Build an agent

The Build tab in agents powered by the GitHub Copilot harness in Microsoft Copilot Studio is the central surface for configuring your agent. It's where you define your agent, what it knows, what it can do, and what its limits are.

Note

Features in this article are powered by the GitHub Copilot harness, which uses Copilot Credits for usage-based billing. Learn more about Copilot Credits in Overview of billing for agents powered by the GitHub Copilot harness.

Usage-based billing applies to using, building, testing, and evaluating agents. These actions might consume Copilot Credits.

What is the Build tab?

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.

As you build your agent, use the Build tab to:

  • Define your agent's details, behavior, and scope.
  • Connect the knowledge sources your agent can ground on.
  • Add tools your agent can use to take action.
  • Set behavioral constraints through the agent's instructions.
  • Select the AI model your agent uses.
  • Review the components that are part of your agent.

Key areas on the Build tab

The Build tab is organized into two main areas: the instructions editor and the components panel.

Instructions editor

The instructions editor contains:

  • Agent name: An editable name field where you set the agent's display name.
  • Agent icon: An icon for your agent. You can replace the default icon and its color with something more representative.
  • Instructions: A rich-text editor where you define the agent's behavior in natural language. Use instructions to describe the agent's role, tone, goals, boundaries, and response style. Learn more in Configure agent details and instructions (preview).

Components panel

The components panel is where you configure the resources and capabilities available to your agent:

How the Build tab differs from the standard harness

In the standard harness, explicit topics, triggers, and branching conversation flows shape the agent behavior. In agents powered by the GitHub Copilot harness, with 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.

This 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.