是千店部署证明。", "missing": "未披露全部对账规则、异常阈值、后端冗余和规模化运行数据。", "exercise": "把金额识别错误与真实结算差异区分开，怎样给员工安排复核优先级？", "g": [ [ "门店与商场单据|纸质、图片及结算信息" ], [ "OCR与视觉识别|提取对账所需内容" ], [ "识别结果比对|第一轮核对、区分差异" ], [ "周期自动化工作流|定时运行、汇总结果", "~规模化冗余|扩展建议：失败重试队列" ], [ "员工差异判断|只处理需人工判断的条目", "对账结果|确认记录、遗漏款项线索" ] ], "links": [ "0a>1a", "1a>2a", "2a>3a", "3a>4a", "4a>4b", "3b>3a" ], "delivery": { "summary": "Cherry Studio Agent + 定制对账前端", "platforms": [ { "name": "Cherry Studio", "role": "项目中的 Agent 能力层" } ], "details": "Cherry Studio 提供 Agent 能力，另外构建用户友好的前端交互与 Agent 链接层，承载周期性对账和异常差异复核。", "evidence": "PDF P100–103，Q3–Q4；P102 图1说明", "caveat": "实际落地范围是上海个位数门店，不是千店部署；Cherry Studio 是能力层，用户前端为单独构建。" }, "acceptance": { "status": "实际业务验证", "summary": "真实账单比对 + 员工复核净差异", "details": "在真实门店账单中识别差异，由员工处理需要判断的部分，同时对比一个结算周期所需工时和确认后的金额差异。", "evidence": "PDF P102–103，Q4", "caveat": "耗时、差异比例是访谈中的使用效果；没有公开合同验收规则、逐单测试结果或千店运行报告。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-015", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "门店与商场单据", "detail": "纸质、图片及结算信息", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "OCR与视觉识别", "detail": "提取对账所需内容", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "识别结果比对", "detail": "第一轮核对、区分差异", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "周期自动化工作流", "detail": "定时运行、汇总结果", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "规模化冗余", "detail": "扩展建议：失败重试队列", "inferred": true }, { "id": "n4a", "col": 4, "row": 0, "label": "员工差异判断", "detail": "只处理需人工判断的条目", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "对账结果", "detail": "确认记录、遗漏款项线索", "inferred": false } ], "svg": "architectures/case-015.svg", "mermaid": "architectures/case-015.mmd", "note": "notes/case-015.md" }, { "id": "016", "title": "跨境快消海外数据流转", "person": "朱恒璟", "book": 24, "pages": [ 237, 246 ], "evidence": "PDF P240–243，Q3与效果说明", "industry": "零售/电商", "originalIndustry": "零售/电商", "business": "运营提效", "types": [ "文档视觉理解", "数据分析/预警", "流程自动化" ], "pains": [ "数据分散", "经营不可见", "重复劳动", "质量与合规风险" ], "summary": "在总部既有系统上叠加AI抽取与校验，统一经销商动销口径，分析结论能够回到原始数据行。", "insight": "先做中国一条产品线，跑通数据入口、可追溯数字和周会决策，再拓展地区。", "boundary": "风险数据由人审核；册子界面为demo数据，数字有脱敏，不视为可复算的生产原始证据。", "missing": "未披露具体模型、数据连接方式和脱敏前指标；不把示意界面当生产截图。", "exercise": "一个SKU销量下降，如何让AI的每个数字都能点回经销商原始报表？", "g": [ [ "总部发货数据|Sell In，较标准化", "经销商动销资料|PDF、图片、Excel" ], [ "字段与口径统一|连接发货、动销和市场费用" ], [ "AI抽取与校验|清洗转换、异常与风险标记" ], [ "既有管理系统|存储、权限、业务流转", "可追溯分析|定位原始行，不无据作答" ], [ "人工最终审核|异常、复杂业务判断", "经营数据看板|支持周会与市场决策" ] ], "links": [ "0a>1a", "0b>2a", "2a>1a", "1a>3a", "3a>3b", "2a>4a", "4a>3a", "3b>4b" ], "delivery": { "summary": "既有管理系统 · 可追溯经营看板", "platforms": [], "details": "保留总部管理系统，将数据抽取、校验和异常标记嵌入原流程；经营看板与问答结果可定位回原始数据行。", "evidence": "PDF P241–243，Q3–Q4", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "样本与业务试点验证", "summary": "正常 / 异常 / 边界样本 + 逐数字溯源", "details": "建立评估样本持续测试；先跑通中国一条产品线，要求数据能进入、数字能追溯、结论能用于周会，再扩展地区。", "evidence": "PDF P241–244，Q3–Q5", "caveat": "原文给出验证方法，但未披露正式验收报告；界面含 demo 数据，收益数字经脱敏，不作可复算的生产原始证据。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-016", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "总部发货数据", "detail": "Sell In，较标准化", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "经销商动销资料", "detail": "PDF、图片、Excel", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "字段与口径统一", "detail": "连接发货、动销和市场费用", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "AI抽取与校验", "detail": "清洗转换、异常与风险标记", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "既有管理系统", "detail": "存储、权限、业务流转", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "可追溯分析", "detail": "定位原始行，不无据作答", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "人工最终审核", "detail": "异常、复杂业务判断", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "经营数据看板", "detail": "支持周会与市场决策", "inferred": false } ], "svg": "architectures/case-016.svg", "mermaid": "architectures/case-016.mmd", "note": "notes/case-016.md" }, { "id": "017", "title": "基金公司 AI Coding 研发转型", "person": "吕昭波", "book": 7, "pages": [ 69, 77 ], "evidence": "PDF P72–74，Q3–Q4", "industry": "金融", "originalIndustry": "金融", "business": "研发提效", "types": [ "组织能力建设", "协作工作台" ], "pains": [ "采纳与协同困难", "判断口径不一致", "知识难传承" ], "summary": "通过真实项目、定制Spec文档和三期跟进，把个人使用AI工具变成产品、研发与业务的协作方式。", "insight": "这是一套研发与组织协作架构，不是生产推理服务；先用实际操作验证团队能力。", "boundary": "官网“文档节省50%以上”不能归为基金项目已核实收益：PDF P74说明该数据来自另一个制造业客户。", "missing": "基金客户未开放完整ROI；具体AI工具组合、版本管理与质量门禁未完整披露。", "exercise": "将一份PRD拆成多份Spec时，谁维护哪份文档？文档冲突时以什么为准？", "g": [ [ "业务与产品需求|老板、研发、产品、HR", "真实操作诊断|验证团队能力与差距" ], [ "定制Spec文档|逐步拆解、客户确认" ], [ "AI Coding实践|围绕真实研发任务使用工具" ], [ "五天半现场培训|教学、咨询、观察与调整", "三期持续跟进|扩展至产品与业务人员" ], [ "共同协作规范|从个人工具到组织方法", "~工程质量门禁|学习补全：测试与评审" ] ], "links": [ "0a>1a", "0b>3a", "1a>2a", "2a>3a", "3a>1a", "3a>3b", "3b>4a", "4b>2a" ], "delivery": { "summary": "定制 Spec 文档 + 现场培训与三期跟进", "platforms": [], "details": "交付研发协作方法、面向真实项目的 Spec 文档和培训陪跑，推动产品、研发与业务团队共同使用。", "evidence": "PDF P71–74，Q3–Q4", "caveat": "飞书 PRD 是原来的文档方式；Codex、Claude Code 等是既有工具背景，不等于本次最终交付的平台清单。" }, "acceptance": { "status": "培训与应用跟踪", "summary": "真实项目操作、团队复用与后续跟进", "details": "通过现场训练和后续三期跟进观察协作框架能否在真实项目中使用、复用；基金客户没有提供完整 ROI 或正式项目验收标准。", "evidence": "PDF P72–74，Q3–Q4", "caveat": "PDF P74 的文档时间节省数据来自另一个制造业客户，不能作为该基金项目的验收效果。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-017", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "业务与产品需求", "detail": "老板、研发、产品、HR", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "真实操作诊断", "detail": "验证团队能力与差距", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "定制Spec文档", "detail": "逐步拆解、客户确认", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "AI Coding实践", "detail": "围绕真实研发任务使用工具", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "五天半现场培训", "detail": "教学、咨询、观察与调整", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "三期持续跟进", "detail": "扩展至产品与业务人员", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "共同协作规范", "detail": "从个人工具到组织方法", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "工程质量门禁", "detail": "学习补全：测试与评审", "inferred": true } ], "svg": "architectures/case-017.svg", "mermaid": "architectures/case-017.mmd", "note": "notes/case-017.md" }, { "id": "018", "title": "央国企财务流程 AI 化", "person": "毛卫俊", "book": 9, "pages": [ 88, 96 ], "evidence": "PDF P90–92，Q3", "industry": "科研机构", "originalIndustry": "政务/公共服务", "business": "财务/人力", "types": [ "文档视觉理解", "Agent编排", "流程自动化", "规则计算/优化" ], "pains": [ "重复劳动", "数据分散", "交付慢", "质量与合规风险" ], "summary": "钉钉接收报销材料，关联出差、部门和项目，AI判断归类并监督风险，RPA在原系统录单。", "insight": "把“判断怎么做”和“实际点系统”分开：AI做判断，RPA做操作，风险交给人。", "boundary": "多AI角色监督不能等同独立审计；不确定业务需人处理，不代表可取消财务审批。", "missing": "未公开监督阈值、审批制度、模型配置和RPA失败处理细节。", "exercise": "发票金额正确但归属项目不明时，系统应继续提交还是停下来？设计异常分支。", "g": [ [ "钉钉报销入口|员工消息与发票", "既有业务信息|出差、审批、部门、项目" ], [ "发票识别与上下文关联|整理财务判断所需信息" ], [ "AI归类判断|科目、项目与财务规则", "AI监督与风险识别|复查不确定结果" ], [ "RPA浏览器操作|创建单据、填写、提交" ], [ "集团原财务系统|承载后续原有业务流程", "人工风险处理|异常关注与专业责任" ] ], "links": [ "0a>1a", "0b>1a", "1a>2a", "2a>2b", "2b>3a", "2b>4b", "3a>4a" ], "delivery": { "summary": "钉钉机器人 + 财务系统 RPA 流程", "platforms": [ { "name": "钉钉", "role": "员工发起业务的机器人入口" } ], "details": "从钉钉入口提交单据，经过业务知识、AI 判断与风险控制后，利用 RPA 操作集团既有财务 Web 系统。", "evidence": "PDF P90–92，Q3–Q4", "caveat": "钉钉是用户入口；集团财务系统品牌没有披露，风险事项仍需监督与人工判断。" }, "acceptance": { "status": "实际使用反馈", "summary": "业务运行与风险复核；正式验收未披露", "details": "原文以多个已落地应用和业务耗时变化说明效果，运行中保留风控监督；没有公开正式验收指标表或签字程序。", "evidence": "PDF P91–93，Q3–Q4", "caveat": "原文未披露正式项目验收清单、完整达标阈值或签字记录；业务复核与使用反馈不等于正式验收通过。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-018", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "钉钉报销入口", "detail": "员工消息与发票", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "既有业务信息", "detail": "出差、审批、部门、项目", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "发票识别与上下文关联", "detail": "整理财务判断所需信息", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "AI归类判断", "detail": "科目、项目与财务规则", "inferred": false }, { "id": "n2b", "col": 2, "row": 1, "label": "AI监督与风险识别", "detail": "复查不确定结果", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "RPA浏览器操作", "detail": "创建单据、填写、提交", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "集团原财务系统", "detail": "承载后续原有业务流程", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "人工风险处理", "detail": "异常关注与专业责任", "inferred": false } ], "svg": "architectures/case-018.svg", "mermaid": "architectures/case-018.mmd", "note": "notes/case-018.md" }, { "id": "019", "title": "非标制造图纸到报价", "person": "Stephen / 崔逸枫", "book": 17, "pages": [ 169, 180 ], "evidence": "PDF P173–177，Q3与挑战", "industry": "制造业", "originalIndustry": "制造业", "business": "运营提效", "types": [ "文档视觉理解", "规则计算/优化", "Agent编排" ], "pains": [ "知识难传承", "重复劳动", "交付慢", "判断口径不一致" ], "summary": "多模态AI先把图纸变成结构化需求，再匹配企业物料、供应商、工艺与报价规则，每单人工校准。", "insight": "端到端直接出价对不上真实成本后，模型被收回到信息结构化，定价回到企业规则。", "boundary": "每单校准后才对外报价；图片中不同行业界面不代表同一家客户全部部署。", "missing": "未披露成本公式、工艺规则完整性、向量库选型和人工回填如何更新模型。", "exercise": "图纸上有模糊的手写公差，怎样阻断它传播成错误的最终报价？", "g": [ [ "客户PDF / CAD|材料、尺寸、结构要求", "企业内部经验|物料、供应商、历史报价" ], [ "多模态解析|字段提取与需求结构化", "工艺与报价规则库|企业自身成本和映射" ], [ "垂类Agent编排|需求、工艺与规则匹配" ], [ "规则匹配与成本计算|不能凭模型常识报价", "校准结果回填|知识、数据或相关模型" ], [ "业务人员逐单校准|不确定字段与真实成本", "对外报价|经过人工校准的结果" ] ], "links": [ "0a>1a", "0b>1b", "1a>2a", "1b>2a", "2a>3a", "3a>4a", "4a>4b", "4a>3b", "3b>1b" ], "delivery": { "summary": "垂直售前 Agent · 图纸解析与报价流程", "platforms": [], "details": "解析非标需求与图纸、调用企业知识和报价逻辑，生成供业务人员校准的报价成果，保留人工最终核定。", "evidence": "PDF P173–176，Q3–Q4", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "分阶段试运行", "summary": "真实订单试用 → 小组试运行 → 部门推广", "details": "先由关键业务人员拿真实订单检验可用性，再由业务小组试运行一至两个月，依据解析、报价质量及响应速度反馈调整后推广。", "evidence": "PDF P175–176，Q3–Q4", "caveat": "这是明确披露的业务验证与推广路径，不等于已公布合同验收清单；每次报价仍需人工校准。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-019", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "客户PDF / CAD", "detail": "材料、尺寸、结构要求", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "企业内部经验", "detail": "物料、供应商、历史报价", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "多模态解析", "detail": "字段提取与需求结构化", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "工艺与报价规则库", "detail": "企业自身成本和映射", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "垂类Agent编排", "detail": "需求、工艺与规则匹配", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "规则匹配与成本计算", "detail": "不能凭模型常识报价", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "校准结果回填", "detail": "知识、数据或相关模型", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "业务人员逐单校准", "detail": "不确定字段与真实成本", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "对外报价", "detail": "经过人工校准的结果", "inferred": false } ], "svg": "architectures/case-019.svg", "mermaid": "architectures/case-019.mmd", "note": "notes/case-019.md" }, { "id": "020", "title": "生物科技销售经营数智化", "person": "跳舞", "book": 18, "pages": [ 181, 188 ], "evidence": "PDF P183–185，Q3", "industry": "生物科技", "originalIndustry": "生物科技", "business": "销售提效", "types": [ "数据分析/预警", "组织能力建设", "协作工作台" ], "pains": [ "数据分散", "经营不可见", "知识难传承", "采纳与协同困难" ], "summary": "先用CRM和数仓统一销售数据，再用BI与AI做经营解释；另以陪练、考试等应用支撑人员能力。", "insight": "销售提成首先是数据问题，不必为了AI而AI；基础数据可用之后再做经营智能。", "boundary": "招聘筛选和自动一面仅按资料记为探索；录用、经营与上岗决定需由负责人承担，不以模型评分代决策。", "missing": "未披露CRM/数仓/BI产品、招聘评估依据和数据权限；20个内部编制属组织建设，非系统节点。", "exercise": "把“本月销售为什么下降”拆成可计算事实与需要销售人员补充的业务解释。", "g": [ [ "客户、订单、绩效|原纸账与个人记录", "销售日报与知识|周报、拜访、培训材料" ], [ "CRM与数仓|先统一销售数据基础", "培训知识与上岗标准|练习、考试、反馈" ], [ "AI经营分析|结合数字与记录找原因", "AI陪练 / 招聘探索|能力练习、筛选辅助" ], [ "BI销售大盘|看趋势、区域与经营结果", "电子化考试|串起训练和验证" ], [ "管理者经营判断|理解原因后调整行动", "~人事复核与申诉|学习补全：高影响决策保护" ] ], "links": [ "0a>1a", "0b>2a", "1a>2a", "1a>3a", "2a>3a", "3a>4a", "0b>1b", "1b>2b", "2b>3b", "3b>4b" ], "delivery": { "summary": "CRM / BI 看板 + 销售辅助与培训流程", "platforms": [], "details": "打通 CRM、ERP、仓储及分析视图，向销售提供经营信息，并将 AI 辅助用于新人训练和考核环节。", "evidence": "PDF P183–185，Q3–Q4", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "实际使用与培训反馈", "summary": "看板业务使用 + 新人训练和考试反馈", "details": "从业务人员能否获得所需信息、训练与考试中的问题等实际反馈进行调整；没有披露整体项目的正式验收报告。", "evidence": "PDF P184–186，Q3–Q5", "caveat": "新人考试是培训结果检查，不等于软件项目验收；访谈中的业务改进不代表独立因果验证。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-020", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "客户、订单、绩效", "detail": "原纸账与个人记录", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "销售日报与知识", "detail": "周报、拜访、培训材料", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "CRM与数仓", "detail": "先统一销售数据基础", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "培训知识与上岗标准", "detail": "练习、考试、反馈", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "AI经营分析", "detail": "结合数字与记录找原因", "inferred": false }, { "id": "n2b", "col": 2, "row": 1, "label": "AI陪练 / 招聘探索", "detail": "能力练习、筛选辅助", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "BI销售大盘", "detail": "看趋势、区域与经营结果", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "电子化考试", "detail": "串起训练和验证", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "管理者经营判断", "detail": "理解原因后调整行动", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "人事复核与申诉", "detail": "学习补全：高影响决策保护", "inferred": true } ], "svg": "architectures/case-020.svg", "mermaid": "architectures/case-020.mmd", "note": "notes/case-020.md" }, { "id": "021", "title": "鞋服电商素材工作流", "person": "潘俊杰（木水）", "book": 12, "pages": [ 121, 130 ], "evidence": "PDF P124–127，Q3", "industry": "零售/电商", "originalIndustry": "零售/电商", "business": "营销增长", "types": [ "内容生成", "文档视觉理解", "协作工作台" ], "pains": [ "重复劳动", "质量与合规风险", "数据分散", "采纳与协同困难" ], "summary": "将商品、素材和生成流程统一起来，AI初检配合人工确认，把一次次生成沉淀为公司级素材资产。", "insight": "电商图片要先保证卖的就是图中的商品；人物、服饰和场景一致性比单纯好看更关键。", "boundary": "核心功能完成后处于POC试用迭代，不能当作已全公司成熟运行或已核验降本。", "missing": "未披露生成模型、素材管理产品、一致性量化标准和正式商业效果。", "exercise": "为同款服饰设计三项验收标准，哪些错即便画面好看也必须退回？", "g": [ [ "商品与品牌资料|商品细节、人物、场景", "员工生产反馈|审美要求与操作问题" ], [ "统一商品管理|商品与素材建立关联" ], [ "AI图像/视频生成|进入商品素材工作流" ], [ "AI第一轮检查|人物、服饰、场景一致性" ], [ "人工视觉与品牌确认|最后判断是否可商用", "公司级素材资产|可复用图片、视频和经验" ] ], "links": [ "0a>1a", "1a>2a", "0b>2a", "2a>3a", "3a>4a", "4a>4b", "4b>1a" ], "delivery": { "summary": "商品 / 素材管理 + 图片与视频工作流", "platforms": [], "details": "把商品、素材和生成流程纳入企业工作方式，形成可供设计人员挑选与使用的图片、视频及素材资产。", "evidence": "PDF P124–127，Q3–Q4", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "原文有验收标准 · POC", "summary": "人物 / 服饰 / 场景一致性，AI 初检 + 人审", "details": "原文明确将三类一致性纳入验收标准，先由 AI 检查，再由人工确认品牌审美与业务可用性；核心功能完成后进入员工试用、反馈和迭代的 POC。", "evidence": "PDF P126–127，Q3–Q4", "caveat": "有验收标准不等于项目已最终验收通过；原文仍在 POC 阶段，没有披露完整量化阈值和正式签字记录。