{"slug": "anthropic-models-in-microsoft-online-services", "title": "Anthropic Models in Microsoft Online Services", "summary": "Microsoft is integrating Anthropic AI models into its enterprise online services, with Anthropic onboarding as a subprocessor to ensure compliance and security under Microsoft's framework. The models will be enabled by default for most commercial cloud customers, excluding EU/EFTA and UK, and full availability is expected by March 2026.", "body_md": "Note\n\nAccess to this page requires authorization. You can try [signing in](#) or [changing directories].\n\nAccess to this page requires authorization. You can try [changing directories].\n\nMicrosoft is introducing a new offering with Anthropic AI models as part of Microsoft Online Services, delivering enterprise-grade commitments and safeguards to ensure secure and responsible use of Anthropic models within your organization.\n\nTo enable this change, Anthropic has onboarded as a Microsoft subprocessor. This change simplifies the experience and strengthens compliance and security under Microsoft's enterprise framework. The Microsoft Customer Copyright Commitment (CCC) applies to Anthropic models used within products covered by the CCC, including Microsoft 365 Copilot and Copilot Studio.\n\nAs a subprocessor, Anthropic will operate with Microsoft oversight through contractual safeguards and appropriate technical and organizational measures. Unless models are labeled \"Preview models with Data Retention,\" the [Microsoft Product Terms](https://www.microsoft.com/licensing/terms?msockid=344e0e6ad66c6b3e19441848d7416abd) and [Microsoft Data Protection Addendum (DPA)](https://www.microsoft.com/licensing/docs/view/Microsoft-Products-and-Services-Data-Protection-Addendum-DPA?lang=18&msockid=344e0e6ad66c6b3e19441848d7416abd) apply to use of Anthropic models through Microsoft's enterprise Online Services. Such use is also covered under our [Enterprise Data Protection](enterprise-data-protection).\n\nFor more information about subprocessor data access, see [Microsoft Data Access Management](https://www.microsoft.com/trust-center/privacy/data-access). To see a list of Microsoft subprocessors, see [Service Trust Portal](https://aka.ms/subprocessor).\n\nMicrosoft will enable Anthropic models on by default for most customers in commercial cloud (excluding EU/EFTA and UK). This update gives users in your organization the ability to use multiple AI models in their Microsoft offerings, such as in Microsoft 365 Copilot, Researcher, Copilot Studio, Power Platform, and Copilot in Microsoft 365 apps. This affirms Microsoft's commitment to offering choice between leading AI models while maintaining enterprise-grade security and compliance.\n\nImportant\n\nAnthropic models deployed in Microsoft offerings such as Microsoft 365 Copilot, Researcher, Copilot Studio, Power Platform, and Copilot in Microsoft 365 apps are currently excluded from the EU Data Boundary, and when applicable, in-country processing commitments. Customers within the EU Data Boundary and customers in the UK will have Anthropic models disabled by default. Anthropic models aren't currently available for use in government clouds (GCC, GCC High, DoD) or sovereign clouds.\n\nAnthropic as a subprocessor is being introduced gradually and isn't yet available to all organizations. During this phased rollout, some features may be limited for your organization. Full availability is expected by the end of March 2026.\n\nNote\n\nFrom time to time, some newly available and advanced models from Anthropic may still be made available with separate controls that allow Microsoft tenant admins to opt in to use preview models under Anthropic's separate [commercial terms](https://www.anthropic.com/legal/commercial-terms) and [data processing agreement](https://www.anthropic.com/legal/data-processing-addendum). For example, different terms apply to Anthropic Preview models with Data Retention, and these models must always be separately enabled even where other Anthropic models are on by default. For more information, see [Preview AI models in Microsoft Online Services](manage-preview-ai-models).\n\n## Anthropic Preview models with Data Retention\n\nSome Anthropic models are subject to retention and storage of [Customer Data by Anthropic](https://support.claude.com/articles/15425996-data-retention-practices-for-mythos-class-models)(\"Preview models with Data Retention\"). For these models, Anthropic acts as an independent data processor, not a Microsoft subprocessor. Preview models with Data Retention are made available for use from within some Microsoft Online Services but subject to [Anthropic's Commercial Terms of Service](https://www.anthropic.com/legal/commercial-terms) and [Anthropic's Data Protection Addendum](https://www.anthropic.com/legal/data-processing-addendum). Your organization's use of Anthropic Preview models with Data Retention is **optional**. Preview models with Data Retention have additional controls in the Microsoft 365 admin center and remain **default-off** for all scenarios, including in regions when different Anthropic models are on by default in commercial cloud and for which Anthropic is a Microsoft subprocessor.