Microsoft
Microsoft's enterprise agent platform, with Copilot Studio for building and governing agents, Microsoft 365 Copilot agents in the productivity apps, per user memory, model choice, computer use and evaluations before publishing.
Microsoft's enterprise agent platform is Copilot Studio, where agents are built, given tools and knowledge, published and then monitored. Agents answer through generative orchestration, which picks among topics, tools, other agents and knowledge sources and chains them in one response, and they can be started by an event rather than a person, so a new row in a table or a message arriving can set one running.
Knowledge is attached at design time from SharePoint, OneDrive, Salesforce, ServiceNow, Azure SQL, Dataverse, Azure AI Search, Bing Custom Search, public or enterprise websites, or uploaded documents, and new knowledge goes into an index without retraining the model. Agents can keep a per-user memory that the user, not the maker, can read and erase. The maker chooses which model runs the agent from a dropdown, including external models, with separate settings for reasoning and for generative responses and published availability by region.
Where no API exists, computer use drives websites and Windows desktop applications through a virtual mouse and keyboard, and escalates to a named human reviewer by email or an inline card when it needs confirmation. Before publishing, evaluations run named test sets of conversations and score the results, and the same runs can be triggered through the Power Platform API inside a release pipeline. After publishing, a monitor tab reports sessions, engagement, credit consumption and response quality.
Agents are governed with the tenant's existing controls, including geographic data residency, data loss prevention, compliance certifications, a runtime protection status view and an automatic security scan. For organizations that onboard Agent 365, agents become identities in Microsoft Entra governed by Conditional Access and access reviews. Agents built in code with the Microsoft 365 Agents SDK, in C#, JavaScript or Python, can be attached and called.
Vendor details
Canonical URL
https://www.microsoft.com/en-us/microsoft-365-copilot/agents
Category
Enterprise operations agent
Subcategory
Platform — Copilot Studio and Microsoft 365 Copilot agents
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Tools call APIs, run automated flows and reach external services, with connections and connectors carrying the credentials. Model Context Protocol makes every tool and resource a connected MCP server publishes available to the agent. Knowledge sources include SharePoint, OneDrive for Business, Salesforce, ServiceNow, Azure SQL, Dataverse, Azure AI Search, Bing Custom Search and Azure DevOps. Agents built with the Microsoft 365 Agents SDK attach by their messaging endpoint, and the Power Platform REST API runs evaluations from a release pipeline.
In practice
A process depends on a system with no API. Computer use drives the website or the Windows desktop app through a virtual mouse and keyboard, and escalates to a named reviewer by email when it needs a decision.
Work should start without anyone asking. An event trigger fires the agent when something happens in a connected system, and generative orchestration works out which tool, topic or other agent to use.
You need to know a change did not make the agent worse. Evaluations run a named set of test conversations and score the results before you publish, and the same run can be triggered from the API inside a release pipeline.
Agentic Index coverage score
14.0 / 14 capabilities · 100%
| Integrations & Tool Calling | Full |
|---|---|
|
Tools are what an agent invokes beyond answering from knowledge. They call APIs, run automated flows, reach external services and fetch real time data, and the orchestrator picks one from the conversation context and the tool's description. Connections and connectors carry the credentials. An agent can be extended with Model Context Protocol so every tool and resource a connected MCP server publishes becomes available automatically, with changes on that server reflected without rework. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/agents-experience/tools-overviewread 2026-09-17 |
|
| Workflow Orchestration | Full |
|
Generative orchestration lets a custom agent answer with the relevant topics, tools, other agents and knowledge sources, fill inputs from earlier conversation history, chain several capabilities together in one response, and act on event triggers, with a triggered agent choosing which tool, topic or agent to invoke. Deterministic work sits beside it as agent flows, and other agents are attached as connected agents and managed from the agent's own Agents page. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/faqs-generative-orchestrationread 2026-09-17 |
|
| Knowledge Grounding & RAG | Full |
|
Knowledge sources are content the maker attaches at design time so every user gets the same grounding. The Add knowledge dialog offers featured sources (public websites, SharePoint, OneDrive for Business, Salesforce, ServiceNow, Azure SQL), advanced sources (Azure DevOps wiki and work items, custom connector, enterprise websites) and direct upload of PDF, Word, Excel and PowerPoint files. Dataverse tables, Azure AI Search, Bing Custom Search, real time connectors and file groups can also be used, and sources can be managed and edited. New knowledge goes into an index without retraining the model, and a file an end user uploads is handled as an attachment to the conversation rather than as a knowledge source. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/agents-experience/knowledge-sources-overviewread 2026-09-17 |
|
| Human Oversight & Guardrails | Full |
|
Human supervision is a configurable setting. When the agent needs confirmation or more information mid execution, it escalates to the named reviewer, sending a review request by Outlook email or showing an inline review card in the activity panel, and a published table shows which underlying models support the feature. Microsoft also covers administration of computer use and its troubleshooting and limitations, and warns explicitly that these agents can meet prompt injection attempts and should run in trusted isolated environments with validation before instructions are executed. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/human-supervision-computer-useread 2026-09-17 |
|
| Security, Identity & Governance | Full |
|
Copilot Studio is covered by Microsoft's standards certifications and regulatory compliance programs, and its customer facing controls include geographic data residency, data loss prevention, environment routing, regional customization, an agent runtime protection status view for makers and an automatic security scan. Where an organization onboards Microsoft Agent 365, Copilot Studio agents are represented as identities in Microsoft Entra and governed with Conditional Access, role based and attribute based access control and access governance workflows, in addition to the existing Power Platform and Microsoft 365 mechanisms. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/security-and-governanceread 2026-09-17 |
