Microsoft says ungoverned AI agents could become corporate ‘double agents.’ Its fix costs $99 a month.

Microsoft's recent unveiling of Agent 365 and Microsoft 365 Enterprise 7 highlights a critical moment in AI governance. With AI agents rapidly integrating into major corporations, the tools to manage these entities lag behind. This gap poses real risks, particularly as nearly a third of AI agents operate without IT approval.

The term 'double agents' aptly captures the potential threat. Imagine AI tools designed to serve a company being manipulated to act against its interests. While Microsoft hasn’t yet reported widespread exploitation, the company's research indicates a looming danger that organizations must confront.

Agent 365 seeks to address these concerns by acting as a centralized control system for monitoring and securing AI agents. Its features reflect a shift in thinking about AI, treating these agents like employees and extending zero-trust principles to them. This approach could redefine how companies view not just AI, but also their overall security posture.

However, the $99 per user monthly fee for the Enterprise 7 package raises questions about affordability and commitment. Companies might hesitate to invest in governance solutions when the tools to create AI agents are readily available and relatively inexpensive.

Ultimately, the challenge lies in whether businesses can adapt quickly enough to govern their AI agents effectively. The future of work won't just revolve around smarter technology but trusted technology. For the 29% of enterprise agents operating without oversight, time is running out to establish that trust.