Assessment integrity changed the moment generative AI became free, fast, and invisible. Institutions and employers that once worried about a hidden phone now face autonomous AI agents that read questions and generate answers in real time, virtual machines that conceal a second computer, and tiny cameras hidden in pens and glasses. Choosing the right proctoring platform is no longer a checkbox — it decides whether your credentials and hiring decisions can survive scrutiny. This guide compares two options: Proctorly, an AI-powered assessment integrity platform built for the AI era, and Mercer | Mettl, a broad, well-established talent-assessment and examination suite.
Table of Contents
Platform Overview
Proctorly is an AI-powered assessment integrity platform with a clear mission: stop AI-assisted cheating before it costs you the exam. It detects AI agents, cheatbots, and virtual machines in real time, then backs every flag with human-reviewed evidence so results hold up under audit. Its detection reaches past the browser into the operating system, and it verifies identity with liveness and facial checks — without storing biometric templates. Proctorly aligns with GDPR and India’s DPDPA and extends the same model to interviews and technical assessments.
Mercer | Mettl is a cloud-based talent-assessment and examination platform operated by Mercer, serving thousands of clients worldwide. Its strength is breadth — psychometric, aptitude, technical, and behavioral assessments, coding tests, and hiring workflows — plus AI-assisted remote proctoring and very large delivery scale. It is ISO 27001 and GDPR compliant and integrates with major LMS, HRMS, and ATS platforms.
In short: Mercer | Mettl leads on catalog breadth and raw scale, while Proctorly leads on the specific detection layers that define cheating in 2026, paired with human-validated evidence.
Why AI-Era Cheating Is Harder to Catch
The most dangerous cheating today happens beneath the surface — at the operating-system level, or just outside the webcam’s frame. A candidate can sit perfectly still while an AI agent reads their questions and feeds answers through a hidden overlay, or while a helper quietly takes control of the screen from elsewhere. A camera pointed at a face can’t see any of it. The new toolkit includes AI agents and cheatbots that generate answers, virtual machines that hide the exam environment, remote-desktop tools that hand over the session, and hidden browser extensions and overlays that leak questions or paint answers on screen. A platform that can’t see these can’t stop them — and this is where a purpose-built integrity engine separates itself from general-purpose proctoring.
Hidden Devices Used During Online Exams
Beyond software, candidates increasingly turn to tiny, concealed hardware that a standard webcam will never catch:
- Pen cameras — a working pen with a pinhole lens that photographs or streams the exam to a helper.
- Smart Glasses with Display — ordinary-looking eyewear with a built-in camera and, a display.
- Micro and wireless earpieces — near-invisible in-ear devices that relay answers by voice.
- Smartwatches and secondary phones — kept just out of frame to receive help.
These devices operate outside the examination system, making them one of the biggest blind spots for traditional online proctoring. Smart glasses, wireless earbuds, smartwatches, AI-enabled wearables, and other connected devices can provide candidates with real-time assistance without ever appearing on the exam screen. As these technologies become more common, relying on a webcam alone is no longer enough.
Proctorly addresses this new generation of threats through its System Integrity Agent (SIA), which extends assessment integrity beyond the examination system to help identify unauthorized connected devices that may be used to facilitate cheating. Since many modern smart wearables and wireless devices operate independently of the candidate’s exam computer, they often remain invisible to traditional browser-based proctoring. SIA is designed to close this visibility gap, helping institutions stay ahead of emerging cheating techniques. Combined with identity verification, behavioural analysis, environment validation, and human-reviewed evidence, Proctorly helps institutions stay ahead of emerging AI-assisted cheating techniques while ensuring every critical decision is backed by evidence rather than automation alone.
Because modern assessment integrity requires greater visibility, it must also be matched with greater responsibility. Proctorly is designed around principles of candidate transparency, data minimisation, defined retention policies, and support for privacy regulations such as GDPR and India’s Digital Personal Data Protection Act (DPDPA).