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-021", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "商品与品牌资料", "detail": "商品细节、人物、场景", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "员工生产反馈", "detail": "审美要求与操作问题", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "统一商品管理", "detail": "商品与素材建立关联", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "AI图像/视频生成", "detail": "进入商品素材工作流", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "AI第一轮检查", "detail": "人物、服饰、场景一致性", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "人工视觉与品牌确认", "detail": "最后判断是否可商用", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "公司级素材资产", "detail": "可复用图片、视频和经验", "inferred": false } ], "svg": "architectures/case-021.svg", "mermaid": "architectures/case-021.mmd", "note": "notes/case-021.md" }, { "id": "022", "title": "制造企业经营流程透明化", "person": "谢尧", "book": 20, "pages": [ 200, 209 ], "evidence": "PDF P203–205，Q3与图2", "industry": "制造业", "originalIndustry": "制造业", "business": "运营提效", "types": [ "数据分析/预警", "流程自动化" ], "pains": [ "数据分散", "经营不可见", "重复劳动" ], "summary": "保留原ERP，按老板关心的订单到回款链路重新组织数据，匹配核对后通过飞书卡片提醒异常。", "insight": "数据已有但关系不可见，也会造成管理盲区；不必替换ERP才能改善经营视角。", "boundary": "正常数据持续匹配，异常由人处理；刚上线后的长期经营改善未充分验证。", "missing": "未披露全量匹配逻辑、异常阈值、同步时延及写回策略。", "exercise": "一笔供应商付款找不到对应验收记录，如何向负责人展示证据而非只发“异常”两字？", "g": [ [ "原有ERP|订单、采购、库存、财务", "经营关注点|价格、库存、应付与回款" ], [ "只读同步与关系重组|不重建原ERP" ], [ "AI数据匹配核对|跨业务环节关联与解释" ], [ "业务标准与异常识别|价格、库存、付款是否对应", "飞书卡片通知|主动提醒相关负责人" ], [ "老板经营看板|看状态、看关系、看异常", "负责人处理异常|作决定并跟进处理" ] ], "links": [ "0a>1a", "0b>3a", "1a>2a", "2a>3a", "1a>4a", "3a>3b", "3b>4b" ], "delivery": { "summary": "飞书异常卡片 + 老板经营看板", "platforms": [ { "name": "飞书", "role": "异常卡片通知入口" } ], "details": "保留 ERP，以只读数据同步串联订单、库存、生产等信息，形成经营看板和可追踪的异常卡片通知。", "evidence": "PDF P203–206，Q3–Q4", "caveat": "ERP 是保留的数据来源；飞书承载异常通知，不据此推断整套经营看板都是飞书原生模块。" }, "acceptance": { "status": "初期上线验证", "summary": "异常追到业务单据；长期效果待验证", "details": "通过跨系统数据核对与异常追踪支持经营判断，观察实际使用和问题跟进；访谈时刚上线，长期经营改善尚未充分验证。", "evidence": "PDF P204–207，Q3–Q5", "caveat": "原文未披露正式项目验收清单、完整达标阈值或签字记录；业务复核与使用反馈不等于正式验收通过。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-022", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "原有ERP", "detail": "订单、采购、库存、财务", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "经营关注点", "detail": "价格、库存、应付与回款", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "只读同步与关系重组", "detail": "不重建原ERP", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "AI数据匹配核对", "detail": "跨业务环节关联与解释", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "业务标准与异常识别", "detail": "价格、库存、付款是否对应", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "飞书卡片通知", "detail": "主动提醒相关负责人", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "老板经营看板", "detail": "看状态、看关系、看异常", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "负责人处理异常", "detail": "作决定并跟进处理", "inferred": false } ], "svg": "architectures/case-022.svg", "mermaid": "architectures/case-022.mmd", "note": "notes/case-022.md" }, { "id": "023", "title": "消费品研发协同与包装审核", "person": "王一棵", "book": 19, "pages": [ 189, 199 ], "evidence": "PDF P193–196，Q3", "industry": "零售/电商", "originalIndustry": "零售/电商", "business": "运营提效", "types": [ "文档视觉理解", "规则计算/优化", "协作工作台", "流程自动化" ], "pains": [ "质量与合规风险", "交付慢", "判断口径不一致", "采纳与协同困难" ], "summary": "设计稿进入产品生命周期工作流后，拆分视觉区域做OCR/VLM处理，规则代码校验数值，人工决定产品表达。", "insight": "热量、NRV等确定性计算固化成代码；模型解释与补录，结论保留来源和置信度。", "boundary": "人决定审美、表达及专业合规结论；业务群观察是材料记载，不等于可无告知收集员工信息。", "missing": "未披露具体法规库版本、OCR/VLM供应商和全量测试集。", "exercise": "把一张包装拆成文字、营养表、艺术字区域：每块用什么方法，如何追溯错误？", "g": [ [ "包装设计稿|文字、营养表、艺术字", "岗位与历史经验|渠道反馈、驳回整改记录" ], [ "视觉模块拆分|不同区域选OCR / VLM", "规则与历史上下文|确定性标准、上线样本" ], [ "解释与补录建议|模型不替代确定性计算" ], [ "脚本化规则校验|NRV、热量、数字一致性", "预审与审计记录|来源、置信度、风险与降级" ], [ "产品/设计人员判断|表达、专业结论与整改", "包装管理 / 产研流程|连接设计、采购和生产" ] ], "links": [ "0a>1a", "0b>1b", "1a>2a", "1a>3a", "1b>3a", "2a>3b", "3a>3b", "3b>4a", "4a>4b", "4b>1b" ], "delivery": { "summary": "包装审核 Skill + 包装 / PLM 协同系统", "platforms": [], "details": "将 OCR、视觉理解与确定性计算规则组合为审核能力，嵌入包装管理及产研协同流程，在设计阶段提前检查并记录反馈。", "evidence": "PDF P193–196，Q3–Q4", "caveat": "飞书群是观察需求和工作过程的场所，不能据此写成最终交付平台；具名 PLM 产品未披露。" }, "acceptance": { "status": "业务质量检查", "summary": "规则计算可复核、依据可追溯 + 人审", "details": "热量、营养成分等规则通过确定性代码计算，审核提示保留来源和依据，结合历史驳回与人工反馈迭代。", "evidence": "PDF P194–197，Q3–Q5", "caveat": "包装的业务审核不等于该 AI 项目的正式验收；原文未公开项目测试阈值及最终验收记录。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-023", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "包装设计稿", "detail": "文字、营养表、艺术字", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "岗位与历史经验", "detail": "渠道反馈、驳回整改记录", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "视觉模块拆分", "detail": "不同区域选OCR / VLM", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "规则与历史上下文", "detail": "确定性标准、上线样本", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "解释与补录建议", "detail": "模型不替代确定性计算", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "脚本化规则校验", "detail": "NRV、热量、数字一致性", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "预审与审计记录", "detail": "来源、置信度、风险与降级", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "产品/设计人员判断", "detail": "表达、专业结论与整改", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "包装管理 / 产研流程", "detail": "连接设计、采购和生产", "inferred": false } ], "svg": "architectures/case-023.svg", "mermaid": "architectures/case-023.mmd", "note": "notes/case-023.md" }, { "id": "024", "title": "工业物料检索与采购减浪费", "person": "马诺", "book": 8, "pages": [ 78, 87 ], "evidence": "PDF P81–83，Q3", "industry": "制造业", "originalIndustry": "制造业", "business": "供应链", "types": [ "知识检索", "流程自动化" ], "pains": [ "数据分散", "供需匹配低效", "判断口径不一致", "质量与合规风险" ], "summary": "先清洗ERP物料主数据，再用语义检索对齐不同叫法，让员工找到已有物料，减少重复录入与误采购。", "insight": "搜索不好只是表面；要追到重复采购、库存积压与报废，才能确定值得优先解决。", "boundary": "最终是否采购、采购多少仍由业务人员决定；没有披露可核验的最终降本金额。", "missing": "未披露嵌入模型、向量库、ERP写回方案和财务损失验证数据。", "exercise": "同一个螺丝有三种叫法时，怎样既提高召回，又避免把相近但不同规格合并？", "g": [ [ "历史ERP物料|重复、缺失、不一致描述", "员工自然语言需求|同物不同叫法" ], [ "数据清洗与标准化|去重、口径对齐、干净底座" ], [ "语义检索|理解业务语义、匹配物料" ], [ "已有物料结果|回到可识别的物料条目" ], [ "业务人员选择|判断是否采购与采购量", "~录入防重复校验|学习补全：保存前候选提示" ] ], "links": [ "0a>1a", "0b>2a", "1a>2a", "2a>3a", "3a>4a", "4b>1a" ], "delivery": { "summary": "ERP 物料数据治理 + 语义搜索能力", "platforms": [], "details": "清洗并标准化物料主数据，提供围绕 ERP 物料的语义检索，帮助采购和仓储找到可用物料。", "evidence": "PDF P81–84，Q3–Q4", "caveat": "原文未披露具名交付平台，不根据技术架构猜填品牌。" }, "acceptance": { "status": "基线与小范围验证", "summary": "历史基线 → 小范围实验 → 客户业务复核", "details": "以历史情况建立基线，小范围验证重复采购、错采购与浪费是否减少，并和客户业务人员确认计算口径和业务价值。", "evidence": "PDF P83–84，Q4", "caveat": "原文披露了业务价值验证方法，但没有公开完整数值、正式合同验收门槛或签字报告。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": "https://fde100.datawhale.cn/cases/cases-024", "source": "双来源", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "历史ERP物料", "detail": "重复、缺失、不一致描述", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "员工自然语言需求", "detail": "同物不同叫法", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "数据清洗与标准化", "detail": "去重、口径对齐、干净底座", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "语义检索", "detail": "理解业务语义、匹配物料", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "已有物料结果", "detail": "回到可识别的物料条目", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "业务人员选择", "detail": "判断是否采购与采购量", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "录入防重复校验", "detail": "学习补全：保存前候选提示", "inferred": true } ], "svg": "architectures/case-024.svg", "mermaid": "architectures/case-024.mmd", "note": "notes/case-024.md" }, { "id": "025", "title": "外贸企业文件资产化与本地适配", "person": "Nature张晨曦", "book": 11, "pages": [ 108, 120 ], "evidence": "PDF P112–115，Q3；仅册子收录", "industry": "商贸/外贸", "originalIndustry": null, "business": "资料治理 / 企业能力建设", "types": [ "协作工作台", "Agent编排", "组织能力建设" ], "pains": [ "数据分散", "知识难传承", "采纳与协同困难", "质量与合规风险" ], "summary": "梳理客户、产品、项目之间的文件关系，把框架和任务包送到客户自己的Codex环境，由现场数据驱动适配。", "insight": "不是把客户数据打包拿走，而是交付约束、规范和任务，让开发在客户环境继续发生。", "boundary": "“文件留在客户环境”不等于完全离线推理；客户侧AI反馈日志给FDE，传输范围需另行约定。", "missing": "未披露具体云模型数据流、文件权限实现和生产知识检索服务。", "exercise": "设计一个不包含原始合同的反馈包，让外部FDE仍能定位问题并迭代任务规范。", "g": [ [ "客户内网与文件|合同、产品、客户资料", "FDE框架与任务包|Harness、原则、规范、Demo" ], [ "文件关系整理|客户—产品—项目关联", "客户自己的Codex|在客户文件环境适配" ], [ "本地开发与适配|按规范运行任务" ], [ "老板测试与AI整理反馈|日志、问题、记录", "FDE调整任务规范|依据反馈更新框架" ], [ "企业文件与工作能力|逐步沉淀Workflow / Skill", "~反馈包脱敏门禁|学习补全：限制对外内容" ] ], "links": [ "0a>1a", "0b>1b", "1a>2a", "1b>2a", "2a>3a", "3a>4b", "4b>3b", "3b>1b", "2a>4a" ], "delivery": { "summary": "客户自己的 Codex · Harness 任务包", "platforms": [ { "name": "Codex", "role": "客户侧任务包运行与适配环境" } ], "details": "把框架、原则、规范与 Demo 组织为 Harness 任务包，交给客户自己的 Codex 适配本地文件与工作流，并用问题日志继续迭代。", "evidence": "PDF P112–115，Q3–Q4", "caveat": "本地文件和客户侧 Codex 不等于模型完全离线运行；最终覆盖范围需以实际任务包与测试为准。" }, "acceptance": { "status": "客户实测与迭代", "summary": "老板亲测 → 问题日志反馈 → 更新任务包", "details": "客户侧完成适配后由老板测试，把实际问题、日志与改进反馈交给 FDE，再更新规范和任务包。", "evidence": "PDF P113–116，Q3–Q5", "caveat": "这是访谈披露的客户测试闭环，没有公开最终签字、全量测试清单或统一达标阈值。" }, "sourceFamily": "datawhale", "sourceLabel": "Datawhale", "kind": "案例拆解", "reviewStatus": "已拆解", "url": null, "source": "仅册子", "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "客户内网与文件", "detail": "合同、产品、客户资料", "inferred": false }, { "id": "n0b", "col": 0, "row": 1, "label": "FDE框架与任务包", "detail": "Harness、原则、规范、Demo", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "文件关系整理", "detail": "客户—产品—项目关联", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "客户自己的Codex", "detail": "在客户文件环境适配", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "本地开发与适配", "detail": "按规范运行任务", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "老板测试与AI整理反馈", "detail": "日志、问题、记录", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "FDE调整任务规范", "detail": "依据反馈更新框架", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "企业文件与工作能力", "detail": "逐步沉淀Workflow / Skill", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "反馈包脱敏门禁", "detail": "学习补全：限制对外内容", "inferred": true } ], "svg": "architectures/case-025.svg", "mermaid": "architectures/case-025.mmd", "note": "notes/case-025.md" }, { "id": "026", "title": "跨境供应链风险 Agent", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/cross-border-supply-risk-agent", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 1, "sourceLabel": "FDE中国社区", "kind": "试点方案", "reviewStatus": "已拆解", "industry": "供应链/采购", "originalIndustry": "供应链", "business": "供应链风险识别与采购审批", "types": [ "Agent编排", "规则计算/优化", "数据分析/预警" ], "pains": [ "数据分散", "判断口径不一致", "交付慢" ], "evidence": "FDE中国社区：跨境供应链风险 Agent，公开页面，读取于2026-09-08", "summary": "汇总订单、库存、供应商交期与物流状态，用规则评估交付风险，给采购经理生成可审批的应对方案。", "insight": "跨系统查询可以并行，追加采购、客户承诺与供应商变更仍需业务审批。", "boundary": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。", "missing": "未披露具名客户、真实上线数据、最终签字记录或完整复现代码；以公开方案为学习材料。", "exercise": "让一条清关延迟订单同时触发风险评分、依据展示与采购审批，如何防止重复下单？", "g": [ [ "订单与风险扫描|用户提问或定时触发" ], [ "库存、交期与物流|按角色查询业务数据", "风险规则|禁运、节假日等约束" ], [ "Agent规划与查询|组合只读业务工具", "规则引擎评分|风险判断保留依据" ], [ "应对方案摘要|改交期、加急或替代料" ], [ "采购经理审批|不自动承诺与下单", "~ERP草稿写回|原文可选高权限步骤" ] ], "links": [ "0a>2a", "1a>2a", "1b>2b", "2a>2b", "2b>3a", "3a>4a", "4a>4b" ], "delivery": { "summary": "风险分析界面 + 审批摘要 / PO 草稿", "platforms": [], "details": "方案交付风险分数、依据字段和应对建议；ERP 写回是可选高权限能力，界面品牌没有披露。", "evidence": "FDE中国社区：架构、工具清单、Demo 脚本", "caveat": "ERP 是系统类别而非具名产品；采购执行须审批、权限校验与幂等保护。" }, "acceptance": { "status": "方案验收目标", "summary": "风险召回 ≥90%；工具成功率 ≥95%", "details": "以 Golden Dataset、越权与无审批采购反例验证；目标还包括解释字段准确率 ≥85%、建议采纳率 ≥60%，配合采购经理试点与工具审计。", "evidence": "FDE中国社区：Golden Dataset、验收指标、30 天计划", "caveat": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。" }, "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "订单与风险扫描", "detail": "用户提问或定时触发", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "库存、交期与物流", "detail": "按角色查询业务数据", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "风险规则", "detail": "禁运、节假日等约束", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "Agent规划与查询", "detail": "组合只读业务工具", "inferred": false }, { "id": "n2b", "col": 2, "row": 1, "label": "规则引擎评分", "detail": "风险判断保留依据", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "应对方案摘要", "detail": "改交期、加急或替代料", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "采购经理审批", "detail": "不自动承诺与下单", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "ERP草稿写回", "detail": "原文可选高权限步骤", "inferred": true } ], "svg": "architectures/case-026.svg", "mermaid": "architectures/case-026.mmd", "note": "notes/case-026.md" }, { "id": "027", "title": "客服工单智能分流 PoC", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/customer-ticket-routing", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 2, "sourceLabel": "FDE中国社区", "kind": "试点方案", "reviewStatus": "已拆解", "industry": "跨行业/客服", "originalIndustry": "客服", "business": "客服工单分类与回复辅助", "types": [ "知识检索", "流程自动化", "内容生成" ], "pains": [ "重复劳动", "判断口径不一致", "交付慢" ], "evidence": "FDE中国社区：客服工单智能分流 PoC，公开页面，读取于2026-09-08", "summary": "对新工单分类并检索 FAQ、SOP 和订单状态，生成回复草稿，再由坐席确认或升级人工队列。", "insight": "第一版以草稿模式进入原流程，退款赔付等高风险工单优先保证不漏转人工。", "boundary": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。", "missing": "未披露具名客户、真实上线数据、最终签字记录或完整复现代码；以公开方案为学习材料。", "exercise": "用退款威胁、过期知识与跨产品线访问三个样本验证分流边界。", "g": [ [ "新工单|用户问题与订单上下文" ], [ "FAQ与SOP知识库|按产品线与地区过滤", "订单状态查询|按权限补充业务上下文" ], [ "分类与信息抽取|标签、优先级与风险", "RAG知识推荐|带引用的处理依据" ], [ "回复草稿|默认不直接发送" ], [ "坐席确认发送|人工修改与品牌话术", "高风险人工队列|退款、赔付等升级" ] ], "links": [ "0a>2a", "1a>2b", "1b>2b", "2a>2b", "2b>3a", "3a>4a", "2a>4b" ], "delivery": { "summary": "工单辅助界面 · 知识卡片与回复草稿", "platforms": [], "details": "交付分类结果、知识引用、订单状态卡片和坐席可编辑草稿，接入人工升级队列。", "evidence": "FDE中国社区：架构、Demo 脚本", "caveat": "未披露工单系统或聊天平台品牌；退款、赔付等敏感事项强制转人工。" }, "acceptance": { "status": "方案验收目标", "summary": "分类 ≥88%；高风险转人工准确率 ≥92%", "details": "以历史工单和人工处理记录建立评估集；方案目标还包括首响时间缩短 ≥30%、坐席采纳率 ≥65%，在 10 名坐席范围试点。", "evidence": "FDE中国社区：验收指标、30 天计划", "caveat": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。" }, "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "新工单", "detail": "用户问题与订单上下文", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "FAQ与SOP知识库", "detail": "按产品线与地区过滤", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "订单状态查询", "detail": "按权限补充业务上下文", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "分类与信息抽取", "detail": "标签、优先级与风险", "inferred": false }, { "id": "n2b", "col": 2, "row": 1, "label": "RAG知识推荐", "detail": "带引用的处理依据", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "回复草稿", "detail": "默认不直接发送", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "坐席确认发送", "detail": "人工修改与品牌话术", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "高风险人工队列", "detail": "退款、赔付等升级", "inferred": false } ], "svg": "architectures/case-027.svg", "mermaid": "architectures/case-027.mmd", "note": "notes/case-027.md" }, { "id": "028", "title": "工厂运营日报智能体", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/factory-operation-report-agent", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 3, "sourceLabel": "FDE中国社区", "kind": "试点方案", "reviewStatus": "已拆解", "industry": "制造业", "originalIndustry": "制造", "business": "工厂早会日报与异常跟进", "types": [ "流程自动化", "知识检索", "数据分析/预警" ], "pains": [ "数据分散", "重复劳动", "经营不可见" ], "evidence": "FDE中国社区：工厂运营日报智能体，公开页面，读取于2026-09-08", "summary": "从生产、订单、库存和停机记录计算日报指标，检索 SOP 解释异常，生成由运营主管确认的报告与待办。", "insight": "指标交给表格工具计算，解释必须引用数据与 SOP，停线和排产不交给模型直接执行。", "boundary": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。", "missing": "未披露具名客户、真实上线数据、最终签字记录或完整复现代码；以公开方案为学习材料。", "exercise": "当生产数据延迟时，日报怎样区分真异常与数据尚未到齐？", "g": [ [ "定时早会任务|生产、订单、库存、停机" ], [ "表格计算工具|统一指标与业务口径", "版本化SOP|异常处理依据" ], [ "RAG与解释生成|引用数据和操作规范" ], [ "日报与待办草稿|指标卡、异常与建议" ], [ "运营主管确认|停线与排产必须人审", "IM或邮件通知|审批、幂等与脱敏" ] ], "links": [ "0a>1a", "1a>2a", "1b>2a", "2a>3a", "3a>4a", "4a>4b" ], "delivery": { "summary": "日报与待办草稿 · 审批后推送 IM / 邮件", "platforms": [], "details": "以定时工作流生成指标卡、异常解释和待办清单，经运营主管确认后通知相关人员。", "evidence": "FDE中国社区：架构、Demo 脚本", "caveat": "IM 和邮件是渠道类别，未指定飞书、钉钉等品牌；停线与排产调整须人审。" }, "acceptance": { "status": "方案验收目标", "summary": "指标准确率 ≥98%；日报生成 <5 分钟", "details": "与人工核算对照；目标包括异常召回率 ≥85%、两周试点的人工采纳率 ≥70%，并覆盖越权查询和停线建议的风险样本。", "evidence": "FDE中国社区：Golden Dataset、验收指标", "caveat": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。" }, "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "定时早会任务", "detail": "生产、订单、库存、停机", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "表格计算工具", "detail": "统一指标与业务口径", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "版本化SOP", "detail": "异常处理依据", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "RAG与解释生成", "detail": "引用数据和操作规范", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "日报与待办草稿", "detail": "指标卡、异常与建议", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "运营主管确认", "detail": "停线与排产必须人审", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "IM或邮件通知", "detail": "审批、幂等与脱敏", "inferred": false } ], "svg": "architectures/case-028.svg", "mermaid": "architectures/case-028.mmd", "note": "notes/case-028.md" }, { "id": "029", "title": "财务报表智能分析 PoC", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/finance-report-analysis", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 4, "sourceLabel": "FDE中国社区", "kind": "试点方案", "reviewStatus": "已拆解", "industry": "跨行业/财务", "originalIndustry": "财务", "business": "财务报表勾稽与分析草稿", "types": [ "规则计算/优化", "知识检索", "内容生成" ], "pains": [ "重复劳动", "判断口径不一致", "质量与合规风险" ], "evidence": "FDE中国社区：财务报表智能分析 PoC，公开页面，读取于2026-09-08", "summary": "解析财务报表科目，由工具计算同比、环比与预算差，再结合会计政策生成可追溯的分析草稿。", "insight": "财务数字由确定性工具计算，结论返回来源单元格；模型不直接出具审计意见。", "boundary": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。", "missing": "未披露具名客户、真实上线数据、最终签字记录或完整复现代码；以公开方案为学习材料。", "exercise": "如何发现跨表公式链接错误，并阻止错误数字被写进分析结论？", "g": [ [ "财务Excel与政策|三大报表、预算、说明" ], [ "科目树解析|校验公式与报表结构", "政策知识检索|会计口径与引用" ], [ "指标计算与异常规则|同比、环比及预算差" ], [ "分析说明草稿|每条结论关联单元格" ], [ "财务编辑复核|不自动出审计结论", "Word报告导出|经人工确认的成果" ] ], "links": [ "0a>1a", "0a>1b", "1a>2a", "1b>3a", "2a>3a", "3a>4a", "4a>4b" ], "delivery": { "summary": "报表上传分析界面 · 财务复核后导出 Word", "platforms": [], "details": "展示科目树、异常计算过程和单元格引用，生成由财务编辑的分析说明，再导出 Word 文件。", "evidence": "FDE中国社区：架构、Demo 脚本", "caveat": "Excel / Word 是输入输出格式，不代表采用某个具名智能体平台；私有化或专网是方案约束。" }, "acceptance": { "status": "方案验收目标", "summary": "计算 ≥99%；引用可追溯率 100%", "details": "与专家标注、对照报表验证；目标还包括异常识别准确率 ≥85%、报告撰写时间缩短 ≥40%，在月结中试点并复盘失败样本。", "evidence": "FDE中国社区：验收指标、30 天计划", "caveat": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。" }, "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "财务Excel与政策", "detail": "三大报表、预算、说明", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "科目树解析", "detail": "校验公式与报表结构", "inferred": false }, { "id": "n1b", "col": 1, "row": 1, "label": "政策知识检索", "detail": "会计口径与引用", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "指标计算与异常规则", "detail": "同比、环比及预算差", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "分析说明草稿", "detail": "每条结论关联单元格", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "财务编辑复核", "detail": "不自动出审计结论", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "Word报告导出", "detail": "经人工确认的成果", "inferred": false } ], "svg": "architectures/case-029.svg", "mermaid": "architectures/case-029.mmd", "note": "notes/case-029.md" }, { "id": "030", "title": "政企公文抽取 PoC", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/gov-document-extraction", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 5, "sourceLabel": "FDE中国社区", "kind": "试点方案", "reviewStatus": "已拆解", "industry": "政务/公共服务", "originalIndustry": "政企", "business": "公文事项提取与督办台账", "types": [ "文档视觉理解", "知识检索", "流程自动化" ], "pains": [ "重复劳动", "数据分散", "质量与合规风险" ], "evidence": "FDE中国社区：政企公文抽取 PoC，公开页面，读取于2026-09-08", "summary": "从公文和会议纪要中提取事项、责任部门与截止时间，保留原文页段引用，由人工校对后写入台账。", "insight": "模糊责任和相对日期必须允许待确认，不能为让表格完整而强行补齐。", "boundary": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。", "missing": "未披露具名客户、真实上线数据、最终签字记录或完整复现代码；以公开方案为学习材料。", "exercise": "纪要只有“有关部门月底前报送”时，哪些字段能确定，哪些必须转人工？", "g": [ [ "公文、纪要与附件|按密级分流和输入拦截" ], [ "OCR与版式解析|表格还原与文档分段" ], [ "Schema约束抽取|事项、责任部门和日期", "原文溯源|定位页码与段落" ], [ "人工校验界面|模糊字段标待确认" ], [ "台账与CSV|确认后写入，不自动发文" ] ], "links": [ "0a>1a", "1a>2a", "1a>2b", "2a>3a", "2b>3a", "3a>4a" ], "delivery": { "summary": "人工校验界面 + 台账系统 / CSV 导出", "platforms": [], "details": "提供解析结果、结构化字段和原文高亮定位；校对后写入台账或导出 CSV，不自动发文。", "evidence": "FDE中国社区：架构、Demo 脚本", "caveat": "原文未披露台账产品品牌；涉密材料阻断公有云处理，写入前必须人审。" }, "acceptance": { "status": "方案验收目标", "summary": "字段准确率 ≥90%；遗漏率 ≤5%", "details": "目标还包括引用段落准确率 ≥95%，四周试点观察人工修改率下降，并用模糊指代、涉密附件等样本检验边界。", "evidence": "FDE中国社区：Golden Dataset、验收指标", "caveat": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。" }, "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "公文、纪要与附件", "detail": "按密级分流和输入拦截", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "OCR与版式解析", "detail": "表格还原与文档分段", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "Schema约束抽取", "detail": "事项、责任部门和日期", "inferred": false }, { "id": "n2b", "col": 2, "row": 1, "label": "原文溯源", "detail": "定位页码与段落", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "人工校验界面", "detail": "模糊字段标待确认", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "台账与CSV", "detail": "确认后写入，不自动发文", "inferred": false } ], "svg": "architectures/case-030.svg", "mermaid": "architectures/case-030.mmd", "note": "notes/case-030.md" }, { "id": "031", "title": "专项攻略┃从零教Agent做标书，我是怎么搭出这条流水线的｜AI实战", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/de-2026072210756", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 6, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数字员工", "originalIndustry": "数字员工", "business": "数字员工主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 6 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：上周有个朋友问我：她一个人要做几份标书，每份100多页，时间紧得要命。她想让AI帮忙，但试了几次发现——", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 6 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 6 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-031.md" }, { "id": "032", "title": "一个文件夹 = 一个 Agent 员工：如何用文件夹重构 AI 时代的工作。", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/de-2026072279165", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 7, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数字员工", "originalIndustry": "数字员工", "business": "数字员工主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 7 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：盯着十几个工作文件夹发了会儿呆。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 7 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 7 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-032.md" }, { "id": "033", "title": "Dify 1.16.0解读：Agent、Linux沙箱和MCP协议升级，Dify开始进入真正的Agent运行时", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/dify-2026072221839", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 8, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/智能体工具", "originalIndustry": "智能体应用", "business": "智能体应用主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 8 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：本文基于 Dify( 官方 GitHub 1.16.0 release note（2026 年 7月 17 日发布）整理，重点看这次版本…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 8 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 8 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-033.md" }, { "id": "034", "title": "制造业AI落地：专家系统、AI知识库、AI智能体，高管，老板必须清楚的3个概念", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/industry-2026072137814", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 9, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "制造业", "originalIndustry": "工业制造", "business": "工业制造主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 9 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：一把手必须亲自盯的5个关键", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 9 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 9 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-034.md" }, { "id": "035", "title": "MooxAI：十来人如何跑出 AI 原生组织", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/aisaas-2026072062704", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 10, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "软件/企业服务", "originalIndustry": "AI SaaS", "business": "AI SaaS主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 10 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：Elric Liu 提到这里时，只是顺口带过。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 10 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 10 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-035.md" }, { "id": "036", "title": "Dify近三个月技术演进分析——从应用工具到AI基础设施", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/dify-2026072016403", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 11, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/智能体工具", "originalIndustry": "智能体应用", "business": "智能体应用主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 11 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：Dify的RAG在三个月内完成了三层叠加。从架构层面看，每一次升级都不是简单的\"加个功能\"，而是在检索管道的特定环节插入了新的处理层。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 11 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 11 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-036.md" }, { "id": "037", "title": "马斯克的xAI推出Voice Agent，可以免费获得+1美国电话号，帮你打造专属AI电话客服", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/cs-2026072093681", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 12, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/客服", "originalIndustry": "智能客服", "business": "智能客服主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 12 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：刚才看到X上的分享，马斯克的xAI（现在应该叫SpaceXAI）最近上线了Grok Voice Agent Builder，可以几乎零门槛、几分钟就能创建自己的AI语音代理，还可以获得一个（+1）美国…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 12 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 12 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-037.md" }, { "id": "038", "title": "WAIC观察：企业 Agent 落地完整指南", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/dify-2026071802534", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 13, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/智能体工具", "originalIndustry": "智能体应用", "business": "智能体应用主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 13 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：7 月 18 日 WAIC 世界人工智能大会多场企业 AI 专题论坛同步开讲，阿里云、腾讯云均把「企业内部标准化搭建 Agent、实现业务全链路降本增效」作为核心分享主题之一。两家云厂商现场完整披露了…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 13 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 13 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-038.md" }, { "id": "039", "title": "你的 AI 项目为什么死在生产环境：3 个真实团队的工程教训", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/cs-2026071729876", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 14, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/客服", "originalIndustry": "智能客服", "business": "智能客服主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 14 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：接个大模型做个 Demo，4 个方框就够了——用户、WebSocket、LLM、再回到用户。架构图看着很 AI。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 14 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 14 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-039.md" }, { "id": "040", "title": "AI数字员工最值钱的8个场景：2026年全球高ROI落地案例全拆解", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/de-2026071706845", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 15, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数字员工", "originalIndustry": "数字员工", "business": "数字员工主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 15 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：AI Agent的讨论，2026年已经不需要再证明\"有用\"了。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 15 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 15 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-040.md" }, { "id": "041", "title": "高价率运营 AI 工作台：约定驱动与 AI 编排的评测优化实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-ai-workbench", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 16, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 16 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：本文介绍了“高价率运营 AI 工作台”中基于“约定驱动 + AI 编排”架构的自动化评测优化实践。针对 LLM Agent 场景下传统测试失效、反馈周期长等痛点，文章详细阐述了将评测体系作为核心工程产…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 16 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 16 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-041.md" }, { "id": "042", "title": "企微AI助理“大圆”上线后，事情开始变得有趣了", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/de-2026071753642", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 17, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数字员工", "originalIndustry": "数字员工", "business": "数字员工主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 17 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：上个月，企业微信智能助理「大圆」正式开启内测。