Spoken answers, follow-ups and answer summaries

Three optional helpers for written screenings: spoken answers and answer summaries use AI, and in English a follow-up question may quote words AI picked from the answer. Each is off until you switch it on, and none of them ever scores, ranks or judges a candidate.

In short

A written screening works without any AI. In Hiring settings you can switch on up to three optional helpers: spoken answers, follow-up questions and answer summaries. None of them ever scores, ranks or judges a candidate; people read every answer and decide.

Before you start

• Who: people who manage hiring switch the helpers on. Every reviewer reads the results.
• Plan: spoken answers need a plan that includes heavier AI work and a transcription service; the others use a little of the AI budget.
• First: each helper is off until you switch it on in "Hiring settings", and an organization can switch AI off entirely.

Switch them on

1. Open Jobs › "Hiring" › "More" › "Hiring settings".
2. Under "How you hire", switch on what you want.
3. Press Save. The next screening you send tells its candidate exactly which helpers apply.

Approve follow-up questions

1. Open the role and its "Interview kit".
2. Under a question, open "Follow-up questions" and tick what may be asked after it. You can tick at most two: a concrete example, their own part, how it turned out, why that approach, or what they would do differently.
3. Optionally write "Your wording". Left empty, the ready sentence shown is used.
4. Save the kit, then send the screening with that question picked from the kit.

Which languages

A language gets a helper only after {{brand}} measured it on answers it was not tuned on. Quoted words must never reach into something private a candidate disclosed. And at least four in five ordinary answers must offer words to quote. The language is the one the candidate applied in, not your screen's.

What you see when you review

• An answer that began as a recording is marked "Transcribed by AI, checked by the candidate".
• Follow-ups sit under the answer they followed, marked "(follow-up)" - or "(follow-up, topic chosen by AI)" when AI picked the quoted words.
• A summary appears a moment after the candidate answers, for answers of a few sentences or more. Read the answer; the summary is only a way in.

Good to know

• Each switch applies to screenings sent from then on; switching one off stops it at once, also for screenings already out.
• Spoken answers are offered only when your plan includes heavier AI work and the platform has a transcription service.
• With AI switched off for your organization, none of the three calls AI.
• Questions you type for one candidate while sending a screening have no follow-ups.

If something goes wrong

• "A question picked is not in the role's interview kit.": add it to the kit first.
• A question about a protected topic (age, family, religion, health and others) cannot be asked or saved. Ask about the work instead.
• Summaries did not come: ask again from the answer; if the AI budget is used up they wait until it resets.

Related

• /help/jobs/answering-screening-questions
• /help/jobs/cv-summaries
• /help/ai/what-the-assistant-reads