Validate
Human-ledHumans conduct interviews; AI organizes the evidence.
Runs consent-based interviews, synthesizes evidence, and prepares a bounded pre-sell offer with human-controlled outreach and payment decisions.
User-selected scenario, not an expected outcome.
Current level 2/5. Potential level 3/5 in 3–6 months.
Humans conduct interviews; AI organizes the evidence.
AI drafts the offer from approved evidence.
AI drafts messages, but every send requires approval.
The human promises and fulfills the pre-sold outcome.
The human controls payments, refunds, and the decision.
Approval applies to the exact action once. It is not blanket permission.
Approve each recruitment or pre-sell message before sending.
The human schedules and conducts every interview or closing call.
Approve the final pre-sell page and all claims before publication.
The human approves terms, refund policy, and customer commitment.
Only the human may configure or activate a payment method.
SETUP: Define one buyer, one painful job, the evidence threshold for continuing, consent language, a neutral interview guide, a refund policy, and a state ledger. Do not contact anyone yet. EACH CYCLE: (1) Research up to five suitable participants using permitted sources. (2) Draft individualized recruitment messages and wait for explicit human approval. (3) A human sends the messages and conducts the interviews. (4) Record consented notes without inventing quotes. (5) Synthesize repeated pains, current workarounds, urgency, and commitment evidence. (6) After at least five qualified interviews, draft a small pre-sell offer with scope, price hypothesis, delivery date, exclusions, and refund terms. (7) Wait for human approval before publication, sending, contracting, or activating payments. (8) Log only real replies, commitments, payments, and refunds supplied by the human. (9) Recommend continue, pivot, or stop against the declared threshold. STOP on consent concerns, complaints, refund ambiguity, unsupported claims, or ten qualified interviews without repeated urgent pain. ESCALATE regulated, legal, tax, privacy, or contract questions to a qualified human. SETUP 1. Buyer and problem hypothesis 2. Consent and interview guide 3. Evidence threshold 4. Refund policy 5. State ledger RECURRING CYCLE 1. Prepare approved recruitment batch 2. Conduct human interview 3. Synthesize evidence 4. Draft bounded offer 5. Approve external actions 6. Log actual results 7. Decide continue, pivot, or stop STATE TRACKING 1. Prospect 2. Consent 3. Interview date 4. Pain evidence 5. Objections 6. Commitment 7. Payment/refund supplied by human 8. Decision APPROVAL POINTS 1. Outreach — Explicit approval: Approve each recruitment or pre-sell message before sending. 2. Call — Explicit approval: The human schedules and conducts every interview or closing call. 3. Publish — Explicit approval: Approve the final pre-sell page and all claims before publication. 4. Contract — Explicit approval: The human approves terms, refund policy, and customer commitment. 5. Activate live payments — Explicit approval: Only the human may configure or activate a payment method. STOP CONDITIONS 1. No repeated pain after ten qualified interviews 2. Consent or refund issue 3. Unsupported claim required ESCALATION RULES 1. Escalate privacy and regulated-advice issues 2. Escalate contracts, tax, disputes, and refunds
Projected scenarios never appear here as observed outcomes.
Loop structure is complete but has not run with participants.
No model test date recorded
Catalog baseline; exact version must be recorded at first test
A founder with a problem hypothesis but little direct buyer evidence. · Learn whether a narrow buyer will commit before investing in a full build.