|
| Observability & Auditability | Full |
|
After publishing, the Monitor tab reports conversation sessions, user engagement with message counts and session duration, credit consumption over a chosen period, thumbs up and thumbs down reactions with comments on specific responses, and response quality indicators over time, filterable by time range. Errors can be reviewed in their own view. Computer use runs have their own monitoring, and where Agent 365 is onboarded it carries centralized observability and policy enforcement across agents rather than only where each was built. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/agents-experience/analytics-overviewread 2026-09-17 |
|
| Memory & State Persistence | Full |
|
Memory gives an agent a separate store per user, held in a dedicated folder in Microsoft managed storage that the agent reads from and writes to. The agent records preferences and context shared during a conversation as files there and applies them in later conversations. The maker turns memory on or off per agent, and a user's memories are private to that user and invisible to the maker. Each user can ask the agent what it remembers, update or forget a specific memory in chat, or view and delete all memories in a portal. Memories are deleted automatically after 28 days without interaction, and turning the feature off stops use without deleting. Memory is a production ready preview for agents on the GitHub Copilot harness, and agents on the Customer Conversation harness have a separate memory capability. Sourcelearn.microsoft.com Copilot Studio memory overviewread 2026-10-25 |
|
| Deployment & Data Residency | Full |
|
In Copilot Studio, geographic data residency covers where data is stored, processed and managed, tied to the environment's geography. Microsoft also covers data locations, the security side of residency, and deliberately moving data across regions when generative features are needed outside the geography or capacity is constrained. Model availability is published per region in tables on the model selection page, with a warning that data in an experimental model may be processed outside the organization's geographical boundary. Virtual network support is also offered. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/geo-data-residencyread 2026-09-17 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
The Agents page in Copilot Studio offers a set of agent templates, each a complete agent a maker creates from and then customizes. IT Helpdesk, Safe Travels, Weather, Website Q&A and Sustainability Insights are custom agents, and Benefits is a Microsoft 365 Copilot agent, each with its own documentation page. The catalog sits inside the product and can be browsed, each template works as soon as it is chosen, and the set is small. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/template-fundamentalsread 2026-09-17 |
|
| Triggers & Channel Coverage | Full |
|
Event triggers let an agent act on its own rather than waiting for a person. A chosen event generates a payload, a connector delivers it, and the agent executes against its instructions and the payload's variables. Topic triggers, by contrast, need a user's phrase. In Microsoft's example an agent starts when a new entry lands in a Dataverse table, and agents also reach users through conversational channels and the Microsoft 365 apps. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/authoring-triggers-aboutread 2026-09-17 |
|
| Model Flexibility & Routing | Full |
|
The maker chooses the primary model for an agent's generative orchestration from a dropdown, with separate settings for deep reasoning, for generative responses and for the prompt builder. An external model can be chosen as the primary model, and Microsoft publishes tables of model availability by region and by scope. Models carry tags for their status, experimental ones are marked as not recommended for production with a limitations page, and makers can continue on a model after it is retired. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/authoring-select-agent-modelread 2026-09-17 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
The Microsoft 365 Agents SDK builds agents in code in C# and .NET, JavaScript on Node.js, and Python, and such an agent is attached to a Copilot Studio agent by its messaging endpoint so the platform can call it. The platform is itself callable. The Power Platform REST API triggers agent evaluations programmatically for release validation, regression testing and CI/CD pipelines, and separate admin APIs delete, quarantine and reassign ownership of agents. Model Context Protocol brings external tools and resources in. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/add-agent-microsoft-365-agents-sdk-agentread 2026-09-17 |
|
| Testing, Debugging & Optimization | Full |
|
Evaluations are named test sets. A conversation is a test case with user messages and optionally expected responses, created by hand, generated with AI, or uploaded from CSV, grouped into an evaluation with a test method, and run to produce scored results that track quality across changes and validate behavior before publishing. Evaluations run under a chosen authenticated user profile so the agent's tools and connections are exercised. The Power Platform REST API runs them programmatically against predefined test sets for release validation, regression testing and CI/CD. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/agents-experience/analytics-agent-evaluation-introread 2026-09-17 |
|
| Browser & Computer Use | Full |
|
Computer use is a tool an agent adds from its Tools page. It interacts with a Windows computer through a virtual mouse and keyboard, working across websites and desktop applications by selecting buttons, choosing menus and typing into fields, described in natural language rather than code, for work where no API exists. It is powered by a computer using agent model combining vision with reasoning, which Microsoft says adapts when buttons or screens change rather than breaking. Microsoft also covers where runs execute, human supervision, monitoring, administration and limitations for computer use. Sourcelearn.microsoft.com/en-us/microsoft-copilot-studio/computer-useread 2026-09-17 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Product updates
Admins can now create custom Model Context Protocol connectors in the Microsoft 365 admin center and deploy them across the organization. Microsoft 365 Copilot then draws on the company's own data through those connectors. Creation sits in the Copilot extensibility section, next to the federated MCP connectors admins could already deploy from the same center.
