Practical review worksheet

How to check an AI-generated business summary

A practical, synthetic exercise for checking sources, calculations and unsupported claims before using an AI-generated business summary.

Illustrative exercise using invented data. This is not a customer result, a product benchmark or a promise of productivity gains.

Source material

In a fictional month, a team submitted 10 proposals and won 3. Two proposals are still awaiting a decision. No figures are available for the previous month.

Draft to review

The team achieved a 40% win rate, improving on last month. The two remaining proposals will probably be won.

Check each claim before using the summary

CheckWhat to verifyResult in this example
NumbersRecalculate using the stated denominator.3 wins / 10 submitted proposals = 30%, not 40%. If using only decided proposals, state that denominator explicitly.
SourcesFind support for every comparison and assertion.No previous-month figures support the claim of improvement.
UncertaintySeparate facts, assumptions and predictions.The two pending decisions are unknown outcomes, not likely wins established by the source.
Human reviewConfirm definitions and intended audience before sharing.A responsible reviewer checks how pending proposals are treated and approves the final wording.

A more defensible summary

The team won 3 of the 10 proposals submitted in this fictional month (30% of submissions). Two decisions are still pending. The source provides no previous-month comparison.

Use this checklist on your own work

  1. Record the source, version and period.
  2. List each material claim and its supporting passage or calculation.
  3. Mark unsupported claims, missing inputs and uncertainty.
  4. Correct the draft and record who approves it before use.

Use synthetic or approved information for practice. Do not paste confidential, personal or restricted business data into an unapproved tool.

Choose a practical workshop →