Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
The Microsoft Copilot harness is a redesigned authoring and runtime environment for building AI agents. It introduces a natural-language-first approach to agent creation, a unified single-surface design, and an enhanced orchestration model that improves response quality and reasoning capability.
Note
The harness used to build this agent uses usage-based billing. Using, building, testing, and evaluating agents might consume Copilot Credits. Learn more in Overview of usage-based billing.
What is the Microsoft Copilot harness?
The Microsoft Copilot harness is built on an enhanced orchestration runtime that replaces the standard model used in Copilot Studio's standard harness agents. Instead of authoring explicit conversation topics, flows, and branching logic, you describe your agent in natural language (preview) and the system generates the underlying configuration.
Agents powered by the Microsoft Copilot harness are available alongside agents using the standard harness. You choose which harness to use when you create a new agent. Agents created with the Microsoft Copilot harness can't be transferred to the standard harness. However Copilot Studio can create an agent powered by the Microsoft Copilot harness from an agent powered by the standard harness.
Learn more in Harnesses in Copilot Studio and in Upgrade agents to the Microsoft Copilot harness (preview).
Key concepts
Agents powered by the Microsoft Copilot harness are organized around their core components. You can configure the following from the Build tab:
Instructions: Define the agent's identity, tone, scope, and behavior. Learn more in Configure agent details and instructions (preview).
Knowledge: Connected data sources and memory give the agent context for answering questions. Learn more in Add knowledge to an agent.
Tools: Tools let the agent perform actions, such as calling APIs and running flows. Learn more in Tools overview for agents.
Skills: Skills define reusable structured behaviors. Learn more in Skills overview for agents.
Model: Choose the AI model that powers the agent's reasoning. Learn more in Select a model for an agent.
Connected agents: Delegate specialized tasks to other agents. Learn more in Connected agents overview.
Memory: Turn Memory on to help an agent remember details from its interactions and use that context for future interactions. Learn more in Memory (preview).
How the Build tab differs from the standard harness
In agents powered by the Microsoft Copilot harness, the Build tab brings the core parts of agent authoring together in one place. Instead of designing explicit topic flows and branching logic, you describe your agent in natural language and connect the resources it needs to respond and act. By using the Build tab's natural-language-first approach, you create agents that rely on enhanced orchestration to interpret your instructions, decide when to use knowledge, and determine when to invoke tools. In the standard harness, explicit topics, triggers, and branching conversation flows shape the agent behavior. The Microsoft Copilot harness's approach reduces the need to author detailed flow logic up front. Instead, you focus on describing your agent, connecting the right components, and defining the boundaries that guide how it operates.
Comparison of features: Standard vs. Microsoft Copilot harnesses
| Category | Feature | Standard harness | Microsoft Copilot harness | Comments |
|---|---|---|---|---|
| Tools | Connectors | Supported | Supported | |
| Tools | MCP | Supported | Supported | |
| Tools | Computer use | Supported | Planned | |
| Tools | Prompt | Supported | Not supported | Replaced by Skills in the Microsoft Copilot harness |
| Tools | Workflows | Supported | Supported | |
| Knowledge | Public websites | Supported | Supported | |
| Knowledge | File upload | Supported | Supported | |
| Knowledge | SharePoint | Supported | Supported | |
| Knowledge | Graph connectors | Supported | Supported | |
| Knowledge | Dataverse tables | Supported | Supported | |
| Knowledge | Azure AI Search | Supported | Not supported | |
| Skills | Skills | Not supported | Supported | GH agents only |
| Memory | Memory | Not supported | Supported | GH agents only |
| Code interpreter | Code interpreter | Supported (opt-in) | Supported (native) | |
| Topics | Triggered on user intent | Supported | Supported | Replaced by Hooks in the Microsoft Copilot harness. Hooks run workflows automatically at specific points in an agent’s lifecycle, such as when a session starts or before a tool run |
| Topics | Other system triggers (for example, Message received) | Supported | Supported | Replaced by Hooks in the Microsoft Copilot harness. Hooks run workflows automatically at specific points in an agent’s lifecycle, such as when a session starts or before a tool run |
| Connected agents | Copilot Studio agents (same environment) | Supported | Supported | |
| Connected agents | Foundry agents | Supported | Supported | |
| Connected agents | A2A protocol agents | Supported | Supported | |
| Connected agents | M365 SDK agents | Supported | Supported | |
| Agent evaluation | Agent evaluation | Supported | Supported | |
| Monitor | Monitor | Supported | Supported | |
| Channels | M365 / Teams | Supported | Supported | |
| Channels | Web site | Supported | Supported | |
| Channels | SharePoint | Supported | Not supported | |
| Channels | Supported | Not supported | ||
| Channels | Dynamics 365 Contact Center | Supported | Not supported | |
| Triggering | Autonomously trigger | Supported | Supported | In the Microsoft Copilot harness, use triggers (preview) or an Agent Node in Workflows |
| Activity / Run history | Activity / Run history | Supported | Supported | Review sessions and runs on the Monitor tab |
| Auth | Microsoft (Entra ID) | Supported | Supported | |
| Auth | No authentication / anonymous | Supported | Supported | |
| Auth | Manual / custom authentication | Supported | Not supported | |
| ALM | Solutions / ALM support | Supported | Supported | |
| Governance | DLP / governance controls | Supported | Supported | |
| Languages | Languages explicitly added & enabled | Supported | Supported out of the box | In the Microsoft Copilot harness: response language is controlled by the agent instructions |
| Entities | Custom entities | Supported | Not supported |
Agent authoring surface
Agents powered by the Microsoft Copilot harness use a tab-based surface for building and managing:
| Tab | Purpose |
|---|---|
| Build | Configure the agent's identity, knowledge, tools, skills, model, and memory |
| Preview | Test the agent interactively in a preview chat interface |
| Evaluate | Create and run test sets to measure agent quality |
| Monitor | Review production health, outcomes, reactions, sessions, and credit usage |
Agent lifecycle
Building an agent powered by the Microsoft Copilot harness follows this general flow:
- Create: Start from the Copilot Studio home page and choose to create a new agent powered by the Microsoft Copilot harness.
- Build: Configure who the agent is, what it knows, what it can do, and what its limits are.
- Test: Use the Preview tab and Evaluate tab to validate the agent's behavior.
- Publish: Deploy the agent to your chosen channels.
- Monitor: Review production health and identify where to improve the agent. You can open Monitor before publishing, but production data appears after people use the published agent.
Orchestration and reasoning
The Microsoft Copilot harness uses an improved orchestration runtime that supports deep reasoning capabilities and higher-quality responses compared to the standard harness—particularly over Microsoft 365 data. Unlike the standard harness, where you can configure orchestration behavior, the Microsoft Copilot harness uses this enhanced orchestration model for all agents. Learn more in Harnesses in Copilot Studio.
Availability
Agents built with the standard harness remain fully supported alongside agents powered by the Microsoft Copilot harness. For information on standard agents, see Create and delete agents and AI-based agent authoring overview. Learn more in Harnesses in Copilot Studio.
Send feedback about the agent authoring experience
To send comments about the authoring experience, select the feedback button 
in the top toolbar of the Microsoft Copilot Studio agent authoring experience. To learn more about support and feedback options, see Find support and give feedback for Copilot Studio.