Does Mercer Mettl detect cheating?

You have an assessment invitation and want to know what it can see before you start. This section lists the checks Mercer Mettl documents in its own help center, so you can prepare instead of guessing. They come as three proctoring modes plus individual settings the organizer chooses.

• Auto proctoring: an AI system watches the live video and audio feed and highlights suspicious activity.
• Live proctoring: a human proctor watches your video feed and your test screen in real time.
• Record and review: the administrator can access the full video recording of your attempt afterwards.

On top of the mode, individual settings can add snapshots of your screen and camera every 60 seconds, a full webcam recording, audio proctoring that detects human speech, a photo of your ID card and face before you begin, and a human authorizer who admits or blocks each candidate. Questions and answer options can also be randomized for each candidate.

Mettl's help center lists the AI flags: face not visible, another person present, away from the test window, looking away, face partially visible, candidate not present, impersonation suspected, and mobile phone detected. Each is a timestamped event in the report, not a verdict.

Those flags, together with how the test finished, feed a credibility index rated High, Medium or Low. Employers can filter results by it and see it in candidate reports; Mettl labels the feature as beta. A low rating tells a reviewer where to look in the recording, and the organization running the test decides what it means.

A commerce graduate taking a Mettl aptitude test at home had a relative walk through the room halfway through, which produced an "another person" flag. The recording showed the context in a few seconds, which is exactly why a quiet, closed room is the best preparation you can make. Breakdowns for other assessment vendors are on the detectability topic hub.

When a test uses the Mettl Secure Browser, your computer goes into a lockdown mode that prohibits third-party applications from running in the background. Copy-paste and print-screen are blocked.

Separately, a setting called browsing tolerance counts every time you navigate away from the test window. The organizer can show warnings, and once you pass the set limit the test is terminated. Of all the settings, this is the one most likely to end an attempt on its own, so close other apps and silence notifications before you start.

Innocent behavior gets flagged too. With audio proctoring on, reading a question aloud to yourself registers as speech, and a family member speaking in the next room can as well. Read silently, test in a closed room, and run the system check early so camera or microphone problems surface before the timer starts.

SubcueAI assists live, spoken interviews. The native macOS and Windows app captures system audio and your microphone and shows suggestions in a floating local overlay; the browser extension's Side Panel does live assist for a call in a Chromium tab, capturing only that tab's audio. No meeting bot joins the call, and nothing is injected into the meeting page.

A Mercer Mettl assessment is a monitored test, not a conversation, and SubcueAI does not read or answer it. A proctored, camera-watched test with a lockdown browser is out of scope, which is a limit of what any tool should do on someone else's assessment, not a setting to change. What SubcueAI records during a live interview is set out on the security page.

The stage worth preparing for is the interview the test unlocks. A timed run on the mock interview page rehearses saying your answers out loud, which is what that round tests.