How Clinical AI Interviews Are Structured and Scored

Clinical AI-training interviews are not standardised. The format depends on the platform, the role and sometimes the individual project. This page sets out what the two platforms ACJ tracks most closely actually document, and the preparation principles that hold whichever format you are given.

Interview structure varies by platform

There is no single standard format for clinical AI-training interviews. The format can be conversational, asynchronous or task-based depending on the platform and role.

Mercor

Mercor currently uses an AI-led conversational interview for many roles. The interviewer listens to your answer and can ask follow-up questions based on what you say.

Mercor’s own talent documentation sets out the mechanics:

• An AI interviewer asks role-specific questions, produces a transcript of your answers and evaluates your performance for that listing.
• Most interviews take around 20 minutes, though some are shorter or longer.
• A free practice interview is available. It is not tied to any application, can be taken repeatedly, and is not shared with companies.
• You can retake the official interview up to three times across all applications that use it.
• A completed interview is reused — if another application needs the same one, your existing result carries across.
• Mercor states that the AI interview is its only screening method. There is no option to request a human interview instead.
• Mercor says it typically takes two to four weeks to hear whether you have progressed.

Source: Mercor talent documentation, checked 5 September 2026.

micro1

micro1’s Zara interview is also conversational and adaptive. Follow-up questions can probe the detail behind your answer, and technical roles can include additional practical or coding components.

micro1 works on a skill certification model rather than a single pass or fail. Its expert FAQ states that once you are certified, your profile becomes eligible for projects that match your verified skills, and that certifying additional skills widens the range of projects you can be considered for. micro1 is explicit that certification does not guarantee placement.

Source: micro1 expert FAQ, checked 5 September 2026.

Other assessments

Some platforms or individual projects use asynchronous written, audio, video or timed task-based assessments instead of a conversational interview. A role may also add a work sample, rubric exercise or specialist evaluation.

Platform formats checked: 5 September 2026. Always follow the instructions shown for the specific assessment you have been invited to complete.

How answers are actually scored

Platforms do not publish every detail of their scoring or review systems, and the mechanism can differ by role. Treat the points below as robust preparation principles — explicit reasoning, instruction-following, safety, scope, clarity and consistency — rather than as a claimed copy of a private scoring formula.

• Explicit reasoning — say why you reached a judgement, not only what the judgement is. Reasoning you leave implicit cannot be assessed.
• Instruction-following — answer the question that was actually asked, in the form it was asked for.
• Safety — name the point at which you would stop, escalate or seek senior input, rather than pressing on.
• Scope — stay inside your own competence and say plainly when something falls outside it.
• Clarity — structure the answer so someone outside your specialty can follow it.
• Consistency — reason the same way about the same problem, as below.

Keep your reasoning consistent

Give consistent reasoning when similar issues recur, and explain why your judgement changes when the facts change. Human or automated review can notice contradictions even when the platform does not disclose a formal cross-answer scoring rule.

Human review: when and how it happens

Human review varies by platform, role and project. Do not assume every interview is read by a person, and do not assume only borderline cases are looked at. What the two platforms document is that human judgement enters at a defined stage, rather than sitting over every interview.

• Mercor documents the AI interview as its only screening method. It says it retains the recording and an AI-generated transcript, and uses them to evaluate your candidacy and recommend your profile to potential employers — so the human judgement sits downstream, with the hiring team, rather than in the interview itself. You cannot obtain a copy of your own transcript.
• micro1 says human recruiters review profiles and match experts to opportunities against verified skills. Applications that meet the interview criteria move to a stage micro1 calls Hiring Manager Review. In a public AMA on 1 September 2026, a micro1 recruiter said the process combines the AI assessment with human judgement and that recruiters can access the interview transcript and recording when reviewing candidates. That is a recruiter’s description in a community forum, not published policy, so treat it as indicative.

The practical consequence is the same either way: write and speak as though a knowledgeable person may read your answer in full, because on both platforms someone may.

Prepare for the interview

• Clinical AI interview guide — how to make clinical reasoning, uncertainty, safety and escalation explicit.
• Practical interview setup — current Mercor and micro1 format notes, setup and practice.

Last reviewed: 7 September 2026 · Platform details checked: 5 September 2026