Tell me about something you shipped with AI's help that you could not have delivered otherwise -- or could not have delivered anywhere near as fast. Walk me through what you built, how you actually used AI, where you decided not to use it, and how you knew the output was correct.
The interviewer is looking for: a concrete deliverable rather than a general productivity claim, judgment about where AI fit and where it did not, the verification you did before shipping, and what became possible that was not possible before.
Note: this is not your favorite project overall -- it is specifically about the AI-assisted work and the judgment behind it.
How to approach it
- Hint 1
The interviewer is assessing the Work area: what you have actually shipped with AI, and whether you treat it as a tool you exercise judgment over. Vague claims like AI makes me faster carry zero signal -- they want one artifact, a before state, and an after state.
- Hint 2
Pick a story with a named deliverable and a real stake: a prototype built in a day that changed what leadership decided, a migration across hundreds of files, a bug investigation you would not have had time for, a design doc or postmortem you drafted and then rewrote. Beyond-codegen uses count for a lot here. Crucially, choose an effort where you also decided NOT to use AI somewhere -- the security-sensitive path, the piece needing context AI could not have -- because that boundary is half the signal.
- Hint 3
Deliver it in CARL. Keep context to about ten percent and lead with the stakes. Spend most of the answer on your actions in first person -- the specific prompts or loops you ran, what you rejected and why, how you verified (tests you wrote, review you asked for, a spec you diffed against). Close with a before/after result and one learning that changed your practice. Include the part that did not work; a flawless AI story reads as a fairy tale.
Ready to answer it out loud?
Record your answer in 2:00 and Preptile scores it 1–10 with specifics — what landed, what you skipped, and what to say next time.
Practising needs an invite code. Join the waitlist and we’ll send you one.