Weekly roundup New infosec products debut July 17

Polygraf AI introduced Meeting Guard this week, a solution that aims to secure enterprise video calls by spotting AI‑generated fraud in real time. The system joins a virtual meeting as a visible participant and runs security checks for every attendee, according to the company’s announcement.

How the tool works during a live session

Unlike many security products that rely on post‑call analysis, this platform claims to deliver “near‑real‑time” insights, meaning the detection happens while the meeting is still in progress. The company says the service does not send any data outside the organization, positioning it as a compliance‑friendly option for both corporate and government users. By keeping all processing on‑premise or within a private cloud enclave, Meeting Guard aligns with strict data‑sovereignty requirements that many regulated sectors enforce.

Broader implications for meeting security

Deepfake technology has become a growing concern for organizations that depend on virtual collaboration. By embedding a detection agent directly into the meeting interface, Polygraf AI hopes to reduce the window of opportunity for malicious actors who might otherwise exploit synthetic voices or video to impersonate executives. The presence of a dedicated participant that continuously validates authenticity creates a deterrent effect, as potential attackers must now overcome an active, automated checkpoint rather than a passive recording.

Related: Boring Normalcy Aids Finance Phishing

From a broader perspective, the rise of tools like Meeting Guard reflects a shift toward embedding security controls into everyday productivity software. As remote work remains prevalent, enterprises are looking for ways to protect communications without adding cumbersome steps that could hinder productivity. Real‑time verification could become a standard feature, much like encryption is for email. This trend mirrors developments in other segments of the security market, where solutions such as Nudge Security’s automated OAuth‑grant analysis and Cloudflare’s continuous behavioral bot detection are moving the point of protection closer to the user interaction layer.

The product also includes an AI‑driven notetaker that creates meeting minutes without transmitting data to external servers. This dual function—security and documentation—targets organizations that need to keep track of discussions while maintaining strict data handling policies. The notetaker parses spoken dialogue, extracts action items, and formats a concise summary that can be stored in internal knowledge bases, all while adhering to the same on‑site processing constraints that govern the fraud‑detection component.

Polygraf AI says the system is ready for deployment in both private and public sectors, though it did not disclose pricing or specific integration details. The announcement did not include independent testing results, so observers will likely wait for third‑party evaluations before judging its effectiveness. Early adopters will be watching how the tool interoperates with existing video‑conferencing platforms and whether its presence as a visible participant influences user behavior or meeting trends.