Agent Skills: Discover Assumptions and Experiments

Discover Assumptions and Experiments

Overview

Surface the riskiest assumptions in an Opportunity Solution Tree and design the smallest tests that can prove or disprove them quickly.

Position in Workflow

Step 4 of product strategy workflow:

1. /discover-outcomes - Define outcomes
2. /discover-opportunities - Identify opportunities
3. /ideate-solutions - Explore solution concepts
4. /discover-assumptions - Validate with experiments (THIS)

Inputs (ask if missing, max 5)

• Target node(s): outcome, opportunity, or solution
• Target users/market
• Existing evidence (data, research, learnings)
• Constraints (time, budget, ethics, legal)
• Decision deadline

Assumption Types

• Desirability: Users want or value it
• Usability: Users can use it successfully
• Feasibility: We can build/deliver it
• Viability: It supports the business model
• Risk/Compliance/Ethics: It is safe and allowed
• Strategic: It aligns with goals and positioning

Workflow

1. List assumptions per node
   • Write assumptions as testable statements.
2. Score risk
   • Impact (low/medium/high) x uncertainty (low/medium/high).
   • Note current evidence strength (none/weak/moderate/strong).
3. Prioritize
   • Select top 3-5 riskiest assumptions.
4. Design experiments
   • Propose 2+ tests per assumption, fastest/cheapest first.
   • Define hypothesis, method, sample, success metric, and decision threshold.
5. Sequence tests
   • Start with tests that can invalidate assumptions quickly.

Experiment Patterns (examples)

• Customer interviews, observation, diary studies
• Survey with behavioral intent + follow-up validation
• Smoke test or landing page
• Fake-door or click-through test
• Concierge or Wizard-of-Oz pilot
• Prototype usability test
• A/B test or pricing experiment
• Limited rollout with manual operations

Output Format

## Assumption Discovery

### Context Summary
[1-3 sentences]

### Assumptions (by node)
- Node: [Outcome/Opportunity/Solution]
  - Assumption: ... (type: desirability)
  - Evidence: ... (strength: weak)
  - Risk: impact high x uncertainty high

### Top Risks
1) Assumption: ...
   - Why risky: ...

### Experiments
1) Assumption: ...
   - Hypothesis: ...
   - Method: ...
   - Sample: ...
   - Success metric: ...
   - Decision threshold: ...
   - Time/cost: ...

### Sequenced Plan
1) ...
2) ...

### Open Questions
- ...

### Next Step
If assumptions are validated, proceed to product planning.

Quick Reference

• Turn beliefs into testable statements.
• Prefer tests that can disprove the assumption fast.
• Document decision thresholds before running tests.

Common Mistakes

• Treating opinions as evidence
• Testing solutions before validating the underlying assumption
• Running expensive tests without cheap falsification attempts
• Vague hypotheses or missing thresholds