December 5, 2025

Executive Summary: A surge in AI‑powered impersonation scams has created a lucrative opportunity for cyber‑security startups. Imper.ai, a Tel Aviv‑based company co‑founded by veteran technologist Noam Awadish, announced a $28 million funding round to build tools that detect deepfake voice and video attacks in real time. By analysing device telemetry and network metadata rather than the media itself, imper.ai aims to protect companies from sophisticated fraud attempts targeting videoconferencing platforms, customer service lines and collaboration tools.

Full Article:

Deepfake videos and synthetic voices have moved from novelty curiosities to dangerous weapons in the hands of scammers. Over the past year, Fortune reports that impersonation scams jumped 148 percent, costing global enterprises billions. The attacks range from fake CEOs demanding wire transfers over Zoom to AI‑generated customer‑service agents asking for credentials. Even large corporations like Jaguar Land Rover have fallen victim, with one deepfake scheme causing a $1.5 billion write‑down. It’s a growing threat that entrepreneurs and investors can no longer ignore.

Enter imper.ai, a startup that believes the best way to spot an AI imposter is not by examining its face or voice but by reading the invisible fingerprints it leaves behind. CEO Noam Awadish, who previously led teams at Mobileye and served in Israel’s elite 8200 cyber unit, argues that deepfake detection is often a losing game. Attackers can tweak an image or audio clip until it slips past filters. Instead, imper.ai focuses on metadata, device IDs, IP addresses, browser characteristics, and communication patterns, to determine if a call or chat session is coming from a legitimate device at the expected location. By triangulating this data across Zoom, Microsoft Teams, Slack and other collaboration tools, the company claims it can flag impostors in real time.

The company’s $28 million funding round was co‑led by Redpoint Ventures and Battery Ventures with participation from Greylock and F2 Capital. Investors are betting that imper.ai’s platform will become a “universal identity layer” for enterprise communications. When a user joins a video call, the software inspects subtle signals, microphone latency, Wi‑Fi jitter patterns, even gyroscope data, matching them against a known profile of the legitimate user’s devices. If anomalies are detected, the call is halted and an alert is sent to security teams. This approach is reminiscent of fraud detection in fintech, where minute behavioral signatures reveal stolen credit cards. But applying it to the sprawling universe of digital collaboration is an ambitious challenge.

Awadish says the funding will double imper.ai’s research team and triple its go‑to‑market organization. The company is already working with banks, law firms and healthcare providers who are prime targets for social‑engineering attacks. Rather than selling to end users, imper.ai integrates with enterprise platforms via APIs, making the technology invisible to employees. Pricing is based on the number of protected identities and the volume of authentication events. As AI deepfakes proliferate, the addressable market for this type of protection will only grow.

For entrepreneurs, the lesson is clear: as AI tools democratize creativity and communication, they simultaneously empower criminals. Businesses that handle sensitive data must invest in new forms of identity verification. Venture capital’s interest in imper.ai signals broader investor awareness of this risk. It also hints at opportunities for startups building adjacent solutions, from AI watermarking to behavioral biometrics. The arms race between deepfake creators and defenders is just beginning, and whoever cracks reliable, low‑friction authentication could build the next cybersecurity unicorn.

Subscribe to our newsletter to stay up to date with the latest breakthroughs in AI, Business, Technology, and Mindset.