Configure Self-Service AI Models
Use the Management Portal to control which AI models are available when users create self-service agents in the Chat User Portal.
Overview
The Self-Service Agent AI Models section on the Configuration page lets administrators choose a curated list of completion models. The selection is stored in Azure App Configuration as a comma-separated list of AI model resource names.
Prerequisites
- Access to the Management Portal
- Write permissions on the FoundationaLLM instance configuration set
- AI models created in Models and Endpoints (only models of type completion appear)
Open the Configuration Page
- Sign in to the Management Portal.
- In the sidebar, select FLLM Platform > Configuration.
- Scroll to the Self-Service Agent AI Models section.
Update the Self-Service Model List
- Use the multi-select control to choose the AI models you want to allow for self-service agent creation.
- Use the search field to filter the model list by name.
- Select Save.
- Confirm the prompt to apply the changes.
The User Portal reads the updated list on the next configuration refresh. Existing agents are not modified.
How Models Are Filtered
Only AI models with the completion type appear in the selector. If a model does not appear, verify that it is configured as a completion model in the AI Models page.
Configuration Reference
| Setting | Value |
|---|---|
| Configuration key | FoundationaLLM:Instance:SelfServiceAgentAIModels |
| Display name | Self-Service Agent AI Models |
| Content type | text/plain |
| Format | Comma-separated list of AI model resource names |
Example value:
AzureOpenAI-GPT4,AzureOpenAI-GPT35
Troubleshooting
| Issue | Resolution |
|---|---|
| Model is missing from the selector | Confirm the AI model exists and is configured as type completion. |
| Changes are not reflected in the User Portal | Refresh the User Portal or sign out/in to reload configuration. |
| Save fails | Verify you have write access to the instance configuration set. |