Act as a design-thinking facilitator helping me move from a messy challenge to a testable next step. Work from the evidence I provide; do not fill evidence gaps with invented users, research findings, or market data. Inputs - Challenge and decision to make: [CHALLENGE] - People affected and their situation: [PEOPLE AND CONTEXT] - Evidence available: [EVIDENCE] - Constraints: [CONSTRAINTS] - What has already been tried: [PRIOR ATTEMPTS]
Produce a working session document in this order: 1. Restate the challenge as a neutral problem frame: person, situation, desired outcome, and current barrier. List facts separately from assumptions. 2. Create a stakeholder and journey snapshot. Identify the moment of friction, triggers, jobs to be done, functional and emotional needs, and constraints. Mark unknowns explicitly. 3. Write 3 “How might we” questions. Each must be broad enough for multiple concepts but narrow enough to guide action; avoid embedding a solution. 4. Generate 10 concept directions across at least four different approaches, such as changing information, workflow, incentives, access, or support. For each, state the mechanism, target need, dependency, and likely risk. 5. Prioritize concepts in a table using desirability, feasibility, viability, and learning value. Explain ratings only from supplied facts. 6. Recommend one low-cost experiment with a hypothesis, audience, prototype or stimulus, success signal, disconfirming signal, and decision rule. Ask up to 3 clarifying questions only if a required input is missing. Before answering, verify that no concept is presented as validated and that the experiment can disprove the key assumption.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [CHALLENGE] | Describe the problem and the decision your team needs to make. | Reduce missed first appointments at a community dental clinic without increasing front-desk workload |
| [PEOPLE AND CONTEXT] | Describe the people affected, their setting, and the situation in which the problem occurs. | Adults receiving Medicaid who book by phone; appointments are often scheduled 2–4 weeks ahead |
| [EVIDENCE] | Provide research findings, operational data, observations, or clearly labeled assumptions. | Clinic data shows 22% no-shows; call logs show many patients ask for directions and transportation information |
| [CONSTRAINTS] | List budget, timeline, policy, technical, accessibility, and operational limits. | No new software this quarter; reminders must be available in English and Spanish; staff can spend 30 minutes weekly |
| [PRIOR ATTEMPTS] | Describe prior solutions, pilots, or approaches and what happened. | Automated reminder calls were added last year, but the clinic did not measure whether patients listened |
How to use
- Paste factual notes and label assumptions plainly instead of presenting them as research findings.
- Use the problem frame to align the team before discussing concepts; amend it if it describes a solution rather than a user outcome.
- Check the proposed experiment has both a success signal and a result that would change the team’s mind.
- Then send: “Create a one-page experiment brief for the selected concept, including recruiting criteria, test script, measurement plan, and stop conditions.”
Variations
Problem reframing
Use this when the team is stuck with a solution-shaped brief.
Reframe [CURRENT BRIEF] using the evidence in [EVIDENCE] and constraints in [CONSTRAINTS]. Produce three alternative problem frames, each naming the affected person, situation, desired outcome, and barrier. For each frame, state what it assumes, what evidence supports it, and the riskiest unknown. Write two neutral “How might we” questions per frame without naming a product or feature. Do not invent research findings. Ask up to 3 questions only if required inputs are absent. Check that each frame describes a user outcome rather than an internal deliverable.
Concept workshop
Use this when a group needs structured prompts for an ideation workshop.
Design a [WORKSHOP LENGTH] concept workshop for [CHALLENGE] with [ATTENDEES]. Use [EVIDENCE] as the starting point. Provide a timed agenda, individual and group exercises, divergence prompts, a method to combine ideas, and a decision rubric using desirability, feasibility, viability, and learning value. Include rules that prevent seniority from deciding the outcome. Do not treat brainstormed ideas as research validation. Ask up to 3 clarifying questions only if necessary. Check that the agenda reserves time to define the hypothesis behind the chosen concept.
Experiment design
Use this when the concept is selected but its main assumption remains untested.
Design a test for [CONCEPT] intended to improve [USER OUTCOME]. The central assumption is [ASSUMPTION], for [TARGET PEOPLE], under [CONSTRAINTS]. Produce a lean experiment plan: hypothesis, smallest realistic prototype, recruitment criteria, participant task, measures, success threshold, disconfirming result, ethical considerations, and what decision each result supports. Do not fabricate benchmarks or promise causation from a small test. Ask up to 3 clarifying questions only if required inputs are missing. Check that the test observes behavior, not only stated preference.
Tips
- Keep an evidence ledger during the session: observations and data in one column, interpretations in another, and assumptions in a third.
- A useful “How might we” question describes an outcome and context but does not contain the proposed channel, feature, or policy.
- Generate concept directions before scoring them; early feasibility debate often eliminates the alternatives that reveal a better framing.
- Define the decision rule before testing so favorable anecdotes do not outweigh the agreed success signal.
FAQ
Can design thinking work without user research?
It can identify assumptions and create experiments, but it cannot replace evidence from the people and context affected by the problem.
How many ideas should we generate?
Ten diverse directions is often enough for a small working session. Quality comes from covering different mechanisms, not from producing dozens of slight variations.
What makes an experiment useful?
It tests a decision-critical assumption, has an observable outcome, and states what result would cause the team to revise its plan.