Understanding the Different Harness Types in Copilot Studio

(1) Share

Report

Posted on by Prasad-MSFT Microsoft Employee

Could someone provide more details on the different harness types available in Copilot Studio, including their key differences, capabilities, use cases, and limitations? It would also be helpful to understand when to use each harness and whether agents built using different harnesses can interact or be connected with each other.

Categories:

General topics

All responses (5)

Answers (0)

Sort by

HR-24060900-0 Microsoft Employee on at

Like
a
(0)

Report
Copy link

Link copied!
Copilot Studio provides different agent harnesses designed for different development and deployment scenarios. The main differences are in how the agent is authored, orchestrated, extended, and integrated with other agents or services.

Declarative/Agent Builder harness – Best for quickly creating agents using natural language, instructions, knowledge, topics, and predefined capabilities. It is suitable for straightforward business scenarios with minimal custom development.

Copilot Studio agent harness – Provides more advanced authoring and orchestration capabilities, including topics, tools/actions, connectors, workflows, knowledge, and integrations. It is a good choice when more control over the agent's behavior and business processes is required.

Custom/Code-based harnesses – Useful when developers need deeper control, custom logic, SDK-based integrations, or advanced orchestration beyond standard low-code capabilities. These are better suited for complex or highly customized scenarios.

When choosing a harness, consider the required level of customization, orchestration, integrations, governance, and development effort. A simpler harness is generally preferable for standard business scenarios, while more extensible harnesses are appropriate for complex requirements.

Agents created using different harnesses can generally work together when supported integration mechanisms are used, such as tools/actions, connectors, APIs, or agent-to-agent communication capabilities. However, interoperability depends on the specific harness, supported protocols, authentication, permissions, and deployment channel.

For a definitive comparison, it would be useful to document the currently supported harness types, their capabilities and limitations, and the supported agent-to-agent connectivity options for the latest Copilot Studio experience.
Suggested answer

Ashlesha-MSFT Microsoft Employee on at

Like
a
(1)

Report
Copy link

Link copied!

Hi Prasad,

Thank you for raising this question.

Copilot Studio currently provides three harness types. A harness is the runtime that coordinates the agent, selected AI model, context, and tools.

1. GitHub Copilot harness
Best for complex, reasoning-heavy, multi-step processes. It supports dynamic planning, retries, connectors, MCP, connected agents, file operations, skills, and memory.

2. Standard harness
Best for predictable, structured scenarios. It uses configured topics, triggers, branches, rules, and agent flows, making it suitable for help desks and repeatable workflows.

3. Copilot chat harness
Best for extending Microsoft 365 Copilot Chat with organizational knowledge from sources such as SharePoint, Teams, email, websites, and Graph connectors. It primarily targets internal Microsoft 365 users.

Agents created with the GitHub Copilot harness can't be converted to the standard harness, or vice versa.

Agents can interact where a supported connection mechanism is available. Copilot Studio supports child agents, connected Copilot Studio agents, and external agents through mechanisms such as A2A. Requirements can include being in the same environment, publishing the connected agent, enabling connections, and granting appropriate permissions. Not every cross-harness combination is documented as supported, particularly for the Copilot chat harness.

Connected-agent solutions can also introduce additional latency and governance requirements. Changes to a connected agent must be published before the calling agent can use them.

For further details, see:

- [Harnesses in Copilot Studio](https://learn.microsoft.com/en-us/microsoft-copilot-studio/harnesses-overview)
- [Add other agents](https://learn.microsoft.com/en-us/microsoft-copilot-studio/authoring-add-other-agents)
- [Connect an existing Copilot Studio agent](https://learn.microsoft.com/en-us/microsoft-copilot-studio/add-agent-copilot-studio-agent)

I hope this clarifies the differences and helps you select the appropriate harness for your scenario.
Suggested answer

Sayali-MSFT Microsoft Employee on at

Like
a
(1)

Report
Copy link

Link copied!
Hello Prasad-MSFT ,

Copilot Studio currently has three harness types, and the right choice depends on what you want the agent to do.

GitHub Copilot harness: Best for complex, multi-step scenarios where the agent needs deeper reasoning, tool orchestration, file handling, skills, memory, or connected agents.

Standard harness: Best for structured and predictable scenarios using topics, rules, branching, knowledge, and workflows. This is generally preferable when you need tighter control over the conversation flow.

Copilot chat harness: Best when the main requirement is extending Microsoft 365 Copilot Chat with organizational knowledge, such as HR policies, onboarding information, or SharePoint content.

Agents can also integrate with other agents and systems through supported mechanisms such as connected agents, workflows, connectors, REST APIs, or MCP. However, agents created with the GitHub Copilot harness cannot be converted to the Standard harness or vice versa, so the harness should be selected based on the intended scenario from the beginning.
Reference Document:
1. Harnesses in Copilot Studio - Microsoft Copilot Studio | Microsoft Learn
2. Agents powered by the GitHub Copilot Harness overview - Microsoft Copilot Studio (GitHub Copilot) | Microsoft Learn
3. Agents powered by the GitHub Copilot Harness overview - Microsoft Copilot Studio (GitHub Copilot) | Microsoft Learn
Suggested answer

11manish 4,388 Super User 2026 Season 2 on at

Like
a
(0)

Report
Copy link

Link copied!

My recommendation would be:

Don't standardize your enterprise on one harness. Standardize the decision process for choosing a harness.

Use Standard when you need control and predictability.

Use GitHub Copilot harness when you need reasoning, autonomy, multi-step execution, tools, files, memory, and agentic orchestration.

Use Copilot chat harness when your objective is extending Microsoft 365 Copilot with organizational knowledge.
Suggested answer

Haque 4,097 Super User 2026 Season 2 on at

Like
a
(1)

Report
Copy link

Link copied!

Hi @Prasad-MSFT ,

I want to highlight 2/3 imporant points:

1. The harness you use affects the billing, features, and capabilities of what you build.

2. Carefule design consideration must be there before choosing harenss types.

3. Agent built on any harness can interact or be connected using A2A protocols. For example, Github Copilot harness agent can delegate specific tasks to standard harness agent or an external SDK-based agent via A2A.

Clean picture is given here, the comparisons of the harnesses.

References:

1. Connect an agent available over the Agent2Agent (A2A) protocol

2. Add a connected agent to an agent

3. Extend Microsoft 365 Copilot with the Copilot chat harness