1. Scope card
Buyer, task, inputs, checked surface, explicit out-of-scope items, deadline, acceptance criteria, and data-sensitivity notes.
Buyer proof layer
Proof packs for human verification work.
Every first-order QA sprint should end with a compact evidence package: scope, checklist, findings, screenshots/source links, uncertainty notes, and a clear recommendation.
2. Evidence log
Screenshots, source links, row IDs, call notes, timestamps, repro steps, and what could not be verified.
3. Issue table
Severity, location, expected vs actual, buyer impact, recommended fix, and confidence level.
4. Final recommendation
Pass, pass with caveats, do not ship, re-run required, or needs a specialist review before buyer use.
Proof-pack matrix by starter SKU
| SKU | What gets checked | Minimum proof artifacts | Decision output |
|---|---|---|---|
| AI Output Verification $99 | Claims, source fit, hallucination risk, missing caveats. | Claim table, source links, corrected/caveated notes. | Trust, revise, or reject. |
| Automation QA Sprint $199 | One app/automation/user flow on desktop/mobile or tool run. | Checklist, screenshots, repro steps, issue severity. | Ready, fix first, or retest. |
| Clay/GTM QA Sprint $199 | Sample enriched rows, source accuracy, personalization risk, duplicates, broken links. | Row-level scorecard, source URLs, risk flags, cleanup recommendations. | Safe to send, segment, or rework. |
| Real-World Check $79 | One phone/local/manual verification task. | Call/source log, timestamps, confirmed/contradicted/unreachable notes. | Confirmed, uncertain, or not confirmed. |
Acceptance rule
A good proof pack makes uncertainty visible. Workers should not invent missing facts; they should mark unknowns clearly.
Founder review while early
For first orders, GoHireHumans should review deliverables before buyer delivery where feasible.
Case-study capture
After acceptance, ask for testimonial or anonymized case-study permission. Do not imply public usage without approval.