Find the assumption behind the idea
Describe the person you want to help, the problem, and their current workaround. Separate evidence from guesses. An AI “cofounder” is a planning role you assign to software, not a person who can validate demand or take responsibility for your business.
Your activity
Give AI your idea and ask for competing explanations of the problem. Request questions that would disprove your assumptions. Do not ask it to invent customer quotes or market statistics.
An assumption map and five real customer interview questions.
Check your work
Can you explain the decisions you made? What did you verify? Identify one remaining uncertainty and one improvement before moving on.
Design a small validation experiment
Pick the riskiest assumption. Choose a low-cost test such as interviews, a manually delivered service, or a prototype. Define what you will observe and what would make you continue, change direction, or stop.
Your activity
Write an experiment for a fictional service that organizes family photos. Define the audience, recruitment method, evidence to collect, cost limit, and decision rule.
A one-week validation plan with a clear decision point.
Check your work
Can you explain the decisions you made? What did you verify? Identify one remaining uncertainty and one improvement before moving on.
Turn learning into a build plan
Use actual feedback to narrow the idea. Ask AI to compare options against your constraints, break the selected approach into small tasks, and flag dependencies. Keep ownership of spending, commitments, and major decisions.
Your activity
Translate your evidence into three milestones: prove the task, make the experience usable, and test it with people. Identify the owner and completion check for each.
A realistic next-step plan with dependencies and unresolved questions.
Check your work
Can you explain the decisions you made? What did you verify? Identify one remaining uncertainty and one improvement before moving on.