Looks at demand, existing tools, complaints, and workflow gaps
Outputs: Product Brief, Value Contract, Fixture Plan
📜 Value Contract
AHADid they get the promise?
How quickly they hit the first real value moment, and how much friction showed up before it.
PAYOFFDid they finish something useful?
Whether they finished the job, and whether the output was good enough to actually use.
RETENTIONDid they create a reason to return?
Whether the session leaves behind a reason to come back: saved work, reusable output, or a next step.
Discovery drafts it. The Conveyor owns the live version and updates it round by round.
📜
handoff ↓
⚙ The Conveyor
Runs the loop, validates outcomes, and decides what to fix next.
🔨 Builder
Builds the app from reusable ValueBlocks. Handles the next fix, then runs its own QA before handing off.
ValueBlocks battle-tested reusable components
📐 Code Quality
Checks code quality, coherence, and regressions
🧑💻 Dev Persona
Checks operator and admin flows that user personas may not hit
🛡 Builder QA
Retests the changed area before the app leaves Builder
🔒 Security Audit
Looks for auth, injection, secrets, and deployment risk
🛡 Conveyor QA
independent gate
🛡
✓ regression suite ✓ key user journeys retested
✓ independent gate before personas run again
🧪 Experience Lab
Personas use the app like real users and return evidence. First-use sessions test the aha and the job-to-be-done. Later revisits test whether the app earned a return.
"aha!"
👩🔬
Priya28s
"...aha"
🧑💻
Jordan52s
"broken"
👨🎨
Alex∞
Evidence
▸ start_session
▸ friction
▸ aha ✓
▸ completed
▸ return_trigger
▸ gave_up
▸ end_session
Session 0
→
📦 Memory
→
Revisits
→
Payoff
·
Retention
Same persona, remembered context, fresh revisit runs
📊 Scoring & Validation
The conveyor turns persona evidence into three core signals, then picks the next issue to fix. Full rounds update the headline score.