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 17 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 17 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-042.md" }, { "id": "043", "title": "5%的企业AI试点能成功，问题不在模型，在打通知识孤岛的公司大脑，Glean把企业知识串成一张可搜索的图谱，让 AI 开始\"思考\"", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/glean-enterprise-brain", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 18, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/知识管理", "originalIndustry": "知识管理", "business": "知识管理主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 18 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：一、前言：一个刺眼的 5%", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 18 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 18 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-043.md" }, { "id": "044", "title": "淘宝闪购-爆品团精排Scaling Up迭代实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-taobao-shangou-scaling", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 19, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 19 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：文章内容基于作者个人技术实践与独立思考，旨在分享经验，仅代表个人观点。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 19 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 19 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-044.md" }, { "id": "045", "title": "从\"机械应答\"到\"服务伙伴\"：得物高可控智能客服的 Agent 工程实践｜AICon 演讲整理", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/cs-2026071515026", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 20, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/客服", "originalIndustry": "智能客服", "business": "智能客服主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 20 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：二、智能客服 Agent 落地的必然性与挑战", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 20 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 20 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-045.md" }, { "id": "046", "title": "AI落地的深水区：3个真实案例的6个月踩坑与脱坑", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/cs-2026071331974", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 21, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/客服", "originalIndustry": "智能客服", "business": "智能客服主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 21 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：这家企业是一家有 500+ 门店的连锁零售品牌，会员超过 500 万。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 21 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 21 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-046.md" }, { "id": "047", "title": "AI时代，咨询顾问如何构建核心能力", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/de-2026071281935", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 22, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数字员工", "originalIndustry": "数字员工", "business": "数字员工主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 22 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：2026 年，百年头部咨询巨头迎来发展里程碑，行业却迎来明显寒冬：过去 18 个月全球范围内咨询团队缩编比例超 10%；一线咨询机构全面调整内部职级体系，多家头部企业统一暂缓应届生薪资上调。 这场行业…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 22 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 22 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-047.md" }, { "id": "048", "title": "当下企业落地AI，瓶颈已然不在大模型——真正的护城河是数字孪生体", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-2026071170935", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 23, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/FDE方法", "originalIndustry": "FDE 实践", "business": "FDE 实践主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 23 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：最近在给几家企业做 AI 落地。做得越深，越确信一件事：", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 23 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 23 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-048.md" }, { "id": "049", "title": "【揭秘】如何打造一支凌晨3点还在交付的AI军团", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/de-2026071094207", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 24, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数字员工", "originalIndustry": "数字员工", "business": "数字员工主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 24 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：凌晨三点，没有人在群里催进度，也没有人在电脑前等结果。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 24 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 24 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-049.md" }, { "id": "050", "title": "WorkBuddy与ERP集成协同落地实施方案", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/de-2026070905682", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 25, "sourceLabel": "FDE中国社区", "kind": "实施方案文章", "reviewStatus": "已拆解", "industry": "跨行业/数字员工", "originalIndustry": "数字员工", "business": "WorkBuddy 与 ERP 集成实施", "types": [ "Agent编排", "流程自动化", "协作工作台" ], "pains": [ "数据分散", "重复劳动", "采纳与协同困难" ], "evidence": "FDE中国社区：WorkBuddy与ERP集成协同落地实施方案，公开页面，读取于2026-09-08", "summary": "一份以 WorkBuddy 和企业微信为操作入口的 ERP 集成方案：API 优先、数据库只读补充、无接口时以 RPA 兜底。", "insight": "交付不是只有智能体，还包含连接配置、字段映射、技能手册、权限矩阵、联调报告与验收确认单。", "boundary": "这是公开试点方案，验收数字是目标，不是已达成的项目业绩；未披露具名客户的最终验收记录。", "missing": "未披露具名客户、真实上线数据、最终签字记录或完整复现代码；以公开方案为学习材料。", "exercise": "选一个销售报价到订单草稿场景，设计权限隔离、审批、重试与人工回退的验收用例。", "g": [ [ "企微中的业务人员|销售、采购、仓储、财务" ], [ "ERP主数据与接口|权限拆分、编码与口径" ], [ "WorkBuddy技能流程|文章提出的编排入口" ], [ "API优先连接|查询与经审批制单", "只读库或RPA|按ERP条件选择替代路径" ], [ "业务草稿与通知|风险事项人工审批", "测试与交付文档|联调、灰度、手册、验收单" ] ], "links": [ "0a>2a", "1a>3a", "1a>3b", "2a>3a", "2a>3b", "3a>4a", "3b>4a", "4a>4b" ], "delivery": { "summary": "方案：WorkBuddy + 企业微信 + ERP 集成包", "platforms": [ { "name": "WorkBuddy", "role": "文章提出的智能体操作入口" }, { "name": "企业微信", "role": "文章提出的员工入口与通知渠道" } ], "details": "建议交付连接配置、字段映射、工作流技能模板、业务 SOP、联调测试报告、权限与脱敏说明、运维手册及上线验收确认单。", "evidence": "FDE中国社区原文：六「测试、灰度上线与切换」；八「项目完整交付物清单」", "caveat": "这是公开实施方案，不是某个具名客户已落地证明。文中的产品功能、连接器路径和 ERP 兼容性需另查厂商资料；不把列举的所有 ERP 都标成已部署。" }, "acceptance": { "status": "方案中的验收流程", "summary": "测试库联调 → 单部门灰度 7 天 → 验收单", "details": "方案要求覆盖八类业务场景，检查单据、同步、异常拦截、权限和预警；问题闭环后灰度试用，再逐部门扩展并交付测试报告与验收确认单。", "evidence": "FDE中国社区原文：六「测试、灰度上线与切换」；八「项目完整交付物清单」", "caveat": "这是文章建议的流程及交付清单，没有提供真实客户签署的验收单；业务提效与三周工期也不作已验证承诺。" }, "nodes": [ { "id": "n0a", "col": 0, "row": 0, "label": "企微中的业务人员", "detail": "销售、采购、仓储、财务", "inferred": false }, { "id": "n1a", "col": 1, "row": 0, "label": "ERP主数据与接口", "detail": "权限拆分、编码与口径", "inferred": false }, { "id": "n2a", "col": 2, "row": 0, "label": "WorkBuddy技能流程", "detail": "文章提出的编排入口", "inferred": false }, { "id": "n3a", "col": 3, "row": 0, "label": "API优先连接", "detail": "查询与经审批制单", "inferred": false }, { "id": "n3b", "col": 3, "row": 1, "label": "只读库或RPA", "detail": "按ERP条件选择替代路径", "inferred": false }, { "id": "n4a", "col": 4, "row": 0, "label": "业务草稿与通知", "detail": "风险事项人工审批", "inferred": false }, { "id": "n4b", "col": 4, "row": 1, "label": "测试与交付文档", "detail": "联调、灰度、手册、验收单", "inferred": false } ], "svg": "architectures/case-050.svg", "mermaid": "architectures/case-050.mmd", "note": "notes/case-050.md" }, { "id": "051", "title": "淘宝内容生态：GrowBrain - 淘宝Agentic内容成长引擎", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-taobao-growbrain", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 26, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 26 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：文章内容基于作者个人技术实践与独立思考，旨在分享经验，仅代表个人观点。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 26 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 26 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-051.md" }, { "id": "052", "title": "来自硅谷一线创业者的 FDE 非共识和落地指南 | AI透镜-行业圆桌06", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-2026070807812", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 27, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/FDE方法", "originalIndustry": "FDE 实践", "business": "FDE 实践主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 27 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：2026 年 6 月 30 日，AWS 宣布投入 10 亿美元组建 FDE 组织，把工程师以 45 天冲刺的方式嵌入客户团队。两天后，微软成立 Frontier Company，投入 25 亿美元、6…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 27 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 27 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-052.md" }, { "id": "053", "title": "Palantir AI FDE如何重构三一工业资产审计范式", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-2026070807468", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 28, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/FDE方法", "originalIndustry": "FDE 实践", "business": "FDE 实践主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 28 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：各位朋友们好，久未见面。最近，参与了一家大型企业的“SAP替换”咨询项目。由于项目正处于极其关键的蓝图规划与架构设计阶段，工作节奏近乎疯狂，导致公众号停更了近三个月。在此，向一直关注本号的朋友致以最诚…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 28 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 28 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-053.md" }, { "id": "054", "title": "Dify真的适合企业级大型AI Agent应用开发落地吗？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/dify-2026070841297", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 29, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/智能体工具", "originalIndustry": "智能体应用", "business": "智能体应用主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 29 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：我估计，很多团队一聊企业 AI 落地，第一个想到的平台就是 Dify。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 29 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 29 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-054.md" }, { "id": "055", "title": "从Palantir发布的案例，我看到三个转型信号", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-2026070708193", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 30, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/FDE方法", "originalIndustry": "FDE 实践", "business": "FDE 实践主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 30 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：Palantir的客户案例，表面上在讲AI落地，实际上每一刀都切在组织和人身上。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 30 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 30 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-055.md" }, { "id": "056", "title": "他给银行写了一年尽调报告，发现AI最难的不是快，是准", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-bank-due-diligence", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 31, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 31 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：如果一份发债说明书里有一个数据错了，后果是什么？", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 31 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 31 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-056.md" }, { "id": "057", "title": "实测扣子3.0：把Claude Code、CodeX拉到一个项目群里干活，夯爆了", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/coze-2026070680379", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 32, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/智能体工具", "originalIndustry": "智能体应用", "business": "智能体应用主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 32 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：学会用Agent、Vibe Coding，几乎承包了科技圈这半年里最热门的流量。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 32 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 32 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-057.md" }, { "id": "058", "title": "3个中小银行推进AI的经验分享：先做小场景，再补数据治理", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-small-banks-experience", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 33, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 33 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：银行业谈AI，正在进入一个更务实的阶段。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 33 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 33 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-058.md" }, { "id": "059", "title": "Cozi 智能体 + 工作流组合拳：复杂场景下的 AI 应用搭建指南", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/coze-2026070582094", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 34, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/智能体工具", "originalIndustry": "智能体应用", "business": "智能体应用主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 34 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：在上一篇我们彻底理清了Cozi 智能体和工作流的适用边界：智能体擅长处理 “不确定、复杂、需要交互” 的场景，工作流适合 “固定、标准化、高可控” 的任务。但很多朋友留言问：“如果我的场景既需要智能体…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 34 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 34 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-059.md" }, { "id": "060", "title": "可信仿真评估：工业 Physical AI 后训练落地的关键基础设施", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/industry-2026070309658", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 35, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "制造业", "originalIndustry": "工业制造", "business": "工业制造主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 35 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：当前工业具身智能面临落地过程中评估体系与真实场景脱节等多重问题，制约后训练专属模型从实验室走向产线落地。