Feature Comparison at a Glance
| Capability | Proctorly | Mercer | Mettl |
|---|---|---|
| Identity verification | Liveness + facial checks; no biometric templates stored | AI-based authentication and ID checks |
| Environment / system inspection | OS-level checks for VMs, remote-access tools, overlays | Secure browser lockdown |
| AI agent / cheatbot detection | Yes — core differentiator | Public documentation does not prominently describe dedicated AI-agent detection. |
| Virtual machine detection | Yes | Public documentation does not emphasize dedicated VM detection. |
| Hidden-device coverage | Supports secondary-camera workflows where enabled by institutional policy. | Standard webcam proctoring |
| Human review | 24/7 trained review of critical incidents | Available; varies by plan |
| Audit-ready evidence | Timestamps, event history, screenshots, review logs | Reporting and analytics dashboards |
| Interview / hiring integrity | Live AI monitoring + resume screening + ATS | Virtual interview and coding tools |
| Enterprise scalability | Scales with LMS/SSO; integrity-focused | Very high-volume delivery |
The pattern is clear: Mercer | Mettl wins on breadth and scale, Proctorly wins on depth of AI-threat detection, hidden-device coverage, and human-validated evidence.
What Strong AI Proctoring Actually Does
Rather than chasing any one product, judge a platform on five things:
- Verifies identity with liveness and facial checks, so the right person is taking the exam — no proxies.
- Validates the environment at the system level, catching virtual machines, remote-access tools, and hidden monitors.
- Detects AI assistance — agents, cheatbots, extensions, and overlays — in real time.
- Puts humans in the loop, so trained reviewers confirm serious flags and honest candidates aren’t wrongly accused.
- Produces audit-ready evidence — timestamps, screenshots, and review logs that hold up if a result is challenged.
On these criteria, Proctorly is engineered around all five as its core purpose, while Mercer | Mettl delivers them as part of a much broader assessment suite.
Why Human Review and Evidence Matter Most
Fully automated verdicts have a catch: false positives. A lone AI flag can wrongly accuse an honest candidate over a network hiccup or an innocent movement — and nothing erodes trust in an exam program faster than punishing the innocent. The fix is a human-in-the-loop model: AI catches issues at scale, trained reviewers judge the ones that matter, and every decision comes with a documented trail. That turns integrity from a claim into something you can actually prove — protecting the institution and the candidate alike. Proctorly makes this human validation and evidence trail a standard part of every critical flag.
Best Fit by Use Case
For universities defending degree credibility and certification bodies protecting the value of a credential, Proctorly’s deep detection plus audit-ready evidence is the stronger fit. For recruitment and hiring, Proctorly unifies identity verification, live interview monitoring, resume screening, and evidence reporting in one integrity-first workflow. Mercer | Mettl is compelling where a broad assessment catalog or extreme concurrent scale — national exams, global hiring drives — leads the requirements, or where a team is already standardized on the wider Mercer ecosystem.
For pure AI-era assessment integrity detecting AI agents, cheatbots, virtual machines, and hidden devices, then proving each flag with human-reviewed evidence.
Proctorly is the stronger platform. Mercer | Mettl remains a solid choice when breadth of assessment types and sheer scale matter most. If your central concern is stopping AI-assisted cheating and defending every result, Proctorly’s focused, evidence-first approach is the better match.
Book a Demo
See how Proctorly detects AI-assisted cheating, validates incidents with human reviewers, and delivers audit-ready evidence for trusted assessments and secure hiring. Explore the Assessment Integrity Platform, AI interview proctoring, and the System Integrity Agent, or read about remote desktop cheating prevention. Questions? Start with the FAQ or visit Proctorly to book a demo.
What’s the biggest difference between Proctorly and Mercer | Mettl?
Focus. Proctorly is purpose-built to detect AI-assisted cheating and hidden devices, and to prove each flag with human-reviewed evidence. Mercer | Mettl is a broad assessment suite with established proctoring and very large scale.
How do you catch pen or glasses cameras and earpieces?
Not with one camera. It takes wider coverage — side-view angles, a pre-exam room scan, and audio and behavioral analysis — to reveal concealed devices and the signs they’re being used.
Why keep a human in the loop if AI already detects cheating?
Because automated flags produce false positives. Trained reviewers confirm the serious ones so honest candidates aren’t wrongly penalized, and their review becomes part of the evidence.
Which is better for university exams?
For stopping AI-assisted cheating and defending degree credibility, Proctorly is generally stronger; Mercer | Mettl suits institutions needing an extensive catalog or extreme exam-day scale.
