Exercise 06: Use generative AI orchestration to interact with your connectors
Scenario
Some customer requests require dynamic actions rather than static answers. Contoso intends to expose line‑of‑business operations (such as creating or updating cases) through plugin actions that the agent can invoke on demand. This exercise shows how to enable generative AI orchestration, let the agent choose the right action at runtime, and test the end‑to‑end flow.
Objectives
After this exercise you’ll be able to:
- Understand the basics of agent tools.
- Use Copilot Studio to request data from another data source using tools, then return the data in a conversational dialog with the end user.
Duration
Estimated time: 20 minutes
Generative AI orchestration
Using generative AI to determine how your agent responds can make the conversation more natural and fluid for users. When a user sends a message, your agent selects one or more actions or topics to prepare its response.
Multiple factors determine the selection. The most important factor is the description of the topics and tools. Other factors include the name of a topic or actions, any input or output parameters, and their names and descriptions. Descriptions make it possible for your agent to be more accurate when it associates a user’s intent with actions and topics.
In generative mode, an agent can select multiple tools or topics at once to handle multi-intent queries. Once tools and topics are selected, the agent generates a plan determining their execution order.
When you test an agent that uses generative mode in Copilot Studio, you can open the conversation map to follow the execution of the plan.