{ "updated_at": "2026-05-19", "canonical_list_url": "https://www.guokezhifei.com/cases", "cases": [ { "id": "silhouette-dashboard", "url": "https://www.guokezhifei.com/cases/silhouette-dashboard", "sort_order": 1, "featured": true, "industry_zh": "新零售 · 原创服饰", "industry_en": "Retail · Original apparel", "industry_slug": "enterprise-services", "product_zh": "BI 与指挥中枢", "product_en": "BI & Command Center", "product_slug": "bi-platform", "title_zh": "可视化 AI 数据看板", "title_en": "Visual AI Analytics Dashboard", "summary_zh": "为一家深耕原创设计的新零售服饰品牌，打通天猫、抖音、小红书等多渠道经营数据，让 GMV、转化与库存风险一屏尽览。", "summary_en": "For an original-design apparel brand, we unified Tmall, Douyin, and Xiaohongshu into one live command center for GMV, conversion, and inventory risk.", "highlights_zh": [ "多平台数据整合", "AIPL 漏斗分析", "AI 异常预警" ], "highlights_en": [ "Multi-platform integration", "AIPL funnel analytics", "AI anomaly alerts" ], "capabilities_zh": [ "全栈数据集成", "AI 异常洞察", "业务 KPI 共创", "可视化指挥中枢" ], "capabilities_en": [ "Full-stack data integration", "AI anomaly insights", "Co-designed KPIs", "Visual command center" ], "image": "https://www.guokezhifei.com/assets/case-studies/silhouette-dashboard.png", "outcomes_zh": [ "经营复盘从「T+1 表格」缩短至分钟级刷新", "大促异常平均提前 2–4 小时预警，支撑备货与投放调整", "管理层一屏掌握全渠道健康度，减少跨部门对齐成本" ], "outcomes_en": [ "Ops reviews moved from next-day spreadsheets to minute-level refresh", "Peak anomalies flagged 2–4 hours earlier, enabling stock and ad adjustments", "Leadership gained one-screen cross-channel health, cutting alignment overhead" ] }, { "id": "silhouette-ai-design", "url": "https://www.guokezhifei.com/cases/silhouette-ai-design", "sort_order": 2, "featured": true, "industry_zh": "新零售 · 原创服饰", "industry_en": "Retail · Original apparel", "industry_slug": "enterprise-services", "product_zh": "行业大模型平台", "product_en": "Industry LLM Platform", "product_slug": "industry-llm", "title_zh": "原创服装 AI 设计赋能", "title_en": "AI-Assisted Original Fashion Design", "summary_zh": "将品牌二十年设计档案与东方美学语料沉淀为服装行业大模型，设计师用自然语言即可生成符合调性的新款草图与面料方案。", "summary_en": "We turned two decades of design archives and Eastern aesthetics into a fashion LLM—designers sketch brand-aligned concepts and fabric directions in natural language.", "highlights_zh": [ "服装大模型", "传统色与面料库", "文生图设计" ], "highlights_en": [ "Fashion LLM", "Traditional palette & fabrics", "Text-to-image design" ], "capabilities_zh": [ "行业大模型训练", "知识图谱构建", "文生图工作流", "设计业务落地" ], "capabilities_en": [ "Domain LLM fine-tuning", "Knowledge graph design", "Text-to-image workflows", "Design ops delivery" ], "image": "https://www.guokezhifei.com/assets/case-studies/silhouette-ai-design.png", "outcomes_zh": [ "概念稿产出效率显著提升，缩短前期探索时间", "新款方案更一致地延续品牌东方美学基因", "设计团队可将精力集中在版型与工艺创新上" ], "outcomes_en": [ "Concept throughput rose sharply, compressing early exploration", "New lines more consistently expressed the brand's Eastern aesthetic DNA", "Designers redirected effort to pattern-making and craft innovation" ] }, { "id": "wulairong-matching-platform", "url": "https://www.guokezhifei.com/cases/wulairong-matching-platform", "sort_order": 4, "featured": false, "industry_zh": "金融科技 · 助贷服务", "industry_en": "Fintech · Lending services", "industry_slug": "financial-risk-control", "product_zh": "行业自动化引擎", "product_en": "Automation Engine", "product_slug": "ai-rules-engine", "title_zh": "智能匹配数据处理中台", "title_en": "Intelligent Matching Data Platform", "summary_zh": "构建助贷业务「智能匹配中台」，统一字段资产、产品档案与用户画像，用可配置规则引擎驱动千人千面的贷款产品匹配与失败归因分析。", "summary_en": "We delivered a lending matching hub—unifying field assets, product profiles, and user personas behind a configurable rules engine for personalized matching and failure analytics.", "highlights_zh": [ "智能匹配引擎", "规则库", "用户画像" ], "highlights_en": [ "Matching engine", "Rule library", "User personas" ], "capabilities_zh": [ "规则引擎", "数据资产治理", "用户画像", "业务自动化" ], "capabilities_en": [ "Rules engine", "Data asset governance", "User personas", "Process automation" ], "image": "https://www.guokezhifei.com/assets/case-studies/wulairong-matching-platform.png", "outcomes_zh": [ "匹配策略迭代从「周级开发」变为「小时级配置」", "失败热点可视化，产品与渠道团队对齐更精准", "整体匹配成功率与可解释性同步提升" ], "outcomes_en": [ "Strategy changes moved from weekly dev cycles to hourly config", "Failure hotspots became visible, aligning product and channel teams", "Match success and explainability improved together" ] }, { "id": "emerson-material-analysis", "url": "https://www.guokezhifei.com/cases/emerson-material-analysis", "sort_order": 5, "featured": true, "industry_zh": "先进制造 · 工业设备", "industry_en": "Advanced manufacturing · Industrial equipment", "industry_slug": "enterprise-services", "product_zh": "Document AI", "product_en": "Document AI", "product_slug": "document-ai", "title_zh": "供应商材质智能分析平台", "title_en": "Supplier Material Intelligence Platform", "summary_zh": "为全球工业制造龙头搭建供应商材质报告智能分析平台，OCR 与大模型多路共识校验化学成分，OneDrive 自动入库，人机协同保障检测一次合格。", "summary_en": "For a global industrial manufacturer, we built supplier material report intelligence—OCR plus multi-LLM consensus on chemical composition, OneDrive ingestion, and human-in-the-loop quality gates.", "highlights_zh": [ "OCR + LLM 共识", "材质报告数字化", "100% 一次合格率" ], "highlights_en": [ "OCR + LLM consensus", "Digital material logs", "100% first-pass yield" ], "capabilities_zh": [ "Document AI", "多模型共识", "企业系统集成", "人机协同质检" ], "capabilities_en": [ "Document AI", "Multi-model consensus", "Enterprise integration", "Human-in-the-loop QC" ], "image": "https://www.guokezhifei.com/assets/case-studies/emerson-material-analysis.png", "outcomes_zh": [ "材质报告数字化入库全自动，人工录入工作量显著下降", "元素检测一次合格率达到 100% 目标（人机协同门控）", "采购与质检可实时检索历史报告，审计链路完整可追溯" ], "outcomes_en": [ "Fully automated digital ingestion; manual entry workload dropped sharply", "100% first-pass yield on elemental checks via human-in-the-loop gates", "Procurement and QC gained searchable history with full audit trails" ] }, { "id": "ocr-center-bank-credit-report", "url": "https://www.guokezhifei.com/cases/ocr-center-bank-credit-report", "sort_order": 6, "featured": true, "industry_zh": "金融科技 · 银行信贷", "industry_en": "Fintech · Bank credit", "industry_slug": "financial-risk-control", "product_zh": "Document AI · OCR Center", "product_en": "Document AI · OCR Center", "product_slug": "document-ai", "title_zh": "OCR Center 金融文档智能解析平台", "title_en": "OCR Center: Document Intelligence Platform for Finance", "summary_zh": "OCR Center PRO 是面向银行与消费金融的通用文档解析平台——征信报告、流水、证照等复杂扫描件均可通过可配置流水线输出 Markdown/JSON，对接授信、风控与贷后系统。", "summary_en": "OCR Center PRO is a general document intelligence platform for banks and consumer finance—credit reports, statements, and certificates run through configurable pipelines into Markdown/JSON for underwriting, risk, and post-loan systems.", "highlights_zh": [ "通用解析平台", "可配置流水线", "输入结果对照" ], "highlights_en": [ "General-purpose platform", "Configurable pipeline", "Input vs. output review" ], "capabilities_zh": [ "通用文档解析平台", "可配置流水线", "Markdown/JSON 输出", "金融场景落地" ], "capabilities_en": [ "General document parsing platform", "Configurable pipelines", "Markdown/JSON output", "Finance deployment" ], "image": "https://www.guokezhifei.com/assets/case-studies/ocr-center-bank-credit-report.png", "outcomes_zh": [ "形成可复用的文档解析平台，新增文档类型以配置扩展为主，显著降低重复开发", "信贷与运营团队从手工抄录转为「提交任务—复核结构化结果」，批量接入开户、授信与贷后环节", "解析全过程可观测、可追溯，为金融机构提供统一的文档智能底座，而非单点项目交付" ], "outcomes_en": [ "A reusable parsing platform—new document types extend by configuration, cutting repeated engineering", "Credit and ops teams move from manual entry to submit-and-review structured output across onboarding, underwriting, and post-loan", "End-to-end observable, auditable parsing—a unified document intelligence foundation, not a one-off project" ] } ] }