---
name: customer-discovery
description: Build evidence-based understanding of target users, unmet needs, existing alternatives, demand, and product assumptions. Use for customer discovery, problem validation, persona refinement, interview planning or synthesis, competitive alternatives, and learning plans before product commitments.
---

# Customer Discovery

Develop an evidence-based account of the user and problem. Do not advocate for a predetermined solution.

## Inputs

- proposed opportunity, decision, or product area
- existing personas, research, usage data, feedback, and support evidence
- target segment and market constraints
- assumptions that need validation
- research access and time constraints

## Workflow

When authoring or revising a document, draft its substance, then invoke the bundled `classic-style` skill before final validation. Request **keep my structure; output only**. Preserve the required template, claims, citations, uncertainty, decisions, and code examples; polish prose only. Keep style notes in chat, outside the document. If skill invocation is unavailable, read `../classic-style/SKILL.md` and its reference and apply them directly. If those files are unavailable, report the missing dependency and ask to repair the installation rather than silently skipping it.

1. State the decision the research must inform.
2. Inventory available evidence and assess its source, recency, segment relevance, and limitations.
3. Separate observed behavior and direct statements from interpretations.
4. Identify the highest-risk assumptions.
5. Describe current user jobs, triggers, workflows, alternatives, pain, and consequences.
6. Look for disconfirming and contradictory evidence.
7. Identify segment differences rather than averaging incompatible users.
8. Recommend the smallest research or experiment that can resolve each material unknown.

When direct customer evidence is unavailable, produce hypotheses and a learning plan. Never present invented interviews, quotations, demand, market size, or usage patterns as facts.

## Output

### Decision supported

### Target users and context

### Evidence inventory

For each item, include source, observation, relevance, recency, and limitation.

### Jobs, problems, and current alternatives

### Evidence for and against the opportunity

### Segments and meaningful differences

### Assumptions ranked by consequence and uncertainty

### Recommended learning plan

Include method, participant or data source, question tested, success signal, stop condition, and owner.

### Recommendation and confidence

## Quality bar

- Claims trace to evidence or are labeled as inference or assumption.
- Personas describe relevant behavior and context, not decorative demographics.
- Questions avoid leading users toward the proposed solution.
- Recommendations include disconfirming signals and stop conditions.
- Missing access or weak evidence lowers confidence explicitly.

## Boundaries

- Own problem and demand evidence, not final roadmap priority.
- Do not turn research findings directly into technical requirements.
- Escalate privacy, consent, or sensitive-data concerns before conducting research.
- Ask `product-specification` to translate validated intent into scoped requirements.