Setting up an AI voice screen
Create an AI voice screen for a role: write questions and a rubric, choose camera, languages and voices, set the welcome screen, publish and invite candidates.
An AI screen is a short, structured voice interview that the agent runs with candidates, in their language and when it suits them. Every answer is rated against your rubric, and a person decides who moves on. Screens live in AI screens.
Create a screen
Choose New screen and pick the role. The screen starts as a draft in English with two questions that suit any role, ready to make your own. A role can have more than one screen, for example one per stage. You can also ask the agent on Home for a screen; its draft has four questions and takes about 13 minutes.
Each screen has four tabs: Questions, Candidates, Settings and Invite.
Write the questions, the rubric and how long it takes
On Questions, each question has:
- the question the interviewer asks
- Reveals: what a good answer tells you
- Strong and Weak: what a strong answer includes and what a weak one looks like. This is the rubric each answer is rated against.
- the minutes it should take, from 1 to 20
- follow-ups: when on, the interviewer follows up when an answer is thin
Reorder questions with the arrows or remove them. Changes save as you type. Emoti adds up the minutes: under 10 is short, up to 25 is comfortable in one sitting, and over 25 is a lot to ask before a first conversation. Write the rubric before anyone takes the screen, so every candidate is judged the same way. See Structured, fair screening.
Camera, languages and the welcome screen
Changes on Settings apply to anyone who starts the screen afterwards.
| Camera | What it means |
|---|---|
| Required | Candidates turn their camera on to start. Use it when you need to see who is answering. |
| Optional | Candidates decide. The scorecard never depends on it. |
| Off | Audio only. Candidates are never asked for camera access. |
Languages and voices: add any of English, Mandarin Chinese, Malay, Hindi, Japanese, German, French, Spanish, Portuguese and Polish. The interviewer speaks and listens in each, with the voice you pick: Warm, mid; Calm, low; Bright, clear; or Steady, measured. The first language is offered first, and candidates choose one on the welcome screen.
Welcome screen: the company name, a headline and a message. Candidates also see the length, the camera setting, and that an AI interviewer asks the questions and people at the company review every answer before any decision. To start, they tick a box to confirm they understand this and that their answers are recorded and transcribed.
The welcome screen also offers Need an adjustment, or rather talk to a person? Candidates can ask for more time for each answer, written answers instead of speaking, a conversation with a person instead, or something else, and the page tells them the request goes to your recruiting team. If a candidate asks for another way to take part, talk to them directly. See Asking for adjustments.
Draft, Live and Paused
| Status | Can candidates take it? |
|---|---|
| Draft | No. Choose Publish when it's ready. |
| Live | Yes. |
| Paused | No new starts until you choose Resume. |
Every question needs wording before you can publish. Choose Preview as candidate to go through the screen as a candidate would. The preview doesn't use your microphone or record anything.
Invite candidates
Once the screen is live, open Invite:
- Share a link: anyone with the link can take the screen. Post it on a job ad or send it yourself.
- Invite by email: add the candidate's name, email and language, then choose Send invite. Emoti sends the invite and one reminder, and the candidate takes the screen when it suits them. What they receive shows the email first: it says an AI interviewer asks the questions, that a person reviews every answer, and that the link lets them ask for an adjustment before they start.
Invited candidates appear on the Candidates tab, and their scorecards arrive as they finish. See Reading screen results and scorecards, and for the candidate's side, Taking an AI voice screen.
Questions people ask about this
Genie answers from this article and the rest of the help center.
Related articles
- Reading screen results and scorecardsWhat a finished AI screen's scorecard shows, how the 0 to 100 score is worked out, and how to advance or reject candidates after a screen.
- Taking an AI voice screenWhat happens in an Emoti AI screen, how long it takes, what you can control, and how to give your best answers.
- Structured, fair screeningHow to screen candidates consistently. Ask everyone the same questions, write the rubric before anyone is reviewed, know what each question is for, follow up fairly, score independently and run debriefs that decide.
- Ranking applicants with an ideal-candidate profileHow Emoti ranks inbound applicants against each role's ideal-candidate profile, how to change the profile, and how to advance, hold or reject applicants in bulk.
- Asking for adjustmentsHow to ask for adjustments to an AI screen or interview, such as more time, a different format, a person instead of an AI interviewer or your camera off, and when to ask.