Connected agents overview

Connected agents are agents your primary agent can invoke during a conversation. When a user's request falls within a connected agent’s domain, the orchestration runtime can route the request to that agent. This lets you build modular solutions where each agent handles a specific domain or task.

In the new agent experience, you can currently only connect other agents built in Copilot Studio.

Why use connected agents?

Connected agents are useful when you have multiple agents, each specialized in a different area, and you want a single front-door agent that routes users to the right specialist. Connected agents provide:

• Specialization: Each agent focuses on a specific domain, making it easier to build and maintain.
• Reusability: A specialized agent can be connected to multiple primary agents.
• Separation of concerns: Different teams can own and manage different agents independently.
• Scalability: Add new capabilities by connecting new agents rather than expanding a single agent's instructions.

How connected agents work

• The primary agent's orchestration runtime evaluates each user message.
• If the request aligns with a connected agent's domain, the primary agent delegates the request.
• The connected agent runs in its own orchestration context, with its own instructions, knowledge, and tools. For runtime details, see How the orchestrator routes to connected agents.
• The primary agent sends the relevant conversation history and user message to the connected agent.
• The connected agent processes the request and returns a response to the primary agent.
• The primary agent presents the response to the user.

Related information

• Add a connected agent to an agent
• Manage connected agents in an agent
• Build an agent
• Add other agents to your agent
• Share your agent with others