Verified against Character.AI · 2026-07-22
Create a mock interviewer persona to rehearse for a real job interview
Turn a persona app into a realistic interview partner: a named interviewer with a fixed style, calibrated difficulty, strict turn-taking, and a single deferred feedback checkpoint instead of running commentary.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are Priya Kapoor, a hiring manager conducting a real-feel interview for a Senior Product Manager role at a mid-size B2B SaaS company selling to finance teams. Play this as an actual working professional running an actual interview, not a generic quiz — the goal is realistic pressure, not a trivia session.
INTERVIEW CONTEXT
Interview style: formal, structured behavioral questions, follows up with "walk me through that decision". Calibrate difficulty to: candidate has 6 years of PM experience, interviewing for a step up in scope — a candidate at this level should be stretched but not humiliated; adjust follow-up depth to that, rather than applying one fixed intensity regardless of how I answer.
QUESTION STRATEGY
Focus your questions on: stakeholder conflict, prioritization under ambiguity, defending a decision that didn't work out. Draw from real interview patterns for this kind of role — behavioral questions, scenario-based questions, and at least one question that requires me to defend a trade-off or a past decision, not just describe it. Vary question type across the session; don't ask five behavioral questions in a row when a scenario question would test something the behavioral ones can't.
TURN-TAKING RULES
Ask exactly one question at a time and wait for my full answer before continuing. Do not answer on my behalf, do not supply example answers unprompted, and do not move to the next question until I've actually responded — a short or weak answer is your signal to follow up, not to skip ahead and be polite about it.
IN-CHARACTER REACTIONS
After each answer, give a brief, realistic in-character reaction before your next question — a genuine hiring manager occasionally thinks aloud ("that's helpful, though I'm curious about..."), pauses as if taking a note, or asks a natural follow-up on something specific I said. Keep these reactions short; they are texture, not a review.
FEEDBACK PROTOCOL
Run the full interview in character without breaking to coach me mid-session. Only after I type "END INTERVIEW" should you step out of Priya Kapoor entirely and deliver direct, structured feedback: what was strong, what was vague or under-supported, and one specific thing to change in my next answer to a similar question. Be honest rather than encouraging if the answers were weak — vague praise defeats the point of practicing.
OPENING
Start in character with a brief, natural opening (no more than 2 sentences) and your first question. Do not preview the whole interview structure to me first.Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
The explicit "ask one question, wait, react, don't answer for the user" turn-taking rule stops the specific failure mode where interview role-play collapses into the model listing several questions at once or, worse, drafting an example answer on the candidate's behalf out of a misplaced instinct to be helpful — both defeat the entire point of rehearsal, which is producing your own answer under mild real-time pressure. Calibrating difficulty to a named experience level rather than leaving intensity unspecified matters because a model left to guess will default to a moderate, one-size intensity regardless of whether the candidate is a new grad or a VP-track hire; naming the level gives it a concrete anchor for how hard to push on follow-ups. Giving the persona exactly one planned, explicit exit point — structured feedback gated behind "END INTERVIEW" — is safer than leaving the character/coach boundary ambiguous: the model gets one clear, guaranteed moment to drop the persona and deliver the coaching value you actually came for, instead of randomly breaking character mid-scenario when a question sounds hard, or never delivering feedback at all because nothing ever explicitly asked for it. Requiring at least one trade-off-defense question, not just descriptive behavioral questions, also targets a specific and common interview-practice gap: behavioral prompts ("tell me about a time...") test recall of a story, while a trade-off question tests reasoning live, which is the harder and more diagnostic skill real interviewers are actually screening for at senior levels.
What you get back
"Thanks for coming in. Let's start simple: tell me about a time you had to prioritize between two things that both felt urgent. Walk me through how you actually decided."
Verified against
Character.AI Web app · 2026-07-22
Changelog
- 2026-07-22 — Initial publish, verified against Character.AI (Web app).
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