Bears on: Integrations
View sourceAgent accounts in Copilot Studio let agents work in Microsoft 365 with their own credentials, so people can assign them work in Teams and Outlook, mention them in Word, Excel and PowerPoint comments, and hand off tasks the agent continues between interactions. Tools run under the agent's own approved credentials and scoped permissions rather than a user's or maker's, and agents keep business process memory across the people they work with.
Bears on: Security / enterprise
View sourceMicrosoft's September 23 release notes add Outlook actions that let Microsoft 365 Copilot delete, move, copy, and categorize email and manage inbox rules and folders through natural language. Copilot requests confirmation before acting on more than five messages or changing inbox rules. The capability covers classic and new Outlook for Windows, web, Mac, and mobile.
Bears on: Agent capability
View sourcePricing
Copilot Business from $18 per user a month billed annually, a first year offer running 1 July to 31 December 2026 (standard $21, or $25.20 billed monthly), with Copilot Chat at no extra cost. Agent usage is billed in Copilot Credits.
Hybrid. Copilot seats are a per user monthly fee, billed on an annual commitment or month to month at a higher rate, and agent usage is metered separately in Copilot Credits.
Included quota
Copilot Chat is free and offers web grounded AI chat with enterprise data protection (no training on prompts), but no access to internal organization data without agents or credits. M365 Copilot Business ($18 to $21 per user a month) adds Copilot in the Office apps, Chat with Work IQ and internal Copilot Studio agents, for up to 300 users. M365 Copilot Enterprise ($30 per user a month) includes everything in Business plus tenant wide Microsoft Graph grounding, Purview governance, advanced admin, bundled Copilot for Sales, Service and Finance, and Dynamics 365 Copilot features. In Copilot Studio, internal agents are free and external agents run on $200 packs of 25,000 messages or Azure pay as you go. Vertical SKUs (for example Copilot for Sales at about $50 per user a month) and Dynamics 365 Copilot vary by existing licensing.
What is public
Microsoft publishes seat pricing and the agent metering structure, with the detailed rates in a separate licensing guide. Copilot Chat is included at no additional cost with eligible Microsoft 365 plans. Microsoft 365 Copilot Business starts from $18.00 per user per month promotionally, originally $21.00, as an auto renewing annual subscription, or $25.20 per user per month billed monthly. Bundled Microsoft 365 suites with Copilot built in are listed at several higher figures. Copilot Credits are the single currency for agent usage, available through pay as you go meters, prepurchase plans and prepaid pack subscriptions, with a published usage estimator and licensing guide, and a GCC plan.
Billing mechanics
Billing runs on two separate lines. Seats are a per user monthly Copilot subscription, billed on an annual commitment or month to month, sold as an add on to a qualifying Microsoft 365 plan, with bundled suites that include Copilot also published. Agent usage is billed in Copilot Credits, the common currency across Copilot Studio, bought through pay as you go meters, prepurchase plans or prepaid pack subscriptions, then allocated to environments and monitored by administrators in the Power Platform admin center. Copilot Credits replaced messages as the unit with no change to prepaid pack quantity or the pay as you go rate. Billing begins at publish on the standard harness and from the start of building on the GitHub Copilot harness. A licensing guide and a usage estimator forecast credit volume from agent type, traffic, orchestration, knowledge and tools.
