AI INCOME EXPERIMENT WORKSHEET — THE WEALTH EXPERIMENT
Put AI to Work. From ideas to results.
Educational template. Blank fields are unknown, not zero. No earnings guaranteed.

1. Buyer and triggering problem:
2. Evidence (source, date, observation; separate assumptions):
3. Current workaround and buyer's spending authority:
4. Offer, price, deliverable, exclusions and acceptance check:
5. Riskiest assumption:
6. Hypothesis: We believe ___ will ___ because ___ by ___.
7. Test method and intended audience:
8. Start/end dates, cash ceiling and total hour ceiling:
9. Primary metric, denominator and preselected decision threshold:
10. Continue if / change if / stop if:
11. Sources and permissions required:
12. Agent task: owner / inputs / output file / acceptance check / next owner:
13. Human approvals before sending, spending, publishing or live changes:

RESULTS (include failed attempts and setup)
Invitations: ___ Responses: ___ Qualified conversations: ___
Eligible visits: ___ Purchases: ___ Accepted deliveries: ___ Refunds: ___
Collected revenue: ___ Direct costs including tools/fees: ___
Allocated overhead: ___ Total hours: ___ Value assigned per hour: ___
Contribution before overhead, tax and owner labor = revenue - refunds - direct costs.
Labor-adjusted result before tax = contribution - overhead - hours × assigned hourly value.
Do not subtract refunds twice if revenue is already net of refunds.
Cash contribution per hour is unavailable if time has not been recorded.
Repeat use/renewal observed: ___ Corrections/support: ___
Observed facts: ___ Assumptions: ___ Missing information: ___
Decision and reasons: ___ Next smallest test: ___ Review date: ___

RESEARCH PROMPT
Use only my sources and anonymized notes. List recurring problems, evidence,
current workarounds, buyers and unknowns. Do not invent quotes, volumes or demand.
Suggest one question that could disprove each idea.

TEAM PROMPT
Break this approved experiment into bounded research, product, design,
development, marketing and QA tasks. Specify inputs, output files, acceptance
checks and dependencies. Stop at missing evidence or budget limits. Ask before
contacting people, buying, publishing, deleting or changing live systems.

REVIEW PROMPT
Calculate rates with denominators from these raw counts. Check cost boundaries,
missing data, confounders and uncertainty. Do not infer significance from tiny
samples. Recommend continue, change, stop or more evidence with reasons.

Start: https://wealthexperiment.live/tools/first-experiment/
Guide: https://wealthexperiment.live/blog/ai-income-experiments-evidence-before-hype/