Proctoring tech just got a serious AI upgrade. Talview, the San Mateo–based talent assessment and proctoring company, has been awarded its first U.S. patent—No. US 12,361,115 B1—for Alvy, an agentic AI platform built to both protect the integrity of remote assessments and actually help candidates through them.
Co-invented by CEO Sanjoe Tom Jose and CTO Subramanian Kailasam, Alvy goes beyond the standard “watch and flag” approach. It actively monitors interviews and exams, detects cheating—even AI-assisted or deepfake-based attempts—and can intervene in real time. If suspicious behavior crosses certain thresholds, Alvy can terminate the session entirely.
Not Just Watching—Thinking
Traditional proctoring tools tend to be passive, relying on human review or rigid rule sets. Alvy operates more like a digital proctor with instincts, using large language models (LLMs) to analyze real-time audio and video data, interpret context, and make autonomous decisions.
Key capabilities include:
- Autonomous decision-making: Acts on the spot without waiting for human intervention.
- Deepfake & AI detection: Flags synthetic media and AI-generated responses.
- Behavioral intelligence: Spots subtle telltale patterns of external assistance.
- Contextual understanding: Uses LLMs for nuanced judgment calls.
- Adaptive learning: Evolves to counter emerging cheating techniques.
- Candidate assistance: Offers real-time clarifications, reducing test anxiety.
This dual role—security and support—is part of Talview’s pitch that AI proctoring should be as empathetic as it is vigilant.
Performance That Moves the Needle
Talview claims Alvy detects suspicious activity at eight times the rate of legacy AI systems, matches 99% of human proctor accuracy, and boosts candidate satisfaction scores by 35%. That’s not just a security upgrade—it’s a user experience play in a market where candidates often complain about cold, impersonal testing software.
“Most platforms are still patching legacy systems,” said Jose. “Alvy reimagines proctoring by bringing genuine AI that can think, adapt, and outsmart even the most sophisticated cheating tools.”
Why This Matters
Remote hiring and certification aren’t going away, and neither are the tools designed to game them. AI-assisted cheating—think ChatGPT misuse, AI-written answers, and deepfake impostors—is becoming a serious threat to credibility in hiring and education. By securing a patent, Talview is not just defending its IP—it’s staking a claim in a growing arms race between assessment integrity and AI-assisted fraud.
Early adopters like Best Buy, Cambridge Assessment, and Cognizant have already deployed Talview’s broader platform, praising both its scalability and its ability to deliver a uniform, human-centered experience at scale.
With Alvy now patented, Talview has a protected technological edge—and in the high-stakes world of exams and interviews, that edge might make the difference between trust and doubt.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
