Great question! Based on the current documentation, yes, this should work.
The key reason is that both pieces use the same foundation. The Azure AI Foundry Agent docs show that existing Foundry agents are wrapped using the standard Agent class with an AzureAIAgentClient as the chat client. The docs explicitly state: "The agent is a standard BaseAgent and supports all standard agent operations" - and that includes run_stream(), which is what the AG-UI integration needs to produce SSE events.
Since add_agent_framework_fastapi_endpoint accepts any Agent instance, you'd pass in your Foundry-backed agent the same way you would an AzureOpenAIChatClient-backed one. The AG-UI layers (EventBridge, Orchestrators) handle the translation to protocol events.
I haven't tested this exact combination end-to-end yet, so if you try it out, I'd love to hear how it goes!