Our backend developer has confirmed that it is possible to call and use each resource individually (we are implementing a RAG pipeline using the LangChain framework with endpoint-and-key authentication).
While there is no issue with service development itself, we would like to test the features provided within the Azure AI Search service directly from the Azure Portal.

We have broken down potential problem areas and tested as follows:
Same-network environment: Configured private endpoints between resources and applied NSG inbound/outbound controls, then spun up a VM on the internal network for testing.
Public-communication environment: Removed the private endpoints to test over public network access.
Permissions: Assigned all roles recommended by the Azure documentation uniformly across all related resources.
Despite making these changes and retesting, we still encountered the same error.
When inspecting the request headers in the browser’s developer tools within the AI Foundry chat playground, we observed that the request path is returning a 404 error.
( Request URL: "https://myresource-test-aifoundry.openai.azure.com/openai/on-your-data-validation?api-version=2023-03-31-preview", Request Method: POST )
We reviewed the relevant API documentation (Link) but could not find any reference to an “on-your-data-validation” path or a “2023-03-31-preview” API version.
Given that our SDK-based RAG setup in the private configuration is functioning correctly and that we have applied every configuration step outlined in the Azure documentation, we believe this to be an issue on the AI Foundry side.
Could you please investigate the root cause of the above 404 error and advise on a resolution?