The useful part is the challenge
A synthetic check can suggest ambiguity, missing context or a counterargument. Those signals can make a draft or a research plan stronger.
The dangerous part is the certainty
Generated people did not see the campaign, buy the product or consent to a quote. A polished scorecard does not change that boundary.
- Label every simulated response.
- Avoid names and documentary framing.
- Do not turn a model score into a market prediction.
- Escalate important unknowns to real evidence.
A good rehearsal improves the real conversation
If the exercise leaves the team with a sharper interview question or a clearer testable assumption, it has done useful work without pretending to be something else.