本文提出基于NVIDIA Isaac Lab的可信仿真评估体系，构建双层验证机制、三层解耦架构与…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 35 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 35 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-060.md" }, { "id": "061", "title": "让 AI 真正接入 CAD：一次 Onshape + LLM 的云台工作流实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/industry-2026070228195", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 36, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "制造业", "originalIndustry": "工业制造", "business": "工业制造主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 36 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：这类项目最难的地方，不是画出一个“看起来像云台”的模型，而是不断回答一些非常具体的问题：", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 36 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 36 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-061.md" }, { "id": "062", "title": "保险极客-团险理赔数字化项目——AI理赔重塑团险数字化：把\"成本中心\"变成\"增长引擎\"", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-insurance-claims", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 37, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 37 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：把\"成本中心\"变成\"增长引擎\"", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 37 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 37 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-062.md" }, { "id": "063", "title": "搭建Skill的“产品平衡法则”：人、大模型、规则脚本", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-2026070109475", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 38, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 38 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：上一篇我介绍了在AI时代爆发的11款开源爬虫工具，好多人问：这些工具，能拿来干嘛？", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 38 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 38 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-063.md" }, { "id": "064", "title": "Firebase 深度接入智能体生态，赋能全栈应用构建", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/aisaas-2026062679046", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 39, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "软件/企业服务", "originalIndustry": "AI SaaS", "business": "AI SaaS主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 39 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：如今，我们正迈入以智能体为主导的开发时代。智能体正从编码助手演变为能够管理整个软件开发生命周期的智能协作伙伴。在 Google I/O 2026 上，我们发布了一系列更新，帮助您在常用的 AI 工具和…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 39 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 39 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-064.md" }, { "id": "065", "title": "Text-To-CAD: AI Agent自主CAD设计工程实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/industry-2026062170821", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 40, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "制造业", "originalIndustry": "工业制造", "business": "工业制造主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 40 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：前几篇，我们一直在讲 CAD AI 的一个判断：", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 40 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 40 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-065.md" }, { "id": "066", "title": "咨询｜为OPC建的OS级操作系统，把中后台都打通的YC创业项目：Result；在中国有没有跑通的可能性？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/aisaas-2026061961803", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 41, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "软件/企业服务", "originalIndustry": "AI SaaS", "business": "AI SaaS主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 41 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：过去两年，“OPC（One Person Company，一人公司）”几乎变成了中国创业语境里的新常用语。你很难再把它简单理解为“个体户”或者“自由职业者”。在大模型、云服务和各种工具的加持下，很多人…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 41 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 41 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-066.md" }, { "id": "067", "title": "面向Skills编程-淘宝企业购端对端研发提效实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-taobao-skills-rnd", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 42, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 42 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：本文提出\"面向Skills编程\"范式——将领域知识、工作流、约束规则封装为可版本化的Skills单元，让LLM在确定性框架内生成代码，实现从\"人写代码\"到\"人沉淀Skills，AI写代码\"的研发范式升…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 42 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 42 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-067.md" }, { "id": "068", "title": "小鹏集团发布X-Foresight：学习世界知识赋能智驾，开创预测模型新范式", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/auto-xpeng-x-foresight", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 43, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "汽车/出行", "originalIndustry": "汽车出行", "business": "汽车出行主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 43 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：大语言模型在编程、数学领域的能力突破，源于文本模态承载了对应领域的全部知识，模型通过海量文本训练即可完成知识学习。知识的承载模态，决定了模型的学习与建模方式。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 43 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 43 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-068.md" }, { "id": "069", "title": "别把 Palantir 的“本体”做成知识图谱：企业数据架构的务实落地指南", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-2026060746835", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 44, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/FDE方法", "originalIndustry": "FDE 实践", "business": "FDE 实践主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 44 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：最近在企业数据服务的圈子里，有一个非常明显的趋势：学 Palantir 的厂家越来越多了。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 44 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 44 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-069.md" }, { "id": "070", "title": "从腾讯文档升级看腾讯在Agent赛道上的节奏", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/aisaas-2026060558476", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 45, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "软件/企业服务", "originalIndustry": "AI SaaS", "business": "AI SaaS主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 45 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：2025 年，姚顺雨写过一篇传播很广的文章，叫《The Second Half》。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 45 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 45 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-070.md" }, { "id": "071", "title": "如何从零打造AI-native服务公司丨Y Combinator", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/aisaas-2026060536821", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 46, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "软件/企业服务", "originalIndustry": "AI SaaS", "business": "AI SaaS主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 46 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：\"未来十年，一些最大的公司根本不会是软件公司。\"", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 46 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 46 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-071.md" }, { "id": "072", "title": "【广发金工】财务分析Skills的创建与应用", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-gf-analysis-skills", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 47, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 47 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：在本文中，我们围绕“行业财报自动化分析”构建了两个互相协同的研究技能（skill）。第一个skill是“ Excel 财务分析Skill”，主要解决行业样本财务数据的标准化拉取、同比环比计算、分组比较…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 47 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 47 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-072.md" }, { "id": "073", "title": "Workbuddy投研迎来史诗级加强", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-workbuddy-research", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 48, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 48 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：最近，workbuddy接入了通达信的MCP服务，让workbuddy的用户终于可以免费用到靠谱的金融数据了。（希望有接入wind的那一天，或者在付费订阅中加入wind数据库的API）", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 48 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 48 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-073.md" }, { "id": "074", "title": "MuleRun Messages上线！Agent从个人工具走进团队协作", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/aisaas-2026060254278", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 49, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "软件/企业服务", "originalIndustry": "AI SaaS", "business": "AI SaaS主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 49 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：今天，MuleRun正式上线Messages。作为MuleRun Enterprise版的AI协作IM，Messages的核心设计是让人类员工与AI Agent在同一个工作空间里像同事一样协作——Ag…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 49 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 49 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-074.md" }, { "id": "075", "title": "扣子 3.0 正式上线：新一代 AI 团队，从扣子开始", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/coze-2026060185314", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 50, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/智能体工具", "originalIndustry": "智能体应用", "business": "智能体应用主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 50 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：今天，扣子 3.0 正式上线，扣子手机端 （iOS / Android）、电脑端（Mac OS / Windows）、网页端（coze.cn）三端全量更新。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 50 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 50 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-075.md" }, { "id": "076", "title": "Anthropic 亲自示范，如何使用金融Skills", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-anthropic-skills", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 51, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 51 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：Anthropic 像是憋了大招，专攻金融行业，我之前都有介绍：", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 51 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 51 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-076.md" }, { "id": "077", "title": "Qwen3-8B +GraphRAG 在医疗领域的应用", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-qwen3-graphrag", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 52, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 52 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：不知道读到这篇文章的朋友们，有没有感受到，目前一些较大的门诊医院中，已经存在了 AI 导诊的功能了，目的就是为了能在就诊前能提前收集一些用户的信息。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 52 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 52 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-077.md" }, { "id": "078", "title": "拦截率从15%到55%：快手智能Oncall系统演进与落地实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/cs-2026042937926", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 53, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/客服", "originalIndustry": "智能客服", "business": "智能客服主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 53 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：在《采纳率从3%到80%：智能单元测试生成的进化之路》中，快手研发效能团队分享了如何通过智能单测架构的多次演进升级，将AI从提效工具进化为质量伙伴。当AI成功在代码源头构筑起质量防线后，如何让它继续向…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 53 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 53 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-078.md" }, { "id": "079", "title": "刚刚，扣子搞了个大的：你的满配 AI 伙伴已就位！（附场景案例）", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/coze-2026040750697", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 54, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/智能体工具", "originalIndustry": "智能体应用", "business": "智能体应用主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 54 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：上次写完扣子 OpenClaw 的文章之后，后台收到最多的一个问题就是：扣子龙虾这么快就来了，然后呢？", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 54 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 54 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-079.md" }, { "id": "080", "title": "给 AI 配张银行卡：Stripe 半年搭完的 Agent 支付全景", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-stripe-agent-payment", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 55, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 55 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：3 月中旬，一个叫 AgentCard 的产品在 X 上刷屏了：给你的 AI Agent 发一张虚拟 Visa 借记卡，Agent 自己刷卡买东西，用完自动销毁。底层用的是 Stripe Issuin…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 55 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 55 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-080.md" }, { "id": "081", "title": "一起奔跑，奔赴下一站：构建AI时代的工业数据基座", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/industry-2026030895068", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 56, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "制造业", "originalIndustry": "工业制造", "business": "工业制造主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 56 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：就在一周前，TDengine TSDB 时序数据库在全球安装实例数超过一百万，遍布全球60多个国家和地区，这是一个相当令人兴奋的数字。2016年底，我看到物联网、工业互联网的高速发展，市场需要有一款高…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 56 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 56 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-081.md" }, { "id": "082", "title": "为什么顶尖投行都选择了 Rogo 这个金融 Agent？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-rogo-agent", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 57, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 57 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：全球投行业每年处理超过 3.5 万亿美元的交易，但驱动这台庞大机器运转的，是数以万计每周工作超过百小时、从事着高度重复性劳动的初级分析师。Vertical Agent 开始加速很多专业领域的工作流，比…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 57 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 57 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-082.md" }, { "id": "083", "title": "从Agent Skills到下一代智能风控中台", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-risk-control-platform", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 58, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 58 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：主要涉及修改其配置文件 SKILL.md 和核心逻辑代码 scripts/scorecardpipeline.py。以下是具体的更新方法，您可以根据需求选择：", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 58 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 58 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-083.md" }, { "id": "084", "title": "刚刚，OpenAI 把“健康”变成长期记忆，苹果生态可能第一个接入", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-openai-long-memory", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 59, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 59 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：就在 “ChatGPT 健康” 刚刚对接 Apple 健康之际，OpenAI 又做了一件非常关键、但很多人还没反应过来的事：", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 59 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 59 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-084.md" }, { "id": "085", "title": "为什么说Ontology才是工业智能的\"操作系统\"？