JobCoach

Turn an AI-assisted portfolio project into an interview you can defend

Use a decision, a rejected suggestion and a piece of evidence to tell the story of your contribution.

By Ashley Kays Rabbitt · November 18, 2026

“The AI built my prototype in an afternoon” sounds quick. It leaves an interviewer wondering what judgment you brought to the work. Try telling the story from the moment you overruled it.

Imagine a candidate redesigning a fictional appointment form. The AI proposes “Booking confirmed” after a customer selects a time. The business rule says staff must confirm availability. The candidate changes the message to “Time requested,” keeps the customer’s details on screen, and adds the next step. That is a defensible design decision even when the prototype has never served a real customer.

An interview answer a reviewer can inspect
SuggestionShow Booking confirmed
ConstraintStaff must check availability
DecisionShow Time requested with next step

Assemble three pieces of evidence

Show the original state, the rule that made it wrong, and the revised state. Identify what you personally chose and what the tool generated. Include any tests you actually ran. If you tried it only with sample data, describe that limit instead of claiming a conversion improvement.

Borrow a principle from technical writing: match your explanation to the knowledge and goal of the audience. A design lead may ask how people interpreted the receipt. An engineering lead may ask how you distinguished a request from a confirmed booking in the data. Prepare a concise answer for both. Google Technical Writing • Learning objectives and audience

Practice with one follow-up: “What would change your mind?” A strong answer might be, “If the shop starts reserving inventory at selection, I would revisit the message and test the new commitment.” It shows you understand why the design works under its current assumptions. That is something a screenshot of an attractive form cannot establish.