A founder believes offices lose enquiries because they lack automation. Respondents describe a different problem: nobody knows who owns the next follow-up. Those accounts support an ownership question, not automatically the founder's preferred technical solution.
For teams analysing discovery, keep an assumption ledger beside the source evidence. The distinction helps prevent a convincing narrative from becoming stronger than the study supports.
What is a founder assumption ledger?
It is a reviewable list of beliefs about the audience, problem, proposed solution and buying process, with the evidence and uncertainty attached to each. An assumption is not a failure; it is something the team has not yet established sufficiently for the decision.
Avoid a binary validated label when the available evidence only supports part of the claim. A problem can recur while the proposed solution remains untested.
Break a broad belief into separate claims
Use a fictional enquiry-management product to create the proposed worksheet below. Each row needs its own evidence rather than borrowing support from another row.
| Claim | Evidence needed |
|---|---|
| Relevant staff lose follow-ups | Concrete accounts and current process |
| The consequence matters | Observed time, cost or operational effect |
| Ownership visibility would help | A task or workflow experiment |
| AI automation is the right mechanism | Comparison with simpler alternatives |
| The organisation will buy | Appropriate commercial commitment evidence |
If interviews support the first two rows, do not report that all five are established. That leap can drive an engineering budget toward an untested mechanism.
Retain evidence that changes the idea
Record effective existing workarounds and participants who do not experience the problem. Their context can narrow the audience or reveal that the opportunity is smaller than the founder assumed.
The repeated-problem guide helps group accounts without flattening differences. The feature-request guide separates suggestions from behaviour and commitment.
Choose the next test from the riskiest uncertainty
A manual ownership queue may test the central task before an AI integration is built. If the main uncertainty is access to the current CRM, a technical feasibility review can answer that separate question.
My StackFlight case study describes collecting original feedback for founder review. It does not claim that a study platform can turn assumptions into verified demand automatically.
Keep the ledger updated when the decision changes. Mark what new evidence supports, what remains unknown and which prior assumption the team has abandoned. This makes a smaller product scope easier to justify without hiding the learning that led there.
Should we remove assumptions from the product brief?
Keep them visible and labelled. Hidden assumptions are harder to test or challenge.
Can one interview validate an entire product idea?
It can provide useful context, but broad product and commercial claims need appropriately scoped evidence.
What should we write first?
List the audience, problem, solution and buying assumptions separately before a discovery review.