Harnesses in Copilot Studio

Whatever you build in Microsoft Copilot Studio, a harness powers it behind the scenes. You design your agent or workflow, while the model you select supplies the reasoning and generation. The harness is a runtime that exists between the two: it determines when to call the model, what components to send it, interprets what comes back, and calls the right tools.

Note

This article applies to multiple harnesses. The harness you use affects the billing, features, and capabilities of what you build.

What runs on a harness

Everything you build in Copilot Studio runs on a harness:

  • The GitHub Copilot harness powers agents and workflows built for reasoning-heavy, multi-step work.
  • The Copilot chat harness for extending Microsoft Copilot.
  • The Customer Conversation harness powers agents that answer questions and resolve issues for customers over voice and chat.
  • The standard harness powers rule-based agents and agent flows.

The rest of this article focuses on how the four harnesses differ so you can choose the right one for what you're building.

Why the harness matters

The harness you build on affects several things you care about as a builder:

  • How your work gets done. Some harnesses follow the rules, branches, and steps you define in an agent or workflow. Others can take a goal, break it into steps, and adjust as they go.
  • How much work you can automate end to end. A more capable harness can handle longer, multistep business processes across tools, files, and intermediate decisions, and recover when something doesn't go as planned.
  • What experiences you can build out of the box. Capabilities like creating and editing files, using skills, remembering context, and orchestrating tool calls depend on the harness.
  • How your usage is billed. Different harnesses use different pricing. Learn more in Licensing for agents powered by the standard harness.

GitHub Copilot harness

The GitHub Copilot harness is the most capable option, built for reasoning-heavy agents and workflows that need to complete complex business processes. Instead of following only a fixed script, it can take a goal, break it into steps, call the right tools across connectors, knowledge, MCP, and connected agents, and adjust when a step fails or a request changes. It natively creates and edits Word, Excel, PowerPoint, and PDF files, supports skills and memory, and runs each task in a secure sandbox governed by Copilot Studio. Agents and workflows on this harness use Copilot Credits. It's the best fit for scenarios such as an accounts payable process, where an agent or workflow reads invoices, matches them to purchase orders, and routes exceptions for approval.

Copilot chat harness

The Copilot chat harness allows you to extend Microsoft Copilot Chat. Connect your enterprise knowledge to Microsoft Copilot Chat so employees get answers grounded in your content without leaving their everyday experience. It runs on current chat models and is designed for scenarios where the priority is connecting people to information, such as an employee onboarding agent that answers questions from SharePoint knowledge. Publish to internal teams. Billing is consumption-based or included in Microsoft Copilot user subscription license.

Customer Conversation harness

The Customer Conversation harness is best when your agent needs predictable responses for defined topics while still providing a fluid, natural-language conversation over voice or chat. It uses a multi-agent architecture with specialized sub-agents that can each have their own escalation path to reach human representatives, like in contact centers or customer service. The harness can call tools across connectors, knowledge, MCP, and connected agents, and it supports skills and memory. Use it when you want to publish an agent on a website so customers can chat with it, or assign it a phone number so customers can call it. If the agent can't resolve a customer's question or issue, it can route the conversation to a human representative through the appropriate escalation path. Agents on this harness use Copilot Credits. Learn more in Agents powered by the Customer Conversation harness overview.

Important

The GitHub Copilot harness is a Microsoft Copilot Studio authoring and orchestration framework, not the GitHub Copilot service. Although the harness shares underlying technology and SDK components with GitHub Copilot, customer data isn't sent to or processed by the GitHub Copilot service when agents run in Copilot Studio. Existing Microsoft Copilot Studio privacy, security, compliance, and data residency commitments continue to apply.

Standard harness

The standard harness is a dependable option for rule-based agents and structured, repeatable workflows or conversations, where you want predictable behavior for well-understood requests. You define the topics, prompts, and paths so the experience responds consistently, and it can draw on your existing prompt library and enterprise knowledge. It's a good fit for scenarios such as an internal help-desk experience that answers common questions and routes simple requests through a workflow.

Learn more in Access standard agents and agent flows.

Compare harnesses

Choose the harness based on what you're building:

  • GitHub Copilot harness when your agent needs to reason through longer tasks, work across several tools, handle files, or automate a real business process end to end.
  • Copilot chat harness when your goal is to extend Microsoft Copilot Chat with your organization's knowledge so employees get grounded answers without leaving their everyday experience.
  • Customer Conversation harness when your agent needs to answer questions and resolve issues for your customers over voice and chat, with the option to escalate to a human representative.
  • Standard harness when your scenario is well-defined and rule-based, and you want consistent, predictable answers.
Considerations GitHub Copilot harness Copilot chat harness Customer Conversation harness Standard harness
Best for Complex, multi-step business processes Extending Microsoft Copilot Chat with enterprise knowledge Customer-facing, contact-center-style scenarios over voice and chat Rule-based agents and structured conversations
How it works Reasons through a goal on its own, step by step Connects enterprise knowledge to Microsoft Copilot Chat Multi-agent architecture that can escalate to a human representative Follows the topics and rules you define
Recovers from problems Retries and finds alternative paths automatically Not a focus Escalates to a human representative through a defined path Follows the paths you've built
Works with files Creates, edits, and reasons over Word, Excel, PowerPoint, and PDF files Not a focus Not a focus Not a focus
Skills and memory Yes Not a focus Yes Not a focus
Publishing Internal teams or external customers Internal teams Website chat or a phone number Internal teams or external customers
Billing Overview of billing for agents powered by the GitHub Copilot harness Consumption or included in Microsoft Copilot licenses Copilot Credits (usage-based) Described in Licensing for agents powered by the standard harness
  • Ultima actualizare pe