Ask Q: Azure Foundry Agents
PATH: Maintenance > AI Configuration
Configure Ask Q to use Azure AI Foundry Agents, enabling access to agents you have built and deployed in your Azure AI Foundry project. Foundry Agents support built-in retrieval-augmented generation and function calling, configured within Azure rather than in Q360.
Applies To
Administrators
Overview
Azure Foundry is a service option in AI Configuration, alongside Azure OpenAI and Custom. Unlike Azure OpenAI, which uses an API key to connect to a model deployment, Azure Foundry connects to a pre-built agent in your Azure AI Foundry project using OAuth application credentials. Retrieval-augmented generation (RAG) is handled inside the agent in Azure — Q360’s built-in Search/RAG fields do not apply to Foundry configurations.
Prerequisites
- Ask Q AI is enabled for your tenant. Contact Solutions360 Customer Care to enable it.
- An Azure AI Foundry project with at least one deployed agent.
- An Azure App Registration with:
- The Azure AI Agent Operator role assigned on your Foundry project
- A client secret created
- The following values from your Azure App Registration: Application (client) ID, client secret value, and Directory (tenant) ID.
- The endpoint URL of your Azure AI Foundry project.
Steps
- Navigate to Maintenance > AI Configuration.
- Click Add.
- Enter a Description. This name appears to users in the Ask Q panel.
- Set Service to Azure Foundry.
- Set Type to Assistant.
- Enter a Sequence number to set the order this assistant appears in the Ask Q panel.
- On the API tab, complete the following fields:
Field What to enter AI URL The endpoint URL of your Azure AI Foundry project (for example, https://<project-name>.services.ai.azure.com) Deployment Name The name of your deployed Foundry Agent Fallback Deployment (Optional) A secondary agent name to use if the primary agent is unavailable OAuth Client ID The Application (client) ID from your Azure App Registration OAuth Client Secret The client secret value from your Azure App Registration OAuth Tenant ID The Directory (tenant) ID from your Azure App Registration - On the AI Config tab, complete optional settings:
Field Description System Prompt Text sent with every request to the agent. Use prompt engineering best practices. Max Tokens Maximum tokens allowed in a single response. Temperature Controls response randomness (0–1 range). Lower values produce more consistent answers. Probability Threshold Alternative to Temperature; uses nucleus sampling to limit token selection. - Click Save.
- Open the Permissions tab and assign the users or groups that should have access to this assistant. Users see the assistant in Ask Q on their next login.
Note: The Search/RAG fields that appear for Azure OpenAI (Search Service Name, Search Index Name, Search API Key) are not available for Azure Foundry. Retrieval configuration is managed within the Foundry Agent in Azure.
Results
The Foundry Agent becomes available in Ask Q for the assigned users or groups. Q360 authenticates with Azure using the OAuth credentials, routes prompts to the agent, and maintains conversation continuity using session IDs managed by Azure. Conversation logs are available through Live Data Designer using the LDView_AIChatLog view.
How Azure Foundry Differs from Azure OpenAI
| Setting | Azure OpenAI | Azure Foundry |
|---|---|---|
| Authentication | API Key | OAuth Client ID, Client Secret, Tenant ID |
| AI URL | Not required | Required — Foundry project endpoint |
| Deployment Name | Model deployment name | Foundry Agent name |
| Built-in RAG search | Configured in Q360 (Search fields) | Configured inside the Foundry Agent in Azure |
Related Articles
- Ask Q: Azure OpenAI (beta)