Cybersecurity startup AegisAI Inc. has launched with $13 million in seed funding co-led by Accel and Foundation Capital to fix email security using artificial intelligence in the face of increasingly sophisticated attacks.

Founded by Chief Executive Cy Khormaee and Chief Technology Officer Ryan Luo, former Google LLC security executives, the startup focuses on using agentic AI that scans incoming emails and eliminates malicious content before it reaches the inbox. The company’s autonomous AI agents work together to detect and eliminate malware, phishing attempts and business email compromise attacks.

The company also said Wednesday that its solution reduces false positives up to 90% compared with traditional rules-based solutions. “We’ve spent almost a decade each protecting billions of users at Google; we’ve seen firsthand how enterprise email defenses are falling behind,” said Khormaee.

He added that the advent of large language models has led to an explosion in AI-driven attacks that are far more sophisticated than in past years, overwhelming the capabilities of most tools and proving more capable of tricking users.

AI has allowed attackers to refine their attack methods, which enables them to create email content that looks and feels more similar to content users might click on.

A report from cloud cybersecurity firm Barracuda Networks Inc. in 2024 showed that email-based attacks, such as scamming and phishing, made up 86% of social engineering attacks. In addition, impersonation rose by 10.6% and conversation hijacking incidents increased by 70%. A CrowdStrike Holdings Inc. report from 2024 showed that AI-generated messages have a greater click-through rate of 54% over human-written ones at 12%.

AegisAI’s platform integrates with Microsoft 365 and Google Workspace, where its AI agents go to work. The agents continuously learn from real-world threat intelligence and share data on malicious behavior across organizations, which allows for rapid detection of emerging threats.

The core platform includes real-time analysis of all messages, a zero-configuration setup that requires minimal maintenance and intelligent false-positive suppression. “We don’t believe in creating more alerts — we believe in creating better security outcomes,” said Luo.

AegisAI enters into a market filled with startups and old industry leaders dealing with the rising generative AI-led email threats. This year, Barracuda Networks launched its own multimodal AI solution for providing context-aware defense against emerging email attacks. Other growing email security providers include Sublime Security Inc. and StrongestLayer Inc.

Images: SiliconANGLE/DALL-E 3, AegisAI