Customer Discovery Prover Connector for AI agents.

It forces your AI client to cite specific evidence from your interviews. If the agent can't provide a quote or a specific role, the capability flags the discovery as 'Assumed' or 'Invented.'

The Mom Test is a methodology for getting honest feedback by asking about past behavior instead of future promises. This Connector ensures your AI doesn't ask leading questions like 'Would you pay for this?'

Yes. It specifically checks if your Ideal Customer Profile is grounded in real data or if it's just a list of demographics like age and interests.

It helps you prepare for funding by ensuring your discovery survives investor scrutiny. It moves you from 'we think people want this' to 'we have evidence that this specific segment will pay.'

It works for any business model. It's particularly effective for B2B because it helps separate complex buyer segments that often get conflated as 'enterprise' or 'SMB.'

The capability provides a verdict matrix. It tells you exactly which pivot failed, whether it was biased validation, conflated segments, or lack of commitment signals, so you know exactly what to fix.

No. It validates that your discovery process is grounded in real data. interview evidence, unbiased methodology, separated segments, and commitment-based WTP signals. It does not replace conversations with customers. It forces you to prove you had them.

A framework by Rob Fitzpatrick for conducting customer interviews that produce truthful data. The core rule: never ask leading questions about the future ('Would you use this?'). Instead, ask about past behavior ('When did you last encounter X? What did you do?'). People lie about future behavior. past behavior is a reliable signal.

Yes. it is designed for pre-product discovery. WTP signals for pre-product include: signed LOIs, paid design partnerships, deposits against future delivery, time commitments (agreed to a weekly feedback session), and reputation commitments (willing to be named as a design partner). You do not need a product to test willingness-to-pay.