Verification checklist
AI output verification checklist.
Five checks to run on AI-generated work before it reaches customers, spreadsheets, CRM data, or decisions. Run them yourself, or paste the list into a human review request as the acceptance criteria.
- Checks
- Five
- Use it as
- Acceptance criteria in a brief
- Deliverable
- Correction list with confidence notes
The checklist
In the order a reviewer works through it.
Compare every factual claim against source material
Names, titles, dates, numbers, links, and citations each get checked, not skimmed.
Mark unsupported claims instead of rewriting them as true
A claim without a source is labeled unsupported, even if it sounds right.
Check what the output is missing
Caveats, exceptions, and context the source includes but the output dropped.
Flag tone, policy, and brand-risk issues
Anything a customer, regulator, or partner would object to is listed with the reason.
Return a correction list with confidence notes
What to fix first, and what the reviewer could not verify.
What the reviewer returns
Ask for this shape so the result can be acted on.
- A short report with pass or fail notes per check
- Specific issues with source links or screenshots
- A prioritized correction list
- Confidence notes and anything that could not be checked
When to hand it to a person
Each check maps to a fixed review service.
Scope note. Human review reduces risk and creates a review record, but it does not promise perfect accuracy, legal compliance, or a business outcome, and it does not replace professional legal, medical, financial, or compliance advice. GoHireHumans is a listing and payment connector; the reviewer is an independent worker.
Turn the checklist into a review request.
Paste it into a brief as the acceptance criteria and email the AI output with its sources.