Interview preparation
Practice in conditions close to the real thing before the big day.
In short
- Full voice simulation: question read aloud, answer by voice or by typing
- Every follow-up question is generated live from the entire conversation
- Scored feedback after each answer, then a full summary at the end
A mic check before you start
Before every simulation, a technical check screen confirms your microphone is properly detected and sound is working, so you don't discover a technical issue once the interview has already started.
How a "dynamic" follow-up question actually works
This is the core mechanic, and it's worth explaining precisely. There's no pre-built question bank CvCy Talent pulls from as the interview goes on. Every question is generated live, one at a time, from the full history of your previous answers. The AI receives the entire conversation and decides, turn by turn, whether to dig into something that stayed vague in your last answer, build on something you just mentioned, or explore a fresh angle when there's nothing specific to follow up on.
In practice, that means two people answering the same first question differently will then get different follow-up questions: the interview genuinely adapts to what you say, it doesn't follow a fixed script.
Answer out loud or by typing, your choice
Every question is shown on screen and automatically read aloud, with a button to replay it. You can answer by typing, or by holding the microphone button: your voice is recorded and automatically transcribed into text, which you can then edit before submitting. Both modes can be combined within the same answer.
Feedback after every answer, not just at the end
As soon as you submit an answer, the AI evaluates it and gives you a satisfaction score out of 100, with one identified strength and one concrete improvement point, in an honest but supportive tone, like a professional coach. That feedback shows up through a visual indicator before the next question even appears, so you adjust your approach as the interview unfolds rather than after the fact.
A full summary at the end
Once the interview is over (after six exchanges, or earlier if you choose to stop), the AI re-reads the entire transcript to produce a structured summary: an overall score, an assessment, a recommendation (hire, maybe, or not yet ready), up to four strengths, and up to four areas to improve. The instruction given to the model explicitly asks it to stay honest if answers were empty, evasive, or low-effort, rather than artificially inflating the score.