一个从救火到预防的真实案例", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/industry-2026011337810", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 60, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "制造业", "originalIndustry": "工业制造", "business": "工业制造主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 60 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：10 月 24 日，在介绍和解读党的二十届四中全会精神的新闻发布会上，科技部部长阴和俊强调，要“全面实施‘人工智能+’行动，全方位赋能千行百业”。这一系列政策信号清晰表明：工业智能化正从局部探索，迈入…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 60 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 60 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-085.md" }, { "id": "086", "title": "智“会”锦囊：大模型重塑咨询响应新范式", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/cs-2026011003571", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 61, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/客服", "originalIndustry": "智能客服", "business": "智能客服主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 61 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：银行数字化转型中，行内系统的稳定高效运行直接关乎业务流转与客户体验。当前部分系统咨询类工单数量居高不下，一线业务人员因系统报错、操作存疑提交的诉求单量大，导致运维与开发人员深陷重复咨询处理，难以聚焦核…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 61 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 61 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-086.md" }, { "id": "087", "title": "京东上线「京东AI购」，AI正在重新定义购物方式！", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/cs-2025122991478", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 62, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/客服", "originalIndustry": "智能客服", "business": "智能客服主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 62 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：就在前两天，京东上线了一款APP-京东AI购，这款APP软件的定位是“会思考，能购物，关心你的生活服务助手”。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 62 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 62 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-087.md" }, { "id": "088", "title": "Glean解析：企业AI搜索的革新者如何重塑知识工作", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/glean-2025122110729", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 63, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/知识管理", "originalIndustry": "知识管理", "business": "知识管理主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 63 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：引导语：企业内部信息如孤岛，员工如何一秒调取全网级知识？Glean如何凭企业智能搜索，在2025年估值飙至近百亿美金，成为硅谷新宠？", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 63 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 63 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-088.md" }, { "id": "089", "title": "技术赋能医药全链路：AI 大模型应用在药企的落地痛点与破局之道", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-pharma-landing-pain", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 64, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 64 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：AI 技术浪潮下的制药行业变革", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 64 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 64 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-089.md" }, { "id": "090", "title": "AI时代的SaaS变革：从工具到结果的跨越", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/aisaas-2025120312697", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 65, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "软件/企业服务", "originalIndustry": "AI SaaS", "business": "AI SaaS主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 65 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：「X-Day 行业圆桌」是腾讯投资和腾讯青腾合力打造，面向“腾讯投后+青腾校友+行业专家”的闭门研讨交流平台，立足于前沿科技、深耕产业链，共创学习，分享认知。本期嘉宾有——", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 65 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 65 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-090.md" }, { "id": "091", "title": "森马如何用阿里云 AI 网关，轻松实现“AI+业务”高效落地", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-semir-ai-gateway", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 66, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 66 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：森马快速实现 AI 转型，通过阿里云 AI 网关（即 Higress 企业版）及注册配置中心 Nacos3.0 实现了多模型多 MCP server 统一接入统一管理统一配置，将存量服务一键转换为 M…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 66 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 66 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-091.md" }, { "id": "092", "title": "当AI开始替用户做决定，品牌如何成为“默认答案”？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-brand-default-answer", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 67, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 67 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：问答智能诞生记：当AI开始替用户做决定，品牌如何成为“默认答案”？", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 67 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 67 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-092.md" }, { "id": "093", "title": "深度解析丨Shopify 如何用“AI团队”搞定万亿级电商商品分类", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-shopify-ai-team", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 68, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 68 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：现在，许多企业都在积极寻找 AI 的落地场景，但真正有价值的应用并不多见。在这方面，Shopify 利用 AI Agent 实现智能化商品分类的案例，有很多借鉴得地方，分享给大家。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 68 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 68 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-093.md" }, { "id": "094", "title": "Palantir 创始工程师深度分享：FDE 模式是 Agent 时代的 PMF 范式", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-fde-agent-pmf", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 69, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/FDE方法", "originalIndustry": "FDE 实践", "business": "FDE 实践主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 69 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：FDE （Foward Deployed Engineer，前沿部署工程师）模式是 Palantir 独特的 GTM 和产品部署模式，但当下它已经成为硅谷最热门的话题之一，在 YC 的一次统计中，目前…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 69 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 69 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-094.md" }, { "id": "095", "title": "AI Agent加速SaaS智能化重构 打造企业服务新格局", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/aisaas-2025101304138", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 70, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "软件/企业服务", "originalIndustry": "AI SaaS", "business": "AI SaaS主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 70 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：在AI浪潮席卷企业服务的今天，一个根本性问题正以前所未有的紧迫感叩击每一家SaaS厂商：当用户不再打开软件、不再点击按钮，而是直接与智能体对话“帮我完成一个任务”，软件的存在意义是否正在被重新定义？", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 70 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 70 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-095.md" }, { "id": "096", "title": "OpenAI上线购物Agent，AI早已改写电商逻辑", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-2025100160497", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 71, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 71 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：“OpenAI进军电商”的新闻引发了很多关注：", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 71 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 71 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-096.md" }, { "id": "097", "title": "五个 AI 落地典型案例：垂直场景、极致性价比与硬件创新", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-five-landing-cases", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 72, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 72 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：「Liblib：中国领先的AI创作者社区」", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 72 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 72 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-097.md" }, { "id": "098", "title": "AI驱动下的医药产业端变革—基于华兴伙伴的洞察 | 华兴观点", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-pharma-transformation", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 73, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 73 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：AI在科研与医疗领域的深度应用，根植于其技术体系的持续迭代。从早期依赖人工设计规则的传统模式，到生成式、多模态AI的应用，如今具备自主感知、决策与执行能力的Agent智能形态，AI技术已完成多轮关键突…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 73 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 73 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-098.md" }, { "id": "099", "title": "从体验到系统工程丨上手评测国内首款 AI 电商 App", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-first-ai-app-review", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 74, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 74 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：产品界面，往往体现了产品的设计哲学，界面是产品的第一入口。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 74 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 74 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-099.md" }, { "id": "100", "title": "Palantir如何驱动采购领域的AI自动化变革？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-procurement-automation", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 75, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "供应链/采购", "originalIndustry": "供应链/采购", "business": "供应链/采购主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 75 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：在上一篇文章《Palantir AIP 驱动医疗行业变革，HCA的AI医院管理实践启示》中，我们深入探讨了Palantir AIP如何驱动美国最大医院集团HCA实现“AI医院”的革命性管理。从手术室调…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 75 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 75 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-100.md" }, { "id": "101", "title": "搞企业AI落地，一张A4纸就能开干，空膜拜Palantir没用｜本体工程重生｜Ontology RAG", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-ontology-rag", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 76, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/FDE方法", "originalIndustry": "FDE 实践", "business": "FDE 实践主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 76 条，读取于2026-09-08（仅目录收录）", "summary": "原站主题：FDE 实践。已收录原文入口，具体交付与验收信息待研读。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 76 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 76 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-101.md" }, { "id": "102", "title": "Palantir Foundry：AI赋能RaceTrac革新便利店体验", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-foundry-racetrac", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 77, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "零售", "business": "零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 77 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：想象一下，当二月寒风凛冽之时，突然有一天阳光普照，气温飙升至近30摄氏度。对于便利店来说，这不仅仅是天气变化，更是顾客行为模式的巨大转变：对冰淇淋的渴望、对冷饮的需求、甚至是对墨镜的即时购买欲都将激增…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 77 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 77 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-102.md" }, { "id": "103", "title": "AI技术在汽车合规导向数模设计中的应用", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/auto-compliant-model-design", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 78, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "汽车/出行", "originalIndustry": "汽车出行", "business": "汽车出行主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 78 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：导读：企业依托人工智能AI技术，深度融合大数据与UG NX二次开发技术，精准识别法规符合性检查项、智能提取设计/校验参数、将设计/校验过程转化为操作便利的工程开发软件工具，实现合规导向的数模智能化设计…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 78 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 78 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-103.md" }, { "id": "104", "title": "1688 AI 原生应用的创新探索与实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-1688-ai-native", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 79, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 79 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：1.AI 在电商场景的应用", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 79 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 79 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-104.md" }, { "id": "105", "title": "AI Agent 在电商内容场景的应用及演进", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-agent-content-scene", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 80, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 80 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：3.1.0 时代落地挑战与解决思路", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 80 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 80 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-105.md" }, { "id": "106", "title": "Palantir FDE的一天：我们不仅仅是工程师，更是问题的解决者", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-fde-a-day", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 81, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/FDE方法", "originalIndustry": "FDE 实践", "business": "FDE 实践主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 81 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：你对硅谷的神秘公司Palantir了解多少？它以服务政府情报机构和大型企业而闻名，以其强大的数据分析平台而著称。在Palantir的众多工程师角色中，有一个独特的岗位——Forward Deploye…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 81 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 81 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-106.md" }, { "id": "107", "title": "AI获客实战指南：从流量到转化的智能升级", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-2025070436097", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 82, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 82 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：很多人都在讲AI获客，本质是什么？", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 82 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 82 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-107.md" }, { "id": "108", "title": "医疗行业怎么落地 AI？Abridge AI：重新定义医疗文档记录的智能化变革", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-abridge-documentation", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 83, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 83 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：在医疗AI领域快速发展的浪潮中，Abridge正在成为最受关注的明星企业之一。