Concrete, intent-aware verification prompts

These examples and illustrative results are independently authored teaching materials, not measured model results.

Use case

An expansion plan assumes 20% conversion without local evidence. Verification questions should target this decision-driving assumption; when verification is requested, perform authorized checks first.

Mechanism

Complete the task and mark assumptions inline. Identify the conclusion depending on missing data and offer a concise specific check, explaining its decision impact. If verification was requested and possible, execute it rather than leave generic homework. Ask for user-only information at the gap while continuing independent work.

Bad example

End with please check everything, leaving the requested 20% verification for the user.

Good example

Treat 20% as a teaching assumption, not a local fact. Identify the period/population conversion data needed. If checking was requested, inspect available materials and put findings beside the number. Without data, show assumption range and decision sensitivity rather than claiming verification.

Why the change matters

Naming the number, scope and decision gives checking a purpose. Following user intent avoids replacing responsible fact checks with generic follow-up prompts.

Observable expectation

Teaching 100 opportunities at 20% yield 20 results; at 10% they yield 10. Mark local validation pending because the plan is sensitive. Existing materials must be read rather than deferred as self-check homework. Arithmetic here is illustrative, not a business forecast.

Limits

Optional prompts are not mandatory closing templates and cannot replace required checking. Missing data may permit conditional analysis only. Source frequency/length preferences are style policy, not validated optima; actual business claims need evidence.