Validate
AI-assistedAI synthesizes interviews; the founder judges demand.
Moves from interviews to one-job software, tests, instrumentation, and a human-controlled launch packet.
User-selected scenario, not an expected outcome.
Current level 3/5. Potential level 4/5 in 3–6 months.
AI synthesizes interviews; the founder judges demand.
AI can implement scoped code under review.
AI drafts outreach and launch copy; humans approve all sends.
The approved application performs the narrow job with monitoring.
The founder owns support, billing activation, incidents, and roadmap.
Approval applies to the exact action once. It is not blanket permission.
Approve every interview or launch message.
Approve domains, hosting, APIs, or other spend.
Approve target, configuration, and rollback before public deployment.
Approve landing page and claims before publication.
Only the owner may activate billing or change prices.
Approve data category, retention, and access before any real customer data.
/goal Goal: validate and build the smallest maintainable software product that completes one narrow job. Begin with a named buyer, current workaround, interview consent, evidence threshold, maximum scope, data classification, and $500 real-spend ceiling. Draft interview outreach but wait for approval; humans send messages and conduct at least ten qualified interviews. Stop if the pain is not repeated, urgent, and costly enough to test. Write a one-job specification with inputs, successful output, exclusions, failure states, manual fallback, and completion metric. Build locally or in a sandbox using the existing stack. Include tests, accessibility where relevant, error handling, secrets hygiene, threat notes, logging, and rollback. Use fictional data until the owner explicitly authorizes real customer data. Prepare a landing page, onboarding, support playbook, and launch checklist without fake customers or revenue. Wait for approval before purchases, deployment, publication, payment activation, or customer-data use. MILESTONES 1. Validate pain: Ten interviews are recorded and the declared threshold is met. 2. Freeze one-job scope: Inputs, output, exclusions, fallback, and success metric are approved. 3. Build and verify MVP: Tests, security notes, failure states, and rollback pass human review. 4. Prepare launch: Landing, onboarding, support, and approval checklist are complete. COMPLETION CRITERIA 1. Evidence threshold met 2. One-job MVP passes tests 3. Security and rollback reviewed 4. No external launch action taken without approval BUDGET CEILING $500 USD EXTERNAL-ACTION GATES 1. Outreach — Explicit approval: Approve every interview or launch message. 2. Purchase — Explicit approval: Approve domains, hosting, APIs, or other spend. 3. Deploy — Explicit approval: Approve target, configuration, and rollback before public deployment. 4. Publish — Explicit approval: Approve landing page and claims before publication. 5. Activate live payments — Explicit approval: Only the owner may activate billing or change prices. 6. Use client data — Explicit approval: Approve data category, retention, and access before any real customer data. STOP CONDITIONS 1. Pain threshold fails 2. Scope exceeds one job 3. Security/privacy unresolved 4. $500 ceiling reached
Projected scenarios never appear here as observed outcomes.
Goal has not completed a recorded validation and MVP run.
No model test date recorded
Catalog baseline; exact version must be recorded at first test
A technical founder testing one narrow software job before a broad build. · Validate a repeated pain and produce the smallest testable MVP with explicit launch gates.