把任务、输入、工具、结果与复核组织成一条清楚的工作路径。
写清谁希望完成什么、结果交给谁以及完成后进入哪一步工作。避免用“建设智能平台”代替实际任务。
列出正在使用的表单、工单、知识库和协作工具，记录数据的输入与输出位置。新功能应尽量在实际工作发生的位置提供帮助。
输出可以是摘要、资料包、待确认建议或转人工请求。不是每项任务都需要聊天界面，也不是每次输出都应自动写回系统。
至少讨论资料不足、目标不清、范围不符与系统失败。分别给出补充问题、修改范围、人工处理和重试的明确路径。
用真实工作问题准备常见、复杂与边界案例，说明每种任务什么结果可以接受。评估从适当的工作目标出发，不只检查文本流畅程度。
交付使用说明、例外说明、资料维护方式与后续问题记录。工作流需要能够被继续理解和改进，而不只在一次演示中成功。
Organise task, inputs, tools, outputs and review into a clear working path.
State who wants to accomplish what, who receives the result and which workflow step follows. Do not use a broad ambition for an intelligent platform instead of a concrete task.
List existing forms, tickets, knowledge libraries and collaboration tools, recording where data enters and leaves. New capability should assist where the work actually happens.
The result may be a summary, material pack, suggestion awaiting confirmation or human hand-off. Not every task needs a chat interface, and not every result should automatically be written back.
Discuss insufficient information, unclear goals, scope mismatches and system failures. Give each a clear route for clarification, scope changes, human handling or retrying.
Prepare common, difficult and boundary cases from real work, specifying an acceptable outcome for each. Evaluate against the task rather than only the fluency of generated text.
Deliver operating instructions, exception guidance, material-maintenance practices and an issue record. The workflow should remain understandable and improvable beyond one successful demonstration.