Intake Should Reduce Ambiguity
Client intake is not a form for collecting every possible detail. It is a filter for deciding whether the problem is real, narrow, and worth scoping.
Use AI after the client submits answers. The model can summarize pain points, identify missing details, and suggest follow-up questions. Keep the final scope decision manual.
Minimum Intake Fields
Collect only what changes the quote:
- What task repeats every week?
- Who does it now?
- How long does it take?
- What inputs are used?
- What output is expected?
- What happens when the task is late or wrong?
- Which tools are already in the workflow?
Draft The Scope
After intake, ask Codex or ChatGPT to produce a one-page scope draft with:
- Problem statement.
- Inputs and outputs.
- Out-of-scope items.
- Security and credential boundaries.
- Acceptance criteria.
- Fixed-price setup option.
The human review step matters because clients often describe symptoms instead of root causes.
Good Automation Candidates
Strong candidates have stable inputs, frequent repetition, measurable time savings, and a clear owner. Weak candidates depend on vague judgment, private systems you cannot access, or platform actions that violate terms.
Operating Rule
Never ask a client to share raw passwords in a form or spreadsheet. Use proper account permissions, documented exports, or a live screen-share handoff.
Related Operator Stack Pages
- Turn intake answers into a bounded offer with the AI automation service scope template.
- Review app and file permissions with the AI automation access review checklist before giving a workflow recurring access.
- Validate whether the request should become a product with the automation-before-SaaS guide.
- Close the project with the AI automation client handoff checklist.