Mercor, micro1 and other AI interviews: what each is like

A Mercor AI interview is a recorded video call, usually about 20 minutes long, with an AI interviewer that asks questions based on your résumé and the job, then transcribes and evaluates your answers. You can retake it up to three times, and the result carries over to other Mercor listings that use the same interview. micro1's interview is a spoken conversation with Zara, its AI recruiter, and lasts 20 to 40 minutes depending on how many skills the role needs. Braintrust's AIR interview arrives as a link by email or text after you apply, and most take 15 to 20 minutes. Each fact here comes from the platform's own pages, read on 15 October 2026, and the table at the end of this page cites a source for every cell.

Mercor and micro1 both recruit experts for project work. Mercor says it uses AI "to efficiently match experts with projects at scale". micro1's expert library describes work that helps "train and improve the performance of AI models".

How the Mercor AI interview works

Mercor's help centre says it "employs an AI interviewer to pose role-specific questions, generate a transcript of your responses, and evaluate your performance for the listing". Most interviews take "approximately 20 minutes, although some may be shorter or longer". Mercor's engineering blog from March 2026 names the interviewer Monty and says "around 10,000 of these conversations happen every day, each one fifteen minutes long".

You start the interview from your dashboard whenever you choose. The help centre lists what to expect:

• English, "unless otherwise specified", with captions on by default.
• Some interviews are technical and go deep into your past work; others "test language or reasoning skills".
• Screen sharing "may be required", and only full-screen sharing works. Some interviews include a whiteboard.
• Long pauses may make the AI cut in. If it interrupts you, "you can let it know you weren't finished", and you can ask it to repeat a question.

You can retake an interview "up to three times across all applications related to that interview". The waiting room asks you to keep retakes for technical problems, hiring managers can see how many you used, and a retake resets your application progress. An attempt that ends well short of its planned length shows as "Interview too short" and needs a retake.

You take each interview only once: "if another application necessitates the same interview, your existing result will be retained." There is also a free Practice Interview, which "does not affect future opportunities" and is "not shared with any companies".

You cannot ask for a human interviewer instead. Mercor says its "AI interview is the standard and only method for screening candidates", and interviews "cannot be conducted via chat, call, or any other medium".

How the micro1 AI interview works

micro1's library says Zara "conducts the interview for all candidates applying to roles on the platform". Each role has skill requirements set by micro1's recruiters from the client's needs, and Zara's questions are "tailored to the specific skills required for that role". The interview runs in real time: Zara asks open-ended questions and you answer out loud.

Interviews "are between 20 and 40 minutes long, depending on how many skills the candidate is evaluated on", and one skill "generally takes around 7 minutes".

micro1's interview guide lists the question types: open-ended technical questions for your chosen role, scenario questions on how you "think, problem-solve, and communicate", a coding challenge for technical roles and a "human data exercise" for annotator roles. Its questions page says "no two candidates receive the same questions" and gives sample questions for roles such as full-stack developer, account executive and project manager.

You get "a real-time result indicating whether you meet micro1's global certification criteria". The session is recorded "so our human recruiters can review results". The talent dashboard shows "Under review" while recruiters assess the interview, which "can take up to 48 hours", then a status such as "Hiring manager reviewing", which can last up to 30 days. If you did not qualify, you can ask for written feedback on "what went well and areas to improve".

For interviews completed after July 2025 that met the certification bar, Zara checks your certified skills against a new role and may assess only the missing skills, or skip the interview entirely.

Braintrust AIR, Alignerr and HireVue

Braintrust's AIR (AI Recruiter) is run by a company's own hiring team. Braintrust's candidate page says you receive a link by email or text, need no download or account, and can take it on a phone or computer. Most interviews last 15 to 20 minutes. "Typically, you get one chance per question", and most interviews let you resume if the connection drops. Braintrust says the AI "analyzes your words, not your appearance, tone, or background", produces a scorecard for the hiring team, and that "all decisions are made by human reviewers".

The hiring team also sets how many attempts you get, for the whole company or for one job. Once you reach that limit, you can ask for one more attempt with a short explanation. The team approves or declines, and its decision cannot be changed.

