Adopt Microsoft Agent 365 as our agent control plane?

Microsoft Agent 365 costs $15 per user monthly and splits governance between Foundry and Agent 365, but autonomous agent loops can trigger a 5x cost increase and introduce data leak vulnerabilities.

2026-07-10 · Counsel verdict · AIssential

The question

Every M365 tenant now has to decide whether to standardize agent governance on Microsoft Agent 365 — paying $15/user/mo for a control plane we don't operate — or roll our own agent registry, identity, and access-control infrastructure.

Counsel's position

Standardize on Microsoft Agent 365 to rapidly secure and govern AI agent proliferation, accepting its cost and control limitations.

Verdict

The verdict: Standardize on Microsoft Agent 365 to rapidly secure and govern AI agent proliferation, accepting its cost and control limitations.

Agent 365 costs $15 per user monthly and discovers local shadow agents

Given your decision on agent governance, Agent 365 provides a centralized registry and policy engine for both cloud and local agents.

• Microsoft takes Agent 365 out of preview as shadow AI becomes an enterprise threat

Microsoft splits agent governance between Foundry for builders and Agent 365 for IT

Standardizing on Microsoft's ecosystem means adopting a dual-control plane strategy where Agent 365 handles organizational governance while Foundry handles development.

• Foundry Control Plane and Agent 365: Two Control Planes Walk Into an Enterprise

Usage-based agent billing triggered a 5x cost increase in one month

While evaluating managed vs. self-hosted infrastructure, consider that autonomous agent loops can rapidly drain shared credit pools under usage-based billing.

• AI Agent Costs Are Becoming an Engineering Crisis

Managed governance platforms ship primitives but leave policy authorship to you

Whether you buy Agent 365 or build your own registry, your institution remains responsible for defining the policies those controls enforce.

• The Agentic AI Governance Stack Got Built This Year - Here Is the Part No Vendor Can Ship

Microsoft's Autopilot agents introduce prompt injection and data leak vulnerabilities

Relying on Microsoft's managed agents requires accepting the security risks inherent in their underlying OpenClaw architecture.

• No longer just a Copilot, Microsoft's AI wants to take the wheel

Read another verdict

• Buy a tool for this process, or build around our own knowledge?
• Our documents are a mess. Clean them up before AI, or after?
• How do we measure the return on an AI workflow — and what baseline is honest?
• Our best people's know-how isn't written down — can AI even use it?
• Automate this workflow, or redesign it before we automate?
• Which process should we point AI at first?
• Our AI pilot works but nobody uses it — fix the workflow or kill it?
• Rent AI from a vendor, or run your own?

Get Counsel for your own decisions →