Talview, headquartered in San Mateo, CA, has been awarded its first U.S. patent — Patent No. US 12,361,115 B1, granted on July 15, 2025, for Alvy, its Agentic AI proctoring solution. This milestone underscores Talview’s commitment to innovation and its impact on corporate hiring, education, and certification. Alvy is designed to both protect and assist, detecting sophisticated cheating tools across remote interviews and high-stakes online exams, while guiding candidates through a fair and transparent assessment experience.

Co-invented by Sanjoe Tom Jose (CEO) and Subramanian Kailasam (CTO), the patent protects a novel AI-based system and method powered by large language models (LLMs) for monitoring users during tasks such as interviews and exams. The system analyzes real-time media data to detect abnormal behavior, generate alerts, and — if thresholds are exceeded — terminate the session to prevent fraud, while also assisting candidates by clarifying instructions, reducing stress through empathetic AI interaction, and ensuring they understand what’s happening at every step.

“Most platforms are still patching legacy systems,” said Sanjoe Jose, CEO of Talview. “Alvy reimagines proctoring by bringing genuine artificial intelligence that can think, adapt, and outsmart even the most sophisticated cheating tools. This patent validates that maintaining trust in the AI age requires intelligence, not just observation.”

Unlike traditional systems that passively monitor, Alvy operates as an agentic system, capable of perceiving, deciding, and acting independently to protect the integrity of remote processes. It is specifically designed to detect and counteract AI-assisted cheating tools such as Cluely, ChatGPT misuse, and deepfake-driven impostors.

Key Capabilities:

• Autonomous Decision-Making: Real-time judgments without rigid rule-based systems.
• Deepfake & AI Detection: Identifies AI-generated content and synthetic media. 
• Behavioral Intelligence: Recognizes subtle patterns of external AI assistance.
• Contextual Understanding: LLM-powered for nuanced threat assessment.
• Adaptive Learning: Continuously evolves to counter emerging methods.
• Candidate Assistance: Provides real-time, context-aware assistance to test takers.

Performance Highlights:

• Detects 8x more suspicious activities than legacy AI proctoring.
• Recognizes 99% of events flagged by both human proctors.
• Reports 35% higher candidate satisfaction due to real-time and contextual assistance.

Proven Industry Impact:

Leading organizations have experienced Talview’s commitment to combining secure, scalable assessment technology with a human-centered candidate experience — the same principles that power Alvy’s patented Agentic AI.

• Best Buy: “We don’t just have a tech delivery mechanism. We have partners.”
• Cambridge Assessment: “We needed something that could be rolled out at scale globally. We have fantastic experience working with Talview.”
• Cognizant: “Everyone receives a uniform experience. We haven’t received feedback suggesting it feels mechanized or inhuman.”

The post Talview Patents Alvy, Agentic AI to Outsmart Cheating Tools first appeared on AI-Tech Park.

Origianl Creator: PR Newswire
Original Link: https://ai-techpark.com/talview-patents-alvy-agentic-ai-to-outsmart-cheating-tools/
Originally Posted: Tue, 12 Aug 2025 08:45:00 +0000