Validate your top idea by talking to real potential customers
This week tests your assumptions through real conversations. AI can help you prepare, transcribe, and organise interviews, but the evidence must come from the customer rather than from the model.
Strong interviews focus on a recent event. Ask what happened, what the person tried, how they chose a workaround, and what the result cost them. Questions about future intentions often produce polite guesses. Questions about the past produce details you can examine.
Do not ask whether someone likes your idea or would use your product. That invites agreement. Ask them to describe the last time the problem occurred and what they did next. Stay curious when the answer challenges your idea.
Define a simple screening rule before outreach. A suitable participant should have experienced the problem recently and fit your target group. Use a short invitation that explains the research topic, the time required, and that the conversation is not a sales call.
After each interview, record facts, quotes, behaviours, and your interpretation separately. Once several interviews are complete, AI can cluster observations and surface patterns. Check every summary against the original notes. Preserve outliers because they may reveal a different segment or an important risk.
Approximately 331,000 views
Why it fits: Shows students how to run useful customer interviews without pitching their solution or asking leading questions.
Prepare five interview questions and conduct one real customer conversation.
The video supports the lesson above — it doesn't replace it.
Review this interview transcript as a careful research analyst. Separate direct observations from interpretation. Identify the situation, current workflow, pain, consequence, purchase triggers, and objections. Quote only words that appear in the transcript. List contradictions and unanswered questions. Do not recommend a product yet.
Week 3 · Case study
Source note: Documented customer example from Replit. The download figure is a reported result, not an independently verified measure. The student-sized version below is an illustrative exercise, not part of the reported case.
Bamboo Classroom was created by a UK primary-school teacher who wanted classroom activities to feel more interactive. Because she understood the users and environment personally, she could describe the problem precisely and test the product with fellow educators.
Replit reports that the application received approximately 1,400 downloads overnight after being shared with the teaching community. The important part is not the download spike. The founder already belonged to a community with the problem, understood its vocabulary, and could reach relevant users without a large advertising budget.
Interview five people from a community you already belong to—students, teachers, freelancers, gym members, gamers, or small-business owners. Ask them to demonstrate their current process before suggesting a solution.
What would you need to observe before deciding which classroom feature to build first?
Respond in your submission →Complete ten customer interviews and submit an AI generated insight summary that you have checked against your notes. Include participant criteria, the script, a pattern table, surprising evidence, and what you will test next.