Say it.

See it.

Say it better.

Say it.

See it.

Say it better.

Say it.

See it.

Say it better.

Savvy is an AI interview coach that listens while you answer out loud, shows you what you said, and hands it back in a form you can improve, then asks you to try again.

[Role]

Product & interaction design, AI behavior design, native prototype

[Duration]

Working iOS prototype, Team

[Tech Stack]

SwiftUI, Apple Speech, LLM feedback

AI UX

Conversational UX

Voice Interaction

Knowing what to change is harder than getting feedback.

You can rehearse an answer fifty times and never see it. So when someone says “be more concise,” there's nothing to apply it to. The problem wasn't getting feedback. It was knowing what to change in the next answer.

Your actual answer

Generic feedback

Question → Answer → Feedback → Retry

Every question completes the loop before you move on.
Retry is part of the product, not a fallback.

01 Understand

02 Speak

03 Feedback

04 Retry

Helpful enough to recover.
Quiet enough to let you think.

The first version showed the full STAR structure. The final version uses lightweight memory prompts — enough to find your place, not enough to read from.

Early Screen

Final Screen

  • Labels became memory prompts, enough to find your place, not enough to read from.

  • The panel hides with one tap, and the coach remembers that.

  • The live transcript stays small so you keep talking instead of editing yourself.

  • In Interview mode it starts hidden.

  • Labels became memory prompts, enough to find your place, not enough to read from.

  • The panel hides with one tap, and the coach remembers that.

  • The live transcript stays small so you keep talking instead of editing yourself.

  • In Interview mode it starts hidden.

All Works