What we will prove with you. On what timeline. With what counts as failure.

The first ninety days

Six milestones, in order.

• Week 0: scope selection.

  Together we pick one product bet that matters to your business, not an infrastructure science project. The goal is a build that would meaningfully change your growth math if it landed, and that you can recognize "good" on without a technical reviewer.

• Week 1: intent brief, non-goals, risk list, first build plan.

  You read the brief and correct what is wrong. Nothing is built yet. If the brief misreads your intent, you find that out before code costs anything.

• Day 14: first usable workflow or prototype slice.

  Not a demo. Something you can put in front of yourself or a friendly user and form a real opinion about. Decision rationale recorded in the ledger.

• Day 30: first failure admission.

  By this point Mandaire has been corrected enough times that the corrections show up in the decision ledger and at least one explicit "I shipped this wrong and here is what I changed" admission is on file. If there is no failure admission by Day 30, that is itself a signal.

• Day 60: corrections changing autonomous decisions.

  At least one correction from Day 1-30 visibly changed a Day-31-60 decision Mandaire would have otherwise made the old way. This is the compounding claim under live test. We point to the specific instance in the ledger.

• Day 90: taste memory, and the predictability test.

  Mandaire produces your Day-90 taste memory. You read it. You decide whether the bullets describe decisions you would actually have made yourself, or whether they describe a stylized version of you that flatters but does not predict. The bet pays off if the answer is the former.

What counts as failure

Five failure modes. If any fire by Day 30, we talk before Day 60.

• Mandaire asks too many low-value questions and the escalation tax is annoying.
• Important decisions slip through to autonomous execution without your approval.
• The artifact trail is impressive but not actually useful when you go to act on it.
• The product does not move closer to a real growth bottleneck. You have shipped well but not toward anything that matters.
• You still need a technical co-founder or contractor in every loop, contradicting the no-engineer claim.

If Mandaire is failing the bet for your specific case, you should hear that from us, and you should be able to walk away with the artifact trail and the code intact. That is what trust-and-openness over lock-in means in practice.