Exercise - Build a Fabric data agent with an ontology
Now you have an opportunity to practice building a Fabric data agent with an ontology.
In this lab, you create a Fabric data agent for a fictitious company called Lamna Healthcare. You connect the agent to an ontology as its data source, configure instructions to improve query accuracy, test natural language questions, and publish the agent so clinical staff can get answers to questions like "How many rooms in the Surgical Services department are currently occupied?"
Note
You need a Microsoft Fabric Capacity (F2 or higher) to complete this exercise. You also need a Fabric administrator to enable the following tenant settings:
- Capacities can be designated as Fabric Copilot capacities
- Data sent to Azure OpenAI can be processed outside your capacity's geographic region, compliance boundary, or national cloud instance (required only if your capacity is outside the EU or US)
- Data sent to Azure OpenAI can be stored outside your capacity's geographic region, compliance boundary, or national cloud instance (required only if your capacity is outside the EU or US)
Launch the exercise and follow the instructions.