Cost watchouts
Copilot is sold as an add on to a qualifying base Microsoft 365 license, so the full cost is the base plan plus Copilot. Bundled suites with Copilot built in are also listed, at several higher price points. Agent building is a separate consumption line. On the GitHub Copilot harness, credits are charged for model tokens, tools including knowledge and MCP servers, and the harness itself, and billing starts during building rather than at publish. The $18 Business figure is promotional against a standard $21.
Variable cost rationale
Seats are fixed. Each Copilot license is a per user monthly fee, billed annually or month to month, so that part of the bill moves only with headcount. Agent usage is metered in Copilot Credits, drawn by every external agent interaction in Copilot Studio, and premium and connector messages draw several each. On the GitHub Copilot harness, credits are charged for model tokens, tools including knowledge and MCP servers, and the harness itself, starting while the agent is still being built. Prepaid packs of 25,000 for $200 and prepurchase plans set a known volume, and administrators allocate credits to environments and monitor them in the Power Platform admin center. Azure pay as you go covers usage once prepaid credits run out, so a large agent fleet keeps billing rather than stopping. Internal Copilot Studio agents are free.
Additional watchouts
The full cost is the base license plus Copilot, not the add on headline. External agent messages are the line item that grows quietly, since premium and connector messages consume several messages each. Copilot is a family of SKUs (consumer and work, Business and Enterprise, included and standalone Studio), which makes buying the wrong one easy. Promotional and Enterprise Agreement pricing expires on set dates. Agents require Azure.
Overage / add-ons
Copilot Studio external agent messages beyond the $200 pack of 25,000 bill at $0.01 for a standard message and $0.02 for a generative message. Premium messages (line of business connectors, Dataverse, complex evaluations) consume several messages each. Azure pay as you go covers usage automatically when prepaid credits run out. Seat licenses are a fixed fee per user per month on an annual or monthly term.
Sales call required
Mixed (some tiers require a call)
Free / trial
Copilot Chat at no extra cost with eligible Microsoft 365 plans, and a one month free trial of the Copilot plans with a card required
Lowest paid plan
Microsoft 365 Copilot Business at $18 per user per month billed annually, a first year promotional price offered until December 31, 2026 (standard $21), as an add on to Microsoft 365
Commercial notes
Microsoft's agent platform is Copilot Studio as the agent builder, the Microsoft 365 Copilot agents that run in the productivity apps, and Agent 365 as the governance plane. It is sold as per user Copilot subscriptions plus credit metered agent usage, self serve for smaller volumes and quoted through enterprise agreements at scale, with a US Government Community Cloud plan published.
Key ambiguities
Copilot Chat is listed at no additional cost with eligible plans, Copilot Business at a promotional $18 against a standard $21 on an auto renewing annual subscription (or $25.20 billed monthly), and several bundled suites with Copilot built in at higher published figures. The $30 per user Enterprise add on figure, the per credit overage rates and the premium message multipliers are not on the pricing page. The detailed agent rates sit in the separately published Copilot Studio Licensing Guide. Copilot is a family of SKUs rather than one product, and its pricing changes often.
Cancellation / refund
Copilot plans bill annually or month to month, and a 7 day cancellation window gives a prorated refund. Copilot Studio packs run monthly or on Azure pay as you go with no commitment, Enterprise Agreements are negotiated, and a base Microsoft 365 license is required throughout.
Support SLA / resale
Support follows the standard Microsoft tiers of the base plan. Copilot is sold direct, through resellers and through Cloud Solution Provider partners, with Purview governance, Entra identity and enterprise data protection (data stays in the tenant, with no model training).
Missing data
Dynamics 365 Copilot and Enterprise Agreement pricing are quoted, varying by existing licensing and volume. Exact Copilot Studio premium message multipliers are not fully published, and vertical SKU bundling shifts over time.
Related vendors
- 11th Estate — Agentic platform whose AI engine scans markets, matches an…
- Abacum — FP&A platform for growing companies whose AI Modeler builds, edits…
- Adopt AI — AI native accounting and bookkeeping firm whose agents run…
- Aera Technology — Decision intelligence platform where an always on agent executes…
- Akro AI — On premise operational intelligence that automates document heavy…
- alfred_ — AI executive assistant that triages the inbox overnight and scores…
Alternatives to Microsoft
The closest documented capability profiles to Microsoft among enterprise operations agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Appian13.5 / 14A lighter documented profile than Microsoft
- Atlassian Rovo13.5 / 14A lighter documented profile than Microsoft
- Pega13.5 / 14A lighter documented profile than Microsoft
- Perplexity13.5 / 14A lighter documented profile than Microsoft
- ServiceNow13.5 / 14A lighter documented profile than MicrosoftMicrosoft vs ServiceNow →
- UiPath13.5 / 14A lighter documented profile than Microsoft
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded