
Tony Ferrigno
Founder & Principal, AIT Advisory Group
MSP Intelligence Brief · April 2026
Microsoft Agent 365 goes generally available May 1. Agent governance now has a dedicated platform, and if you're an MSP, this is the conversation your clients need to have right now.
For the past two years, AI agents have been accumulating inside enterprise environments: untracked, ungoverned, and increasingly dangerous. IDC projects 1.3 billion agents in circulation by 2028. Today, 80% of the Fortune 500 are already running Microsoft agents. And most of them have no real visibility into what those agents are doing, who authorized them, or what data they're touching.
Ungoverned agents accumulating risk inside environments MSPs are supposed to manage is a managed services problem. And Microsoft just shipped the solution.
"The best way to manage agents is to extend the infrastructure you use for managing users." - Charles Lamanna, President, Business Apps & Agents, Microsoft
Microsoft Agent 365 is exactly what it sounds like, Microsoft 365, but for your agents. It's the unified governance, identity, security, and management layer that sits above your entire fleet of AI agents, whether they were built on Microsoft platforms, deployed from partner ecosystems, or discovered lurking as shadow agents you didn't even know existed.
It works the same way your RMM does, but for AI agents instead of endpoints and users. Every agent needs an identity, an access policy, and an audit trail before it touches your client's environment. Agent 365 enforces that.
500K+ agents Microsoft manages internally with Agent 365 · 65K+ daily agent responses at Microsoft in just 28 days · 1.3B agents IDC projects in circulation by 2028
This isn't a dashboard bolt-on. Agent 365 extends the full Microsoft security and management stack, including Defender, Entra, Purview, and M365 Admin Center, to cover AI agents as first-class managed entities. Here's what that means in practice:
Agent Registry: Complete inventory of all agents including shadow agents you didn't deploy. If it has an Entra Agent ID, it's in the registry.
IT-Defined Guardrails: Set who can create, onboard, and manage agents. Assign sponsors for continuous supervision. Standard policy templates included.
Access Governance: Enforce least-privilege access. Agents only get access to apps and resources required to complete their specific tasks.
Runtime Defense: AI-powered blocking of prompt injection attacks, malicious traffic, and real-time data exfiltration prevention.
Compliance & Audit: Full logging, reporting, and content safety controls to detect, retain, and investigate every agent interaction.
Performance Measurement: Track agent speed, quality, and ROI. Know which agents are delivering value and which ones are just burning compute.
Agent 365 governs agents across your entire agent environment, not just Microsoft-built ones. Your clients aren't running a single-vendor AI stack, and Agent 365 doesn't assume they are.
Partner agents from Adobe, ServiceNow, Workday, NVIDIA, Databricks, Manus, Genspark, Glean, n8n, and Cognition are all supported with more being added continuously. This is a governance platform, not a vendor lock-in play.
Agent 365 goes generally available on May 1, 2026. The pricing is licensed per human user, not per agent. Agents acting on behalf of a licensed user are covered under that seat.
Microsoft 365 E7, the Frontier Suite, is the strategic bundle to watch. It's the first Microsoft SKU built specifically for the agentic AI era, and for clients already deep in the M365 stack, the per-seat math works strongly in its favor compared to buying components separately.
Managed Intelligence is an industry framework for transforming how organizations adopt, govern, and operationalize AI at scale. As a leading authority on this framework and the voice behind the industry-defining tagline "We Put AI in IT™," Tony Ferrigno brings 35+ years of MSP executive leadership to the work of modernizing and operationalizing Managed Intelligence for the agentic AI era.
Microsoft Agent 365 is not just a new product to resell. It is the technical foundation that makes Managed Intelligence executable for MSPs, providing the governance layer, the identity controls, the security posture, and the performance visibility that transforms AI agents from an experiment into a managed, accountable, revenue-generating service.
This is what "We Put AI in IT™" means in practice. Not AI as a concept. AI as infrastructure. AI as a managed service. AI as a competitive advantage your clients can measure.
The agent governance opportunity isn't coming; it's already here. Every M365 client you have is a potential Agent 365 conversation. Every organization running Copilot is accumulating agents they may not be tracking. Here's where to focus first:
The question isn't whether your clients will adopt AI agents. They already have. The question is whether those agents are governed, secure, and delivering measurable ROI, or running unsupervised, accumulating risk, and quietly becoming your next incident report.
Microsoft Agent 365 is the infrastructure layer that closes that gap. And Managed Intelligence, operationalized through the "We Put AI in IT™" lens, is the strategic framework that turns that infrastructure into a differentiated, scalable managed service practice.
MSPs who move now get a real first-mover advantage in their markets. Those who wait will spend the next two years trying to close a gap that keeps widening. The control plane for agents is here. The only question is who's managing it for your clients: you, or someone else.