Microsoft warns ungoverned AI agents as corporate “double agents,” launches $99/month

Microsoft’s rollout of Agent 365 and Microsoft 365 Enterprise 7 arrives at a moment when “shadow AI” is proliferating faster than any other security concern. VentureBeat notes that the number of un‑governed AI agents doubles roughly every 18 months, creating blind spots that traditional security operation centers (SOCs) simply cannot see. By bundling a centralized control plane with the company’s most advanced security stack, Microsoft hopes to turn that exponential growth curve into a manageable, observable surface.

Agent 365, priced at $15 per user per month, functions as the “control plane for agents,” giving IT, security, and business teams a single dashboard to monitor an agent’s visibility, permissions, and risk profile. According to ZDNet, the tool lets administrators spot risky agents before they become insider threats, flagging anomalous behavior and surfacing unauthorized data accesses in real time. The platform also integrates with Microsoft 365 Copilot’s Wave 3, which adds model diversity from OpenAI and Anthropic, allowing enterprises to enforce policy across a broader set of generative models.

The “Frontier Worker Suite,” officially named Microsoft 365 Enterprise 7, packages Agent 365, Copilot, and Microsoft’s latest security suite into a $99‑per‑user‑per‑month license. Vasu Jakkal, corporate vice president of Microsoft Security, told VentureBeat that the timing is deliberate: “These agents are no longer experimental. We’re seeing them deeply embedded in organizations… the visibility gap creates business risk.” The suite is positioned as a one‑stop solution for enterprises that have already let AI agents into their operational fabric but lack the tools to audit or constrain them.

Adoption metrics underscore the urgency. VentureBeat reports that more than 80 % of Fortune 500 companies now use AI agents, yet nearly a third of those deployments are unsanctioned. That mismatch between usage and governance fuels the “double‑agent” scenario Microsoft warns about—agents that, while performing legitimate tasks, can be hijacked or repurposed by malicious actors to exfiltrate data or sabotage systems. By providing a unified view of agent activity, Microsoft aims to prevent such agents from acting as covert insiders.

Analysts have long warned that the rise of “shadow AI” will outpace existing controls. A recent VentureBeat piece highlighted how 79 % of generative‑AI attacks now bypass legacy defenses, slipping through unnoticed because they originate from trusted internal agents. Microsoft’s new offerings are a direct response to that trend, promising to close the gap between rapid AI deployment and the slower evolution of security tooling. If the suite delivers on its promise, enterprises may finally have a way to keep their AI assistants on a leash—turning potential double agents into transparent, governed collaborators.