Unable to Create Foundry Agent In Portal when using an AI Gateway

[bradveng]

Technical Feedback

Make it optional when creating a Foundry Agent in the portal whether a model deployment happens or not.

Currently, when creating a Foundry Agent in the portal, if a model is not present in the Foundry project for the agent to use one will be deployed for you. Please make this optional for the user to choose.

Enterprises utilise Azure Policy that blocks AI model deployments and use AI Gateways for governance and monitoring that has models centralised within a different Foundry resource to that which Agents are deployed into.

This blocks enterprises using an AI Gateway and the Foundry Model Gateway Connection feature from carrying out rapid prototyping using Azure Foundry.

Desired Outcome

When creating a Foundry Agent in the portal, make Model Deployment optional for the user (where a deployment does not already exist in the Foundry project).

Current Workaround

Unable to find a suitable workaround using the portal.
Rapid prototyping using the Foundry Portal is not possible in this scenario.