Configure generative answers by using Azure AI Search with Foundry
Unlock the complete study guide + 1,040 practice questions across 16 full exams.
Bundled into the existing Designing and Building Integrated AI Agent Solutions in Copilot Studio premium course — no separate purchase.
14-day money-back guarantee — no questions asked.
Included in this chapter:
- Two wirings for one sentence about grounding
- Wiring the Classic data route on the node
- A preview surface with a retirement date attached
- The runtime parameters that decide grounding
- Restricting a node to the approved index
- What the orchestration mode changes
- Proving it with two question classes and a citation check
- How these choices show up in questions
The knowledge-source route and the Classic data route
| Decision | Knowledge-source route | Classic data route |
|---|---|---|
| Where you configure it | Add knowledge, from Overview, Knowledge, or a generative answers node's Properties | The generative answers node's Data source pane, under Classic data |
| What Copilot Studio stores | A search data connection naming one Azure AI Search vector index | An Azure OpenAI Service connection whose Azure-side deployment holds the index |
| Scope of the configuration | The agent, and any node that searches the agent's sources | That one node only |
| Appears in the node's Knowledge sources list | Yes, and can be selected with Search only selected sources | No; Microsoft's note directs you to Classic data instead |
| Where grounding strictness is set | Copilot Studio settings, plus the index's own relevance configuration | Connection properties, as Azure OpenAI On Your Data runtime parameters |
| Documented status | A knowledge source with its own how-to article, carrying neither a preview banner nor a retirement date | Preview on the Copilot Studio side; the Azure-side service is deprecated with a published retirement date |
| Guide page that owns it | connect-azure-ai-search | This page |
Decision tree
Cheat sheet
Unlock with Premium — includes all practice exams and the complete study guide.
References
- Knowledge sources summary - Microsoft Copilot Studio
- Connect your data to Azure OpenAI for generative answers (preview) - Microsoft Copilot Studio
- Using your data with Azure OpenAI in Microsoft Foundry Models (classic)
- Add a generative answers node - Microsoft Copilot Studio
- Orchestrate agent behavior with generative AI - Microsoft Copilot Studio
- Test your agent's knowledge sources - Microsoft Copilot Studio
- Add Azure AI Search as a knowledge source - Microsoft Copilot Studio
- FAQ for generative answers - Microsoft Copilot Studio