Skip to content

PRODUCT WALKTHROUGH · VERIFIED DEMO ATTEMPT

What an online proctoring report tells a reviewer

An exam score, an AI trust index and a reviewer decision answer different questions. This walkthrough uses an existing completed attempt from the demo workspace to explain how to read the results, events and available evidence together.

An anonymised editorial walkthrough of a real demo attempt, rechecked on 17 November 2026. The walkthrough layout is adapted. A separate downloadable PDF preserves the original Russian interface with personal data removed. This is not a customer performance case study.

Walk through a report for your exam

Open the anonymised report PDF

Three pages: an English reading guide and the original two-page export from the demo workspace. The report interface remains in Russian. The name, photos, screen thumbnails, timestamps and programme name are removed; scores, events and the system-generated conclusion are unchanged.

Export checked on 17 November 2026. 3 pages · under 1 MB. The PDF does not include playable recordings or expanded analysis panels. Read the first-page notes on the limits of AI conclusions.

Three different results

25Points for answers
51AI trust index · elevated risk
Not selectedReviewer decision

Logged event types: full-screen exit; background noise; lost focus; external-camera focus. Environment data is unavailable.

Read the academic result first

The Results tab shows the questions, the pattern of submitted answers and an academic summary. The example attempt has 25 points. This describes assessed answers; it does not establish whether all exam rules were followed. A reviewer should use the programme’s marking rules and examine the underlying answers when a result is disputed.

Use the AI report to prioritise inspection

The example displays an AI trust index of 51 and an elevated-risk label. It also exposes sections for answer intensity, log activity and speech-transcription analysis. The number is not presented here as a probability of cheating or a measured detection rate. Read the underlying events and available evidence before reaching a conclusion. Different programmes can require different review rules.

Connect each event with its context

The action log in this attempt contains a full-screen exit, background-noise alerts, lost focus and an external-camera focus event. These event types are shown in the summary without participant names or timestamps. Noise may have an ordinary explanation; a focus event may need a technical check. Compare the event with permitted resources, instructions and the available recording. Do not treat the event label alone as proof of misconduct.

Check what evidence is actually available

The Control tab exposes identity-photo, main-camera, second-camera and screen areas in the inspected workspace. Availability depends on the configured exam and the material retained for the attempt. In this example, the environment section explicitly says there is no usable environment data and that no environment analysis was performed. That limitation should remain visible during review. Missing evidence must not be presented as a successful check.

Check the scope of each AI conclusion

A conclusion about the audio applies to the audio available for analysis. The absence of a detected audio violation cannot establish that no outside help was used through another channel. Check the source, the time covered and any missing material for each statement. Compare the summary with the event log and available recordings. If the sources disagree or do not support a conclusion, record the uncertainty for the authorised reviewer rather than treating the wording as a confirmed finding.

Keep the reviewer’s decision explicit

The interface asks whether the reviewer agrees with the AI report and exposes two decision controls. Neither was selected in the inspected example, and we did not change the attempt. An authorised reviewer must consider the evidence and the institution’s rules. Define escalation and appeal procedures before the exam. An AI summary is an input to that process, not a substitute for it.

What to verify in your own demonstration

Bring a representative assignment and agree permitted resources, device configuration and reviewer roles. Complete a safe test attempt, then inspect the academic score, event detail, available recordings and missing-data messages. Check who can access or export materials and agree retention in the contract. For education, certification and HSE knowledge assessments, the final academic or qualification decision remains with the authorised organisation.

Explore the related workflow

YOUR PROGRAMME. YOUR NEXT STEP.

Walk through a report for your exam

An exam score, an AI trust index and a reviewer decision answer different questions. This walkthrough uses an existing completed attempt from the demo workspace to explain how to read the results, events and available evidence together.

Request a demo