A "How might we..." question is the bridge between a defined problem and ideation in design thinking. It reframes an obstacle as an opportunity: open enough to allow several solutions and narrow enough to guide the search. The generator applies seven transformations to your problem, such as amplifying the good, removing the bad and questioning an assumption, and checks for an embedded solution.
How to use
- Describe the problem or insight as you observed it, without proposing a solution.
- Enter the user, the need (starting with a verb) and the obstacle (as a noun phrase).
- Optionally, add what already works, the current way, an assumption, a real constraint and an analogy from another sector.
- Read the questions by transformation and the checks: embedded solution, need too broad or too narrow, generic user.
- Pick 3 to 5 questions for the ideation session and export as PNG, PDF, DOCX or Markdown.
What each field asks for
| Field | What to enter |
|---|---|
| Problem or insight | What you observed, with whom and why it matters. No solution yet. |
| User | Who lives the problem, as specifically as possible. |
| Need | What the person wants to achieve, starting with a verb. |
| Obstacle | What gets in the way today, written as a noun phrase. |
| What already works (optional) | Something people like that is worth amplifying. |
| How it is done today (optional) | The current way of solving it, which could change. |
| Assumption to question (optional) | A rule everyone accepts without discussion. |
| Constraint (optional) | A real limit, starting with "without" or "with". |
| Analogy (optional) | Something from another sector that already solves a similar problem. |
Example
A fictional online school notices that working students drop the course in week two because they lose their place. With the need "finish the course in short breaks during the day" and the obstacle "the loss of their place when they resume a lesson", the generator creates 13 questions in nine groups. Among them: "How might we remove the loss of their place when they resume a lesson?" and "How might we help working students finish the course in short breaks during the day without the assumption that each lesson must be watched from start to finish?". All five checks pass.
When to use
- After research synthesis, when the problem is defined and the team needs to start generating ideas.
- When a request arrives with the solution already decided and you need to get back to the need behind it.
- To prepare an ideation session with questions that pull in different directions.
- To check whether the framing of the challenge is too broad or too narrow.
Frequently asked questions
What is a "How might we" question?
It is a question that starts with "How might we" and turns a problem into an opportunity to create. "How" assumes a solution exists, "might" makes clear it is still a possibility, and the rest of the sentence points to someone’s need.
How do I know if the question is too broad?
If it would fit any product, such as "improve the experience", it is too broad. A good question says who the person is and what they want to achieve, in a recognisable situation.
And too narrow?
When it already points to the solution or carries an exact numeric target, there is little room for different ideas. Leave the target to the success metric and keep the question open.
What is an embedded solution?
It is when the problem already names the answer, such as an app, a notification or a screen. The generator looks for those words and flags them, because they close off options before ideation.
How many questions should I take to the ideation session?
Three to five is usually enough. Prefer questions that pull in different directions, not variations of the same idea, and give each one a fixed amount of time.
Is AI required?
No. The questions by transformation and the checks are produced by rules, instantly. When AI is available, it only suggests extra questions, which also go through the embedded-solution check.