DIRECT ANSWER
Which AI agents belong to Microsoft?
Microsoft ships agentic products across multiple enterprise stacks. Researcher and Analyst serve Microsoft 365 Copilot workflows; Security Copilot Agents target security operations. Copilot Studio, Foundry Agent Service and Agent Framework occupy separate builder and runtime layers.
This table links existing agent profiles. These are not duplicate copies and trust claims are never inherited automatically.
What product distinctions matter most?
Copilot is an umbrella brand, not one uniform agent runtime. A Security Copilot capability cannot be transferred to Microsoft 365 Copilot Researcher. Fabric Data Agent is another distinct surface.
How does AgentenCode assess these products?
Verify licenses and tenant conditions, Microsoft 365 data access, role-based permissions, auditability and the boundary between Copilot Studio and Foundry. Cloud and enterprise security properties require product-specific evidence.
Individual profiles use the existing 36-control AgentenTrust framework. A missing signal does not mean a failed control. Vendor assertions are never presented as independent security measurements.
IN-DEPTH ANALYSIS
How do the agents in this ecosystem differ?
Microsoft 365 Copilot, Copilot Studio, Foundry and Security Copilot serve distinct administrative and data boundaries. Product family association does not imply entitlement across these services.
Product identity
Keep Researcher/Analyst in the Microsoft 365 Copilot scope.
Execution context
Distinguish Copilot Studio creation/deployment from Microsoft Foundry managed agents.
Trust boundaries
Treat Security Copilot governance evidence as security-product-specific, not global Microsoft evidence.
What matters most before adoption?
Copilot is an umbrella brand, not one uniform agent runtime. A Security Copilot capability cannot be transferred to Microsoft 365 Copilot Researcher. Fabric Data Agent is another distinct surface.
Verify licenses and tenant conditions, Microsoft 365 data access, role-based permissions, auditability and the boundary between Copilot Studio and Foundry. Cloud and enterprise security properties require product-specific evidence.
What does evidence-based assessment mean here?
AgentenCode records provider links as curated product associations. Evidence for API access, data residency, logging, approvals, hosting and model selection applies only to its verified agent and scope. The number of products in one ecosystem is not a security or quality rating.
Primary sources by product
| Agent / platform | Product type | Publisher source |
|---|---|---|
| Microsoft Copilot Researcher | AI agent / feature | Official publisher source ↗ |
| Microsoft Copilot Analyst | AI agent / feature | Official publisher source ↗ |
| Microsoft Copilot Studio | Agent platform | Official publisher source ↗ |
| Microsoft Foundry Agent Service | Agent platform | Official publisher source ↗ |
| Microsoft Agent Framework | Agent platform | Official publisher source ↗ |
| Microsoft Fabric Data Agent | AI agent / feature | Official publisher source ↗ |
| Microsoft Security Copilot Agents | Agent platform | Official publisher source ↗ |
These links do not establish shared feature support or automatically transfer Trust signals.