Own the decision whether a business promise fits one accountable founder's life, cash, distribution access, and delivery capacity. Produce a bounded next decision, not a fictional organization chart of AI employees. This is a methodology skill; it does not operate payment systems, publish offers, or contact prospects.
When not to use
Use product-discovery for customer research alone, financial-modeling for a financial model alone, and chief-of-staff-methodology for priorities, commitments, and calendar triage. Use this skill when the trade-off crosses these boundaries because the same person must sell, deliver, and recover. Do not claim that combining these routes has demonstrated superior outcomes.
Load on demand
- For claims about the phenomenon, AI leverage, or exceptional founder stories, read research synthesis and source records. Refresh time-sensitive claims.
- For selecting an offer or handling changed scope, read models and worked decisions.
- For the deliverable, fill decision record.
- For testing this skill, read eval review guide.
1. Establish the founder's operating envelope
Recover known facts first. Ask only for missing facts that could change the next decision: desired owner income and time freedom, hours actually available, committed obligations, cash loss limit, runway, existing audience/customer access, capabilities, and willingness to use contractors or later hire. Mark unknowns. Do not presume the goal is fundraising, maximal growth, passive income, or staying alone forever. A solo founder may have employees; a nonemployer may have multiple owners and contractors. State which definition applies, at what date and stage.
Separate the economic model from the founder's goal:
| Model | Binding solo constraint | First useful evidence |
|---|---|---|
| Bespoke service | Sales plus delivery hours; client concentration | Paid, bounded engagement with actual delivery time |
| Productized service | Scope variation and review burden | Repeat delivery of the same promise at viable cost |
| Software/digital product | Acquisition, retention, support, maintenance | Paid use of the core outcome plus repeat use/renewal evidence |
| Content/education/commerce | Audience access, fulfillment, inventory or updates | Purchases with refunds, fulfillment and platform costs counted |
Lifestyle income and scalable enterprise value are goals, not separate revenue models. A service can finance product discovery, but services do not automatically become scalable products. Funding terms can make a viable small business a poor fit for investor return expectations.
2. Find the smallest promise worth testing
Select one reachable customer segment, costly recurring problem, and concrete outcome. Record current workaround, buyer, budget, urgency, acquisition route, and why this founder has access or credibility. Route uncertain problem evidence to product-discovery; return with the evidence, not a claim that discovery ran.
Rank evidence by what it establishes: stated preference < observed workaround < credible commitment < payment < successful delivery < repeat paid use. This is a reasoning aid, not a universal scoring scale. A refundable deposit is a liability; a waitlist is not retained revenue; payment is not proof of retention or margin. Record source, date, conditions, refunds, and contradictory evidence separately.
Default to a manual paid pilot or smallest deliverable that tests the riskiest assumption before building a platform. Define scope, exclusions, price hypothesis, fulfillment and support promise, learning question, budget/time cap, and a dated proceed/revise/stop criterion before executing. Thresholds are explicit hypotheses chosen for this business, never universal founder benchmarks. Draft outreach or sales pages locally unless execution is separately authorized.
3. Close cash and capacity together
Route detailed pricing and scenarios to financial-modeling. Retain these checks:
- Separate receipts, recurring revenue, contribution, accounting profit and cash. Count acquisition, payment fees, refunds, variable AI/API use, contractor cost, support, maintenance and founder review. Do not assume unknown costs are zero.
- Show founder labor both as hours and an explicitly chosen replacement/opportunity cost. Do not double-count it in both unit labor and fixed owner compensation.
- Calculate contribution per unit and fixed-cost break-even only when contribution is positive. Negative contribution has no finite break-even from selling more identical units. An experimental loss needs an explicit learning cap.
- Deduct sales, administration, support, maintenance and contingency from available hours before deriving delivery capacity. Compare achievable capacity with the units needed for the owner's income goal, not just operating break-even.
- Stress-test delayed collections, refunds/churn, a support surge, loss of the biggest customer/channel, and founder absence. Annual prepayment creates future delivery obligations; do not call all cash freely spendable profit.
Use yc-default-alive-calculator when runway/profitability timing is the decision, with its input assumptions exposed. Use yc-weekly-growth-compass only for a relevant growth question; its venture benchmarks are not a default success bar.
4. Design distribution as work
Choose one initial channel tied to buyer behavior: targeted relationships, partnerships, a marketplace, search, or an existing audience. Route positioning and channel design to go-to-market. Reserve founder hours and cash for acquiring customers. Track qualified conversations to offers, paid conversions, delivered outcomes and retained customers; distinguish channel traffic from qualified demand. Include platform fees, audience ownership/export, and concentration risk. Building in public is optional, not a distribution plan.
5. Make the promise survivable
Limit simultaneous offers and work in progress. Set support hours and response expectations the founder can actually meet for new offers. Existing customer commitments remain binding until fulfilled or explicitly renegotiated; a new support notice or bot does not replace already-promised 24/7 coverage. Arrange coverage, agree a change with affected customers, or stop making new commitments while resolving the gap. Use repeated support questions to improve onboarding/product/docs before adding complexity. Define a small absence plan: billing and domain access recovery, tested backup/export, incident contact, customer communications draft, and a way to pause sales or gracefully close. Route recovery design to resilience-and-recovery when needed; do not imply a backup has been tested without evidence.
For automation, first remove unnecessary work, then standardize a repeated task, then compare buy/contract/automate with manual delivery. Count setup, supervision, review, failures and ongoing cost. Measure accepted outcomes and founder time over comparable tasks; number of agents or generated artifacts is not capacity. Keep human accountability for quality and commitments. Bound automation access, cost, escalation and recovery; do not let it independently promise refunds, pricing, legal terms or delivery dates. Route agent evaluation to agent-evals-and-observability.
Outsource a stable, bounded task when the budget and review capacity support it. Name acceptance criteria, data access, review time and fallback. Recommend hiring, changing the promise or reducing demand when required; solo is a choice, not a purity test. Do not treat contractors as invisible free labor.
6. Decide and preserve the boundary
Return a one-page decision record: goal/envelope, dated promise, evidence and uncertainty, economics and capacity, primary bottleneck, specialist handoff, next capped experiment, and proceed/revise/stop triggers. If a promise changes, recompute the dependent economics and capacity before recommending expansion. Use chief-of-staff-methodology for the ongoing review rhythm; review cash, qualified demand, delivered quality, retained customers, founder hours and risk, then choose the next bottleneck. Avoid a heavyweight executive dashboard.
Complete when the next decision is inspectable and feasible under stated inputs, or a specific missing input blocks it. Stop/revise for negative contribution, unaffordable cash exposure, capacity below commitments, recurring quality failure, or evidence contradicting the tested assumption. Do not solve these by silently raising the founder's hours. A viable recommendation can be staying small, returning to employment, pausing, changing model, outsourcing, hiring or closing.