Constructor Proctoring modes: finding the right fit for every assessment
23 May, 2025
Not all online exams are created equal—and neither should the way we secure them. While some assessments require the flexibility of fully automated AI proctoring, others demand human supervision and multi-layered controls. The key is to balance integrity, user experience, and resource efficiency.
That’s where Constructor Proctor’s flexible proctoring modes come in, giving institutions, ministries, and organizations the freedom to select the right level of supervision based on exam stakes and context.
Mode 1: Automated proctoring (Soft Mode)
This is your AI-first, lightweight approach.
Fully automated process where AI monitors video, audio, and screen activity, calculates the proctoring score, and flags violations.
Candidates are identified automatically, and conclusions are made purely by AI—making it ideal for quizzes, regular exams, or large-scale practice tests where efficiency matters more than strict supervision.
Mode 2: Post-review proctoring (Medium Mode)
Best of both worlds—automation combined with human review.
The AI monitors during the exam, but after the session, human proctors review the video recordings, analyze the flagged events, and confirm the final results.
Recommended for certification, entrance exams, and regulated industry assessments, where the review adds an extra layer of credibility.
Mode 3: Live proctoring (Restricted Mode)
The most secure mode, combining AI with real-time human monitoring.
How critical is this exam?
What’s the risk if someone cheats?
How much human involvement do I want?
AI proctoring capabilities across all modes
No matter the mode, Constructor Proctor leverages its proprietary AI toolkit to ensure accuracy and fairness:
Face recognition
Gaze direction
Identity verification
Speech detection
Absence detection
Unauthorized person detection
Use of earphones or smartphone
Tab change alerts
Which mode is right for you?
Ask yourself:
How critical is this exam?
What’s the risk if someone cheats?
How much human involvement do I want?
