Agentic security is the billion-dollar challenge for some clever startup to solve
High time to stop kicking the security can down the road, investor tells The Reg
AI security, once an afterthought, is now a critical challenge for startups as increasingly capable agents find creative and unexpected ways to accomplish their objectives. Matt Hartman, chief strategy officer at Merlin Group, notes that security for agents, or "agentic security," is crucial as organizations rush to incorporate AI tools into their production environments.
Todd Graham, managing partner at Microsoft’s M12 venture fund, echoes this sentiment, stating that the security landscape for non-human actors, or agents, must be addressed rapidly.
This need for agentic security arises as companies move quickly to adopt AI tools and allow them to access critical data and applications. However, startups face the challenge of differentiating themselves in a rapidly evolving market where the cost of building technology has fallen, and the value shifts towards differentiated capabilities and marketability. Hartman emphasizes that startups must do more than just build an agentic AI security product, as the bar for differentiation continues to rise.
End-users are seeking agentic identity and governance products, similar to how SaaS companies created solutions like Okta for human identities. Graham believes that a company capable of providing comprehensive solutions for agentic identity will be the next unicorn in this space. The issue of securing non-human accounts, such as service accounts with high privileges and non-expiring passwords, also remains a significant challenge for security teams.
AI endpoint security, akin to CrowdStrike for AI, is another area ripe for innovative startups. Graham believes that this is a hot area that is still early enough for someone to build a quality solution, even though existing endpoint and antivirus vendors will likely develop products to fill this void.
Written by urgent.news from The Register Science's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.