{"schemaVersion":"1.0","updatedAt":"2026-07-12","service":{"problemSignals":{"zh":[{"id":"repetitive-entry","title":"重复录入与整理","description":"同一批信息需要在表格、聊天、邮件或业务系统之间反复复制。"},{"id":"system-handoffs","title":"多系统无法衔接","description":"流程依赖人工下载、转发、核对和重新上传，状态无法持续追踪。"},{"id":"scattered-information","title":"信息分散难以追踪","description":"资料、判断依据和处理结果分散在个人经验与不同文件中。"},{"id":"repeated-judgment","title":"重复判断与录入","description":"相似规则每天由人工重新判断，结果还要再写回其他系统。"},{"id":"fragile-automation","title":"自动化失败后只能重来","description":"脚本或流程一旦中断，缺少恢复点、异常记录和人工接管机制。"}],"en":[{"id":"repetitive-entry","title":"Repeated entry and cleanup","description":"The same information is repeatedly copied across spreadsheets, chat, email, or business systems."},{"id":"system-handoffs","title":"Disconnected systems","description":"The workflow relies on manual downloads, forwarding, checks, and uploads, with no durable status trail."},{"id":"scattered-information","title":"Scattered information","description":"Source material, decision rationale, and outcomes are scattered across files and individual knowledge."},{"id":"repeated-judgment","title":"Repeated judgment and recording","description":"People repeat similar judgments every day and then re-enter the result elsewhere."},{"id":"fragile-automation","title":"Fragile automation","description":"When a script or workflow stops, there is no checkpoint, exception record, or human takeover path."}]},"method":{"zh":[{"id":"diagnose","title":"诊断现状","description":"还原真实输入、处理步骤、责任人、异常与交付结果。"},{"id":"human-boundary","title":"设计人机边界","description":"明确 AI 可以处理什么、关键决策由谁审核、哪些动作必须人工确认。"},{"id":"validate","title":"验证真实流程","description":"使用真实样本验证输入、异常、恢复、审核与输出，而不是只演示理想路径。"},{"id":"deliver","title":"交付并持续优化","description":"交付可运行系统、操作边界和追踪证据，再根据真实使用调整。"}],"en":[{"id":"diagnose","title":"Diagnose the current workflow","description":"Map real inputs, processing steps, owners, exceptions, and deliverable outcomes."},{"id":"human-boundary","title":"Design the human boundary","description":"Define what AI may handle, who reviews critical decisions, and which actions require confirmation."},{"id":"validate","title":"Validate the real workflow","description":"Validate inputs, exceptions, recovery, review, and outputs with real samples rather than an idealized demo path."},{"id":"deliver","title":"Deliver and improve","description":"Deliver a working system, operating boundaries, and traceable evidence, then improve it from real use."}]},"suitableWork":{"zh":["高频、规则相对稳定、需要跨文件或跨系统流转的业务流程","需要 AI 处理非结构化信息，但关键结果仍需人工审核的流程","需要保存状态、异常、恢复记录和可审计输出的自动化改造"],"en":["High-frequency workflows with relatively stable rules that move across files or systems","Workflows where AI handles unstructured information while critical results still require human review","Automation that must preserve state, exceptions, recovery records, and auditable outputs"]},"boundaries":{"zh":["不销售固定行业模板；每个客户从实际流程诊断开始。","不承诺所有步骤无人值守；关键决策和高风险动作保留人工审核。","不使用虚构指标、客户身份或未经批准的项目结果作为销售证明。","初步提交不等于自动报价或必然安排会议。"],"en":["No fixed industry templates are sold; every engagement starts from the actual workflow.","Not every step is promised to be unattended; critical decisions and high-risk actions retain human review.","No invented metrics, customer identities, or unapproved project outcomes are used as sales proof.","An initial submission is not an automatic quote or a guaranteed meeting."]}},"contact":"/contact"}