这家来自匹兹堡的初创公司通过其革命性的环境式AI技术，正在彻底改变医生与患者对话的记录方式，为医疗行业带来前所未有的效率提…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 83 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 83 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-108.md" }, { "id": "109", "title": "AiCure公司，用AI提升患者依从性", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-aicure-adherence", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 84, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 84 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：AiCure, LLC 成立于2009年12月1日，总部位于美国纽约市，是一家专注于将人工智能、计算机视觉和大数据分析应用于临床试验与患者管理的移动软件即服务（SaaS）平台公司。其核心使命是通过 A…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 84 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 84 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-109.md" }, { "id": "110", "title": "医院如何找到大模型落地的“优先场景”？（附视频）", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-hospital-priority-scene", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 85, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 85 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：医疗层面和技术层面形成有机联动，以系统性思维审视问题。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 85 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 85 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-110.md" }, { "id": "111", "title": "Tempus AI：AI 医疗的一个数据引擎？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-tempus-data-engine", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 86, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 86 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：因为对 AI 医疗方向的持续兴趣，最近开始研究 Tempus AI。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 86 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 86 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-111.md" }, { "id": "112", "title": "智能AI电子病历系统产品设计解析：从政策到落地的实战", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-emr-product-design", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 87, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 87 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：前面关于医疗信息化的系统也写了不少，自从DeepSeek出来以后，未来信息化往AI智能方向转已经是未来的趋势了，所以今天观海就从智能AI电子病历系统这个开始，按照信息化系统方向进行解析，希望给各位老师…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 87 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 87 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-112.md" }, { "id": "113", "title": "医疗大模型，集体下沉", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-medical-llm-sinking", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 88, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 88 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：2025年以来，DeepSeek通过开放生态加速了算法研发与临床场景的深度融合。医疗大模型摒弃了“技术至上”的思维，逐渐进入实用主义阶段。此背景下，对提高诊疗效率和质量有迫切需求的基层医疗，成为大模型…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 88 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 88 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-113.md" }, { "id": "114", "title": "Agent-RAG + GraphRAG 双框架，让个性化治疗方案准确率高达 100%", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-agent-graphrag-treatment", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 89, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 89 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：论文：Developing an Artificial Intelligence Tool for Personalized Breast Cancer Treatment Plans based o…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 89 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 89 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-114.md" }, { "id": "115", "title": "DeepSeek医疗多场景应用分析：基于420余家医院的实证研究", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/health-deepseek-420-hospitals", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 90, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 90 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：（1）症状分诊与科室匹配（74次）", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 90 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 90 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-115.md" }, { "id": "116", "title": "Glean：企业AI搜索，估值46亿美元，ARR一年翻4倍", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/glean-enterprise-search", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 91, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/知识管理", "originalIndustry": "知识管理", "business": "知识管理主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 91 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：在 9 月份完成了 2.6 亿美元的 E 轮融资后，主打企业内部 AI 搜索的 Glean 估值达到 46 亿美元。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 91 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 91 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-116.md" }, { "id": "117", "title": "解锁大模型和 AI Agent 在车端的革新应用和落地方法", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/auto-agent-in-car-landing", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 92, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "汽车/出行", "originalIndustry": "汽车出行", "business": "汽车出行主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 92 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：近年来，人工智能技术的发展进入新阶段，大模型在各行各业掀起了智能化变革浪潮。在汽车领域，人工智能（以下简称 AI）的技术驱动能力完美契合了当下汽车的发展需求，作为新一代智能终端，汽车的智能化发展目前有…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 92 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 92 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-117.md" }, { "id": "118", "title": "淘宝AI试衣素材高效写入IC实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-2024101825830", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 93, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 93 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：随着电商行业的快速发展，用户对于在线购物体验的要求也在不断提高。特别是在服装领域，用户希望能够更加直观地了解商品的实际穿着效果。为此，淘宝试衣项目应运而生，通过AI技术为用户提供个性化的试衣体验。本文…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 93 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 93 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-118.md" }, { "id": "119", "title": "Palantir 大模型AIP产品应用于客户服务RAG Copilot的案例", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/palantir-aip-rag-copilot", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 94, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/客服", "originalIndustry": "智能客服", "business": "智能客服主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 94 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：一名医疗器械公司员工在Palantir AIP平台上尝试复现其在Azure上使用GPT和RAG构建的客户服务引擎CoPilot，寻求改进其RAG模型在准确性和效率方面的策略，并得到了社区成员关于改进分…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 94 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 94 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-119.md" }, { "id": "120", "title": "案例：AI大模型如何应用在整车及关键部件质检中？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/auto-vehicle-quality-inspection", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 95, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "汽车/出行", "originalIndustry": "汽车出行", "business": "汽车出行主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 95 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：我国已成为世界新能源车产销第一大国，2023年我国新能源产销分别完成958.7万辆和949.5万辆。新能源车所用到的许多零部件与内燃机原有主要零部件不同，以电力作为动力来源，必须具有电池系统、电驱系统…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 95 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 95 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-120.md" }, { "id": "121", "title": "企业如何落地AI Agent？腾讯、百度等5家头部企业实践案例", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/enterprise-ai-agent-5-cases", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 96, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数字员工", "originalIndustry": "数字员工", "business": "数字员工主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 96 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：大模型的三大涌现能力——上下文理解、指令遵循和逐步推理，推动机器对人的赋能从完成特定任务的Tool形态（例如RPA、OCR），加速向Copilot和Autopilot的形态演进。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 96 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 96 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-121.md" }, { "id": "122", "title": "CPS AI Agent：解锁工业数据融合，加速企业数字化转型新纪元", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/auto-cps-agent-data-fusion", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 97, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "汽车/出行", "originalIndustry": "汽车出行", "business": "汽车出行主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 97 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：该产品集成了多个关键组件，如应用互联智能体（ASB）、开放互联智能体（OSB）、数实融合模型智能体以及工业要素标识解析智能体等，共同构建了一个全面、智能的工业数据处理体系。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 97 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 97 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-122.md" }, { "id": "123", "title": "关于可落地的LLM+BI项目真正的一些梳理", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/bi-2024070130165", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 98, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数据分析", "originalIndustry": "数据分析/BI", "business": "数据分析/BI主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 98 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：在工作中接触/使用大模型已经超过1年了，大部分是做分析/BI相关的场景。过去半年做指标平台+大模型应用项目，最终也有客户实际落地，梳理下自己的一些总结。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 98 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 98 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-123.md" }, { "id": "124", "title": "Quick BI站稳Gartner ABI挑战者象限，打造“人人能用的BI”", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/bi-2024070189574", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 99, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数据分析", "originalIndustry": "数据分析/BI", "business": "数据分析/BI主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 99 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：近日，全球权威IT研究和咨询机构Gartner商业智能与分析平台魔力象限报告（Gartner Magic Quadrant for Analytics and Business Intelligenc…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 99 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 99 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-124.md" }, { "id": "125", "title": "《2024中国制造业智能BI解决方案与案例》报告正式发布", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/bi-2024062501429", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 100, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数据分析", "originalIndustry": "数据分析/BI", "business": "数据分析/BI主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 100 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：制造业作为我国经济的重要支柱，正面临着全球化竞争加剧、消费者需求多样化、资源环境约束加强等多重挑战。在这样的背景下，如何有效利用数据，提升决策质量，优化资源配置，成为制造业数字化转型升级的关键。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 100 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 100 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-125.md" }, { "id": "126", "title": "TOP100 Summit：融合ChatBI与HeadlessBI：腾讯音乐构建下一代数据分析平台的实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/bi-2024062258907", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 101, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数据分析", "originalIndustry": "数据分析/BI", "business": "数据分析/BI主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 101 条，读取于2026-09-08（仅目录收录）", "summary": "原站主题：数据分析/BI。已收录原文入口，具体交付与验收信息待研读。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 101 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 101 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-126.md" }, { "id": "127", "title": "AI+BI融合新篇章：问答式BI 实践与探索", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/bi-2024062147983", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 102, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数据分析", "originalIndustry": "数据分析/BI", "business": "数据分析/BI主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 102 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：AI与BI的结合确实为数据分析领域带来了诸多新的可能性，特别是LLM（大型语言模型）在BI产品中的应用，极大地提升了数据分析的效率和便捷性。以下是对您提到的不同模式的进一步探讨：!(../../ass…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 102 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 102 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-127.md" }, { "id": "128", "title": "AI时代的企业变革：数字员工实战策略", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/digital-employee-strategy", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 103, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数字员工", "originalIndustry": "数字员工", "business": "数字员工主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 103 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：数字员工不仅仅是自动化工具的延伸，它们代表了一种全新的劳动力形态，能够处理大量数据，执行复杂任务，并提供决策支持，从而显著提升企业的工作效率和服务质量。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 103 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 103 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-128.md" }, { "id": "129", "title": "MongoDB和AI 赋能行业应用：制造业和汽车行业", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/auto-mongodb-manufacturing", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 104, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "汽车/出行", "originalIndustry": "汽车出行", "business": "汽车出行主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 104 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：随着人工智能（AI）在制造业和汽车行业的集成，传统的价值链正在经历一场革命性的转变。工业物联网（IoT）的引入使得企业能够从资产中收集和分析大量数据，这不仅提升了库存管理的智能化水平，还使得预测性维护…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 104 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 104 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-129.md" }, { "id": "130", "title": "AI大模型如何重塑汽车后市场？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/auto-aftermarket-reshape", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 105, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "汽车/出行", "originalIndustry": "汽车出行", "business": "汽车出行主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 105 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：汽车后市场当前正面临着数字化转型的重要契机，AI大模型可以凭借其强大的数据处理以及整合能力，对汽车行驶数据和维保记录等进行深度分析，进而显著提升汽车后市场的服务效能与质量。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 105 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 105 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-130.md" }, { "id": "131", "title": "大模型VS小模型，工业质检怎么选？