Does region difference affect in integration of AI search in Foundry agent

Question

Does region difference affect in integration of AI search in Foundry agent

Hari Krishna, Gudditi 0

I have created my Azure AI search service in region Central US and I am trying to use it as knowledge in my Azure AI foundry Agent which is in region EastUS. Will this region difference affect the integration of Search service in my Agent knowledge?

If this doesn't affect, then my Agent is unable to use the AI search knowledge to answer questions. I am using basic tier which is semantic enabled search service and the I am using GPT 4.1 model. Please let me know how can I fix this.

Praneeth Maddali 12,670 Reputation points Microsoft External Staff Moderator

2025-10-29T14:59:32.76+00:00
Hi @Hari Krishna, Gudditi

Thank you for providing details about your setup. To better understand the issue and offer an accurate solution, could you please confirm the following:

Are you using AI enrichment (built-in cognitive skills) in your pipeline? This will help determine if same-region co-location is needed.

Is your Azure AI Search service set up with a private endpoint or is it publicly accessible? This affects Agent connectivity.

How is the Agent connected to the Search service? Please specify whether you’re using an API key or managed identity (RBAC).

Can you provide the exact error message or a screenshot from the Agent playground or Observability tab? This will help identify if the problem relates to connectivity, schema, or semantic ranker.

Does your index have both searchable text fields and a vector field with embeddings? Missing fields or embeddings can lead to retrieval issues.

Is semantic ranker enabled on your Search service? This is required for semantic answers in the Basic tier.

Which model is your Agent using (e.g., gpt-4.1, 2025-04-14,gpt-4.1-nano, 2025-04-14, GPT‑4o)? This helps confirm region support for the model.
Hari Krishna, Gudditi 0 Reputation points

2025-10-30T02:44:24.4666667+00:00
Hi @Praneeth Maddali

Thanks for quick response.

Below are the details:

No we are not using AI enrichment

Azure AI search service is publicly accessible

Agent connected to search service by API key

When checked in AI playground answering flow, the task adherence metric is failed with error AgentEvaluationEvaluatorFailed: BaseException. But how can I debug further to check what the issue is?

Yes index have searchable and vector fields

Yes semantic ranker enabled in search service

My Agent is using GPT-4.1, 2025-04-14

Your answer

Praneeth Maddali 12,670 Reputation points Microsoft External Staff Moderator

2025-10-29T14:59:32.76+00:00

Hi @Hari Krishna, Gudditi

Thank you for providing details about your setup. To better understand the issue and offer an accurate solution, could you please confirm the following:

Are you using AI enrichment (built-in cognitive skills) in your pipeline? This will help determine if same-region co-location is needed.

Is your Azure AI Search service set up with a private endpoint or is it publicly accessible? This affects Agent connectivity.

How is the Agent connected to the Search service? Please specify whether you’re using an API key or managed identity (RBAC).

Can you provide the exact error message or a screenshot from the Agent playground or Observability tab? This will help identify if the problem relates to connectivity, schema, or semantic ranker.

Does your index have both searchable text fields and a vector field with embeddings? Missing fields or embeddings can lead to retrieval issues.

Is semantic ranker enabled on your Search service? This is required for semantic answers in the Basic tier.

Which model is your Agent using (e.g., gpt-4.1, 2025-04-14,gpt-4.1-nano, 2025-04-14, GPT‑4o)? This helps confirm region support for the model.
Hari Krishna, Gudditi 0 Reputation points

2025-10-30T02:44:24.4666667+00:00

Hi @Praneeth Maddali

Thanks for quick response.

Below are the details:

No we are not using AI enrichment

Azure AI search service is publicly accessible

Agent connected to search service by API key

When checked in AI playground answering flow, the task adherence metric is failed with error AgentEvaluationEvaluatorFailed: BaseException. But how can I debug further to check what the issue is?

Yes index have searchable and vector fields

Yes semantic ranker enabled in search service

My Agent is using GPT-4.1, 2025-04-14