Add documentation for Anthropic as a subprocessor in Azure SRE Agent

Author: Craig Shoemaker

articles/sre-agent/anthropic-subprocessor.md

@@ -0,0 +1,89 @@
+---
+title: Anthropic as a subprocessor in Azure SRE Agent
+description: Learn how Anthropic operates as a Microsoft subprocessor in Azure SRE Agent, including model selection, data residency, and policy controls.
+author: craigshoemaker
+ms.author: cshoe
+ms.topic: conceptual
+ms.date: 03/23/2026
+ms.service: azure
+---
+
+# Anthropic as a subprocessor in Azure SRE Agent
+
+Azure SRE Agent supports multiple AI model providers for investigations, incident response, and operational automation. Anthropic is one of the available providers and operates as a Microsoft subprocessor.
+
+Anthropic operates under Microsoft's oversight with contractual safeguards and technical and organizational measures in place. The Microsoft [Product Terms](https://www.microsoft.com/licensing/terms) and [Microsoft Data Protection Addendum (DPA)](https://www.microsoft.com/licensing/docs/view/Microsoft-Products-and-Services-Data-Protection-Addendum-DPA) apply when you use Anthropic models through Azure SRE Agent.
+
+For more information about subprocessor data access, see [Microsoft Data Access Management](https://www.microsoft.com/trust-center/privacy/data-access). For a list of all Microsoft subprocessors, see the [Service Trust Portal](https://aka.ms/subprocessor).
+
+> [!IMPORTANT]
+> Anthropic models in Azure SRE Agent aren't covered by Microsoft's [EU Data Boundary](/privacy/eudb/eu-data-boundary-learn) commitments. When you select Anthropic, your data (prompts, responses, and resource analysis) might be processed in the United States.
+
+> [!NOTE]
+> For customers in the EU, EFTA, and UK, Azure OpenAI is the default provider. Anthropic is available as an opt-in choice. Anthropic isn't available in government clouds (GCC, GCC High, DoD) or sovereign clouds.
+
+## How model selection works
+
+You choose which AI provider powers your agent. The two options are:
+
+- **Azure OpenAI** — Covered by EU Data Boundary commitments. The default for customers in the EU, EFTA, and UK.
+
+- **Anthropic** — The default for all other regions. Not covered by EU Data Boundary commitments.
+
+When you select Anthropic, your data (prompts, responses, and resource analysis) might be processed in the United States.
+
+## Default settings by region
+
+| Region | Anthropic | Azure OpenAI | Notes |
+|--------|-----------|--------------|-------|
+| **Most commercial regions** (US, APAC, etc.) | Default | Available | No data residency restrictions |
+| **EU, EFTA, and UK** | Available (opt-in) | Default | Anthropic isn't covered by EU Data Boundary |
+| **Government clouds** (GCC, GCC High, DoD) | Not available | Default | Anthropic isn't available in government or sovereign clouds |
+
+## Verify the active model provider
+
+To check which AI model provider your agent is currently using:
+
+1. Go to the [Azure SRE Agent portal](https://sre.azure.com).
+1. Select your agent, and then go to **Settings**.
+1. Under **AI Model Provider**, view the active provider.
+
+## Enable Anthropic in EU Data Boundary regions
+
+If your organization is in the EU, EFTA, or UK and you want to use Anthropic:
+
+1. Go to the [Azure SRE Agent portal](https://sre.azure.com).
+1. Select your agent, and then go to **Settings**.
+1. Under **AI Model Provider**, select **Anthropic**.
+1. Review the data residency notice and confirm.
+
+> [!NOTE]
+> By selecting Anthropic, you acknowledge that your data might be processed outside the EU Data Boundary, including in the United States.
+
+## Restrict Anthropic with Azure Policy
+
+Tenant administrators can use Azure Policy to enforce Azure OpenAI as the only provider. When a policy is active:
+
+- Anthropic appears in the model picker but is disabled.
+- Users see: *"Disabled by your organization's administrator."*
+- Users can't switch to Anthropic regardless of region.
+
+If no policy is set, users in EU/EFTA/UK regions can opt in to Anthropic through the model picker.
+
+## Data handling
+
+When you use Anthropic models in Azure SRE Agent:
+
+- Anthropic processes data under Microsoft's direction and contractual safeguards.
+- The [Microsoft DPA](https://www.microsoft.com/licensing/docs/view/Microsoft-Products-and-Services-Data-Protection-Addendum-DPA) and [Product Terms](https://www.microsoft.com/licensing/terms) apply.
+- Neither Microsoft nor Anthropic uses your data to train AI models.
+- Your data is isolated by tenant and Azure subscription.
+
+For more on how Azure SRE Agent handles data, see [Data residency and privacy](data-privacy.md).
+
+## Related content
+
+- [Data residency and privacy in Azure SRE Agent](data-privacy.md)
+- [Anthropic as a subprocessor for Microsoft Online Services](/copilot/microsoft-365/connect-to-ai-subprocessor)
+- [Microsoft subprocessor list (Service Trust Portal)](https://aka.ms/subprocessor)
+- [EU Data Boundary for the Microsoft Cloud](/privacy/eudb/eu-data-boundary-learn)

articles/sre-agent/toc.yml

@@ -160,6 +160,8 @@ items:
         href: starter-prompts.md
       - name: Data privacy and residency
         href: data-privacy.md
+      - name: Anthropic subprocessor
+        href: anthropic-subprocessor.md
       - name: Billing
         href: billing.md
       - name: Frequently asked questions