Customer discovery

Customer discovery

Surveys give you breadth: hundreds of answers to the questions you thought to ask. Customer discovery gives you depth: an hour with a real person, learning how they think about the problem — including the parts you didn’t know to ask about. This is where the why behind the problem lives.

Interview scripts

You don’t interview people with zero preparation. The discovery engine generates interview scripts tailored to your idea: a structured set of questions built for customer interviews, generated by AI from your idea’s context.

Each script is tuned by parameters like the stage of your product, the type of application (for example, B2B software), and the market you’re targeting — so a script for an early-stage B2B tool asks different questions than one for a consumer product. You can edit questions, and the generator can produce new versions of a script as your understanding evolves.

Sessions

A session is one interview with one person:

• Pick a script to follow, and optionally attach a contact from your audience panel.
• Record the essentials as you go — who they are, their role, the date, and the duration.
• Keep notes during the call.
• Paste a transcript or summary afterwards, and note the overall sentiment of the conversation.
• Track the session’s status (scheduled, completed, and so on) so your interview pipeline stays organized.

Sessions are lightweight on purpose: the value is in doing many of them and capturing the pattern, not in perfect records of one call.

Insights

After each interview, capture the insights — the specific things the person said that matter. Each insight is tagged with a category:

• Pain point theme — a problem they described.
• Job-to-be-done theme — a task they’re trying to accomplish.
• Opportunity area — a gap or opening you spotted.

You can rate each insight’s intensity (how strongly they felt it) and tag it for later filtering. The more interviews you log, the more these insights accumulate into a picture of what your market actually needs.

Themes

A single interview can be an outlier. Themes are the patterns that emerge across many interviews — the pain points and jobs that come up again and again, from different people, in different words. The discovery engine helps surface these recurring patterns automatically, so you don’t have to re-read every transcript to notice that “I waste an hour a week on this” appeared in seven of ten interviews.

Note

A theme that shows up once is a story. A theme that shows up five times is a feature. When you’re deciding what to build, weight recurring themes far above isolated comments.

The jobs-to-be-done map

The JTBD map organizes your insights around the jobs your customers are trying to get done — the functional, social, and emotional tasks behind their behavior. Instead of a list of complaints, you get a map that shows:

• What jobs people are hiring a product for.
• Where the current solutions fall short.
• Which jobs are most important (and most underserved).

That map is directly useful: it tells you what your product should actually be for, and it gives you the language to describe it in landing pages and validation tests that speak the customer’s language.

A practical flow

1. Generate an interview script for your idea.
2. Recruit 5–10 people who match your target customer — pull from your audience panel.
3. Interview them, log each session, and capture insights right after.
4. Let themes emerge across interviews.
5. Build the JTBD map to see the jobs clearly.
6. Turn the strongest themes into survey questions or fake-door tests to check them at scale.

Aim for quantity: five honest interviews beat one polished one. The patterns you need will usually appear by interview five or six — if the story is still unclear then, keep going.