Live proctoring: the difference between being recorded and being there

There is a difference between someone who reviews what happened and someone who is there when it happens. Both can observe. Both can document. But only one of them can respond in the moment – can ask a question, read a situation, and make a judgement call before it becomes an incident. That difference sounds small. In the context of a high-stakes exam, it is anything but. .

When institutions look at online proctoring, they typically encounter three approaches.

The first is recorded proctoring. The exam session is captured on video and reviewed by a person afterwards. This can work for lower-stakes assessments, but it means that anything that happens during the exam – a technical issue, an ambiguous moment, a question from the candidate – only gets attention after the fact.

The second is AI-based proctoring. An algorithm monitors the session and flags behaviour it considers unusual – eye movements, background noise, screen activity. In many systems, these flags are visible to the candidate during the exam: pop-up warnings, alerts, on-screen notifications. For someone concentrating on a high-stakes test, that kind of interruption can be deeply unsettling.

The third is live proctoring. A real person is present throughout the session, in real time. They stay in the background – ensuring a calm, focused exam experience. But they are there: ready to step in if a situation needs clarification, if a candidate needs help, or if something genuinely requires attention.

We chose the third approach. Not because the others cannot work – but because we believe that when something truly important is at stake, being there is better than looking back.

When someone takes an exam with Certory, they are connected to a proctor within five minutes – often faster. No software download, no plugin, no admin rights required. A standard browser is all it takes.

The exam begins with a thorough onboarding: the proctor verifies the candidate's identity, checks the environment, and makes sure everything is set up for a smooth session. It is a process designed to give both sides confidence before the exam starts – and it is one of the things our candidates and institutions value most.

Once the exam is underway, the proctor stays in the background. No pop-ups, no automated warnings, no visible flagging. Two cameras – PC and mobile device – give them a complete view. Everything observed is documented with a timestamp. Notes and video are stored securely for 90 days. The proctor is there whenever something comes up – a technical problem, an unclear situation, behaviour that needs to be addressed. And when nothing comes up, they guarantee what matters most: a calm, focused exam experience. This is the opposite of AI-based systems that flag and interrupt throughout the session.

If anyone later asks what happened during that exam, there is an answer. Not a probability score – a report, compiled by someone who was there.

This is one of the most practical advantages of live supervision – and one that is easy to underestimate.

In a recorded or AI-flagged model, an ambiguous moment – a glance to the side, an unexpected noise, a brief disappearance from frame – gets reviewed after the exam is over. Without context, it can easily be marked as suspicious. In the worst case, it leads to a failed attempt for a candidate who did nothing wrong. A live proctor can address the same situation on the spot: ask a question, note the context, and resolve it there and then. That distinction can make the difference between a fair outcome and an unnecessary failure.

For institutions, this also means something else entirely: protection of intellectual property. If a candidate photographs the exam during the session, a live proctor can see it and intervene immediately. In a post-exam review, the same footage might reveal the same thing – but by then, the exam content has already been compromised. Real-time presence is the only model that allows institutions to act before the damage is done.

Live proctoring can trigger an immediate association with being watched. Someone observing you during an exam – that sounds like pressure. But the way live proctoring works at Certory is deliberately designed to minimise that pressure. The proctor stays in the background. There are no pop-up warnings. No on-screen alerts. The candidate knows someone is there – but they are not being interrupted, distracted, or confronted with automated suspicion flags while trying to concentrate.

A person can understand context in ways that an algorithm cannot. They can tell whether someone is nervously tapping on their desk or whether something genuinely unusual is happening. They can exercise judgement where judgement is appropriate.

86% of our candidates rate their experience as positive. Fewer than 5% need support during their exam. That is not a coincidence – it is what happens when supervision is designed around people, not around suspicion.

We are not writing this to dismiss other approaches. There are different models, and different exam formats have different needs. But for exams where something truly important is at stake, we believe it makes a difference whether someone is there or not.

A difference for the quality of supervision.

A difference for the trust in the result.

And a difference for the person who is giving it their best.