Alignerr's onboarding page says some jobs need "a Zara™ AI interview or a domain-specific assessment", with instructions sent to you if so. Zara is also the name micro1 gives its AI recruiter.

If an employer has sent you a HireVue link, see our HireVue interview guide.

Using AI tools and sharing questions

Mercor allows AI for grammar and wording only. Its help centre says to "refrain from utilizing large language models to compose your responses, rate model responses, generate rationales or explanations, or evaluate code", and its LLM Usage Policy warns that "submissions that are evidently produced by artificial intelligence will be identified and flagged". micro1's candidate terms (last updated 21 July 2026) say you will answer "without any third-party assistance, except where explicitly permitted by your interviewer or the interview instructions".

Both platforms keep questions confidential. Mercor lets you "discuss general experiences" but not "specific questions, screenshots, or any confidential information". micro1's terms say interview content is "strictly confidential" and may not be copied or screenshotted, and that breaking the terms can lead to "temporary, or permanent withdrawal" of your right to use its platform. Questions are also generated for each candidate, from your résumé and the job at Mercor and from the role's skills at micro1, which says "no two candidates receive the same questions". A list of "real questions" posted online is unlikely to match yours.

What happens to your recording

Mercor keeps "video/audio recordings and an AI-generated transcript of official interviews to evaluate your candidacy and to recommend your profile to potential employers", and says it does not sell interview data. Its help centre says interview data "will not be employed to train AI models"; its data page says it may train Mercor's "internal recommendation models" but never "third-party large language models". You cannot download your recording or transcript. You can delete your account and its data from Account Settings, unless you have an active contract, a payout in progress or an unanswered offer, and Mercor says deletion "does not remove data from companies that may have already accessed it".

By accepting micro1's candidate terms, you consent to micro1 recording "your answers, audio, video, and platform interactions" and sharing the recording and your results under its privacy policy.

How to prepare

The platforms' own advice is mostly practical setup. Our guide on how to prepare for an AI interview covers the rest.

• For Mercor, use a computer with Chrome, test your camera and microphone, and sit where your full face is visible. Headphones and an external microphone are recommended. Keep pen and paper nearby, and a calculator if your field needs one.
• For micro1, run the system check linked from its help centre, close other apps and tabs, and choose "one core skill or job type" to be evaluated on. Its Interview Prep tool simulates an interview first.
• For Braintrust AIR, find "a quiet, well-lit space where you won't be interrupted".
• For all of them, prepare to talk through real work you have done. Mercor's questions come from your résumé and the job, and micro1's from the skills the role asks for.

How we compared

We read each platform's own candidate pages on 15 October 2026: Mercor's talent help centre, data page, LLM policy and engineering blog; micro1's interview guide, questions page, expert library, help centre and candidate terms; Braintrust's two AIR support articles; and Alignerr's onboarding page. Every cell in the table cites one of them. "Not stated" means we found nothing on that point in the pages we read, not that the platform does not do it. Third-party prep sites and forum posts were left out, because they cannot be checked against the platform. The table ranks no platform.

These pages change, so read the current version and the instructions in your own invitation before you start.

How a ROUND.1 interview differs

ROUND.1, a hiring marketplace for employers and job seekers in Hong Kong and Singapore built by h.work, runs its own AI first-round video interview. Its AI information page sets out how it works.

• Each candidate takes one interview, which ROUND.1 scores and reuses for the other roles the candidate chooses.
• It runs in a web browser with a camera and a microphone, and lasts about 20 minutes by default (roles can be shorter), 30 at most. The candidate first consents to the processing of their video and voice under Hong Kong's PDPO and Singapore's PDPA.
• There is one attempt and no restart. If the connection drops, the same interview can resume and continue, once by default (a role can allow more resumes or none).
• Scores are given competency by competency, each backed by transcript quotes. Accent, pronunciation, grammar, fluency, personality traits and protected characteristics are not scored. Each role uses one language, English or Mandarin Chinese.
• A company first sees the scored evaluation without the candidate's name or contact details, which open when it shortlists the candidate. A person at the company decides whom to hire.