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/industrial-qc-large-vs-small", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 106, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "制造业", "originalIndustry": "工业制造", "business": "工业制造主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 106 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：良品率受到多种因素的影响，包括原材料质量、生产设备状态、工艺技术水平、员工技能水平、管理水平等。因此，要提高良品率，企业需要从多个方面入手，例如优化生产工艺、精益生产流程、采购更先进的生产设备、提高操…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 106 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 106 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-131.md" }, { "id": "132", "title": "透视环世物流：一个AI降本的产业典型样本", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/huanshi-logistics-ai-cost", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 107, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "制造业", "originalIndustry": "工业制造", "business": "工业制造主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 107 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：在过去的多年时间里，与其说低代码完成的数字原生的普惠，不如说其最强的能力恰是能帮助企业用最低的门槛、最高的效率构建出与自身适配的流程和业务应用，在底层帮助企业构建出一个被极致拆分和分子化的软件开发中台…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 107 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 107 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-132.md" }, { "id": "133", "title": "大模型上车！30家国内外车企大模型应用现状", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/auto-30-oems-llm-status", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 108, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "汽车/出行", "originalIndustry": "汽车出行", "business": "汽车出行主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 108 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：大模型正在推动汽车行业向智能化转型，这不仅涉及到车辆功能的升级和用户体验的增强，还包括生产流程、研发设计、营销售后等全产业链的数字化和自动化。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 108 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 108 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-133.md" }, { "id": "134", "title": "如何用大模型升级智能客服？字节、美敦力、工行等7家企业实践案例", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/smart-cs-7-enterprise-cases", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 109, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/客服", "originalIndustry": "智能客服", "business": "智能客服主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 109 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：大模型已经在多个领域展现出强大的语言理解和生成能力。在智能客服场景中，这种能力被广泛应用于理解用户提出的问题、生成与用户需求匹配的回复，甚至是预测用户行为和提供个性化推荐。这不仅极大地提高了客户服务的…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 109 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 109 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-134.md" }, { "id": "135", "title": "企业生成式AI：2024 年企业的 10 多个用例和最佳实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/enterprise-genai-usecases-2024", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 110, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/组织转型", "originalIndustry": "企业转型", "business": "企业转型主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 110 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：与中型市场公司或初创公司相比，生成式人工智能（GenAI）为企业提供了新的机遇，包括：", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 110 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 110 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-135.md" }, { "id": "136", "title": "为什么说AI Agent的机会在To B？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ai-agent-tob-opportunity", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 111, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/组织转型", "originalIndustry": "企业转型", "business": "企业转型主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 111 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：Agent 诞生短短一年，已成为科技巨头与创业公司多方角逐的兵家必争之地。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 111 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 111 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-136.md" }, { "id": "137", "title": "传统企业如何落地应用AI", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/traditional-enterprise-ai-adoption", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 112, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/组织转型", "originalIndustry": "企业转型", "business": "企业转型主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 112 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：AI当前被企业高度热议，那么企业该如何选择AI产品、利用AI并产生价值，也是一件让企业CIO们头痛的工作，粉丝朋友们的需求老杨必须解决，于是在5.28日晚上特邀湖南某企业技术大咖胡总进入直播间为大家讲…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 112 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 112 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-137.md" }, { "id": "138", "title": "未来已来！蔚来、极氪、上汽、北汽等十大车企借助大模型革新汽车产业", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/auto-ten-oems-llm", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 113, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "汽车/出行", "originalIndustry": "汽车出行", "business": "汽车出行主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 113 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：大模型以数据为驱动，链接企业和车辆用户，贯穿研发生产、营销、运营、购车、用车、养车、修车、售后全场景，提供简单、高效的数字化服务。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 113 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 113 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-138.md" }, { "id": "139", "title": "万木健康程锦：医疗行业AI落地实践——利用 AIGC技术提升基层医患服务水平 | 嘉程创业流水席216席精彩回顾", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/wanmu-health-ai-practice", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 114, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 114 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：近日，嘉程创业流水席第216席【探讨生成式AI在垂直行业的最佳落地案例】，邀请了万木健康创始人&CEO程锦分享，主题是《让“轻”服务实现“高”价值——医疗AIGC变革全新格局》。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 114 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 114 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-139.md" }, { "id": "140", "title": "肖钢：推进AI大模型在金融行业应用的五项建议", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-xiaogang-suggestions", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 115, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 115 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：“智能金融”指人工智能（AI）和金融行业的深度融合。与传统的判别式AI小模型相比，生成式AI大语言模型的问世标志着AI与金融融合发展面临新的转折点。当前，大模型已被国内各类金融机构引进，作为提振内部生…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 115 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 115 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-140.md" }, { "id": "141", "title": "2024中国“大模型+智能客服”最佳实践案例TOP10重磅发布", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/smart-cs-top10-2024", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 116, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/客服", "originalIndustry": "智能客服", "business": "智能客服主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 116 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：随着人工智能技术的迅猛发展，智能客服已经成为各行各业中不可或缺的一部分。在这个数字化时代，客户期待获得即时、个性化、高效的服务体验，而大模型技术的崛起为实现这一目标提供了全新的可能性。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 116 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 116 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-141.md" }, { "id": "142", "title": "大模型和生成式AI在医疗行业的落地应用｜非凡观点", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/genai-healthcare-landing", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 117, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "医疗健康", "originalIndustry": "医疗健康", "business": "医疗健康主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 117 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：作为一名西医医生，我认为AI与医学一直紧密相连。自从AI的概念在1953年的达特茅斯会议上提出以来，AI就一直在医疗领域扮演着重要角色。在医疗影像诊断和医疗器械方面，AI应用广泛。经典AI时代过后，我…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 117 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 117 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-142.md" }, { "id": "143", "title": "AI 大模型在金融行业应用深度分析 2024", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/llm-finance-deep-analysis-2024", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 118, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 118 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：➢当前AI与金融的结合主要有两条技术路径：①通用模型+金融语料训练金融大模型，②金融垂类大模型。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 118 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 118 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-143.md" }, { "id": "144", "title": "大模型时代下，数字员工演进全景图：RPA/IPA/Agent", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/digital-employee-evolution", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 119, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "跨行业/数字员工", "originalIndustry": "数字员工", "business": "数字员工主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 119 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：数字员工三阶段：RPA/IPA/Agent", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 119 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 119 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-144.md" }, { "id": "145", "title": "大模型在金融行业的应用场景和落地路径", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/llm-finance-scenarios-path", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 120, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 120 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：这是最好的时代，也是最坏的时代。尽管大模型技术在金融领域具有巨大的应用潜力，但其应用也面临不容忽视的风险和挑战。本文将深入研究大模型在金融领域的数据隐私和安全风险、模型可解释性和透明度、监管和合", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 120 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 120 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-145.md" }, { "id": "146", "title": "AI产品分析：淘宝问问——AI加持的电商购物体验新可能", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/ecom-taobao-wenwen", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 121, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "零售/电商", "originalIndustry": "电商零售", "business": "电商零售主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 121 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：尽管近几年阿里、京东、腾讯等传统的中国互联网巨头负面缠身（总市值上的跌幅也充分体现这一点），但这并不妨碍他们在AI技术和产品应用方面的探索和实践。", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 121 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 121 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-146.md" }, { "id": "147", "title": "山证科技总经理分享：AI与大模型在金融领域的创新应用实践", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-shanzheng-practice", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 122, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 122 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：大模型自问世就得到了广泛关注，并迅速引发了业内的热潮。从行业应用的角度来看，金融行业的渗透率应该是最高的。对于券商行业而言，传统业务模式存在诸多瓶颈，尤其在提升工作效率、服务效率、服务质量及降低人力成…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 122 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 122 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-147.md" }, { "id": "148", "title": "基于LLM的多Agent框架在金融市场数据的应用", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-multi-agent-market", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 123, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 123 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：Agent 金融市场 异常检测", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 123 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 123 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-148.md" }, { "id": "149", "title": "大模型落地金融行业，如何闯关最后一公里？", "person": "原文未核验具名受访者", "book": null, "pages": null, "url": "https://fdechina.ai/cases/insights/finance-last-mile", "source": "FDE中国社区", "sourceFamily": "fdechina", "sourceOrder": 124, "sourceLabel": "FDE中国社区", "kind": "行业文章", "reviewStatus": "目录收录", "industry": "金融", "originalIndustry": "金融", "business": "金融主题阅读", "types": [], "pains": [], "evidence": "FDE中国社区案例目录，第 124 条，读取于2026-09-08（仅目录收录）", "summary": "原站简介：AI 大模型引领千行百业加速升级。在金融行业，大模型正以其出色的数据处理和分析能力引领着一场技术变革。那么，目前大模型在金融行业的应用现状如何？大模型在金融行业的落地应用面临哪些问题和挑战？如何打通大…", "insight": "从原文区分项目实践、实施建议、产品介绍与行业观点，再判断哪些内容可复用于自己的项目。", "boundary": "原站案例目录包含文章、工具解读与方案，不等同于具名客户已交付项目；标题中的效果数字只是原文表述，未在此独立验证。", "missing": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "exercise": "阅读时找出：业务使用者、最终交付物、具体验收用例与仍由人负责的环节。", "g": [], "links": [], "delivery": { "summary": "待研读 · 暂未确认交付载体", "platforms": [], "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 124 条，读取于2026-09-08（仅目录收录）", "caveat": "没有逐篇核验不等于原文没有披露；不从标题中的工具名猜填交付平台。" }, "acceptance": { "status": "待研读", "summary": "待研读 · 暂未确认验收方法", "details": "本次仅收录目录，正文的交付平台、验收方法与效果尚未逐篇核验。", "evidence": "FDE中国社区案例目录，第 124 条，读取于2026-09-08（仅目录收录）", "caveat": "这是一条目录索引，不把标题、预期效果或行业观点写成项目验收事实。" }, "nodes": [], "svg": null, "mermaid": null, "note": "notes/case-149.md" } ], "stats": { "industry": [ [ "零售/电商", 25 ], [ "金融", 16 ], [ "医疗健康", 14 ], [ "制造业", 13 ], [ "跨行业/客服", 11 ], [ "跨行业/数字员工", 10 ], [ "汽车/出行", 9 ], [ "跨行业/智能体工具", 8 ], [ "跨行业/FDE方法", 8 ], [ "软件/企业服务", 8 ], [ "跨行业/数据分析", 5 ], [ "跨行业/知识管理", 3 ], [ "跨行业/组织转型", 3 ], [ "政务/公共服务", 3 ], [ "供应链/采购", 2 ], [ "电信通信", 1 ], [ "法律服务", 1 ], [ "工程租赁", 1 ], [ "建筑设计", 1 ], [ "科研机构", 1 ], [ "跨境物流", 1 ], [ "跨行业/财务", 1 ], [ "商贸/外贸", 1 ], [ "生物科技", 1 ], [ "消防维保", 1 ], [ "营销服务", 1 ] ], "types": [ [ "流程自动化", 18 ], [ "Agent编排", 13 ], [ "规则计算/优化", 12 ], [ "文档视觉理解", 11 ], [ "内容生成", 10 ], [ "数据分析/预警", 9 ], [ "协作工作台", 9 ], [ "知识检索", 9 ], [ "组织能力建设", 6 ] ], "pains": [ [ "重复劳动", 23 ], [ "数据分散", 17 ], [ "交付慢", 14 ], [ "质量与合规风险", 14 ], [ "采纳与协同困难", 11 ], [ "判断口径不一致", 11 ], [ "知识难传承", 9 ], [ "经营不可见", 6 ], [ "供需匹配低效", 5 ] ], "source": [ [ "FDE中国社区", 124 ], [ "双来源", 23 ], [ "仅册子", 1 ], [ "仅官网", 1 ] ], "originalIndustry": [ [ "电商零售", 16 ], [ "金融", 16 ], [ "医疗健康", 14 ], [ "数字员工", 10 ], [ "智能客服", 10 ], [ "汽车出行", 9 ], [ "工业制造", 8 ], [ "智能体应用", 8 ], [ "AI SaaS", 8 ], [ "FDE 实践", 8 ], [ "零售/电商", 7 ], [ "数据分析/BI", 5 ], [ "制造业", 4 ], [ "企业转型", 3 ], [ "政务/公共服务", 3 ], [ "知识管理", 3 ], [ "互联网/软件", 2 ], [ "财务", 1 ], [ "法律服务", 1 ], [ "供应链", 1 ], [ "供应链/采购", 1 ], [ "建筑设计", 1 ], [ "客服", 1 ], [ "跨境电商", 1 ], [ "跨境物流", 1 ], [ "零售", 1 ], [ "企业服务", 1 ], [ "生物科技", 1 ], [ "消防维保", 1 ], [ "政企", 1 ], [ "制造", 1 ] ], "kind": [ [ "行业文章", 118 ], [ "案例拆解", 25 ], [ "试点方案", 5 ], [ "实施方案文章", 1 ] ] } }