\n\n## Manage Anthropic's Claude model settings in the Microsoft 365 admin center\n\nMicrosoft is making Anthropic models available by default in certain regions. In Microsoft 365 Copilot (web, desktop, and mobile), UI indicators will show when Claude models are in use. In Copilot Studio, creators must select the model during agent creation. In capabilities such as Edit with Copilot in Microsoft 365 apps or Researcher, users can select **Claude**.\n\n## Opt-in regions and exclusions\n\nIn some regions, Anthropic's models aren't available by default. For these regions, the toggle will appear but the default is set to **Off**. These regions include the [European Union (EU), the European Free Trade Association (EFTA)](/en-us/privacy/eudb/eu-data-boundary-learn), and the United Kingdom (UK).\n\nIn addition, Anthropic models aren't available in government clouds (GCC, GCC High, DoD) as there's no FedRAMP certification in place yet. They're also not available in other sovereign clouds. No toggle will be present for government or sovereign clouds.\n\nNote\n\nOn April 3, 2026, Microsoft introduced a new Microsoft 365 admin center setting **Copilot in M365 apps with Anthropic models** in EU/EFTA and UK to enable Anthropic as the default model for Copilot in Microsoft 365 apps. For more information, see [Copilot in Microsoft 365 apps with Anthropic models](copilot-anthropic-apps).\n\n## Opt-in to use Anthropic's models\n\nIf your organization is in a region that has Anthropic as a subprocessor set to **Off** by default, you can choose to opt in so Anthropic's models are available for your organization. You must be a member of the global administrator role to perform this task. For more information, see [About admin roles](/en-us/entra/identity/role-based-access-control/permissions-reference#global-administrator) and [Global administrator](/en-us/entra/identity/role-based-access-control/permissions-reference).\n\nGo to the Microsoft 365 admin center and select\n\n**Copilot**->** Settings**->** View all**.Select\n\n**AI providers operating as Microsoft subprocessors**.On the\n\n**AI providers operating as Microsoft subprocessors** page, under**Available subprocessors for your organization**, select** Anthropic**, and** Save**.Under\n\n**Choose who can access Anthropic models for Copilot and generative AI experiences**, select your users or groups and** Save**.\n\nNote\n\nYou can restrict user access to AI provider subprocessors by assigning permissions to specific users or Microsoft Entra ID security groups in the Microsoft 365 admin center. These assignments are applied at the provider level and enforced across Microsoft 365 Copilot and Copilot Studio experiences. When access is limited by user or group membership, only the assigned users can use Copilot features or agents that rely on that AI provider. Review existing user or group assignments and update policies or configurations as needed. For more information on user and security group access, see [Assign AI provider access to users and groups](copilot-ai-provider-user-sec-group-access). For more information on creating security groups, see [Create a security group](/en-us/microsoft-365/admin/email/create-edit-or-delete-a-security-group).\n\nIf your organization is in the European Union (EU), the European Free Trade Association (EFTA), or the United Kingdom and you previously opted in to use Anthropic models under Anthropic's separate commercial terms and data processing agreement, you'll need to opt in again. The toggle will be set to **Off** by default.\n\nSome features are only available when Anthropic models are enabled. If you turn off Anthropic as a subprocessor, certain features may no longer be accessible.\n\n## Opt-in to allow use of Preview models, including Preview models with Data Retention\n\nPreview models are the latest models that are available for exploration and testing but aren't recommended for production use. Review the [limitations of preview models](manage-preview-ai-models) before enabling use of Anthropic Preview models with Data Retention for your tenant.\n\nAnthropic Preview models with Data Retention are advanced models (such as [Claude Fable 5 and Claude Mythos 5](https://support.claude.com/articles/15425996-data-retention-practices-for-mythos-class-models)) that require data retention by Anthropic. This means data is stored by Anthropic and not subject to your Microsoft Customer Agreement including commitments in the Product Terms and DPA. Anthropic Preview models with Data Retention are default-off for your tenant, even if Anthropic is enabled as a subprocessor. No users will have access to these models until the tenant admin explicitly opts in to enable use of Preview models with Data Retention. Use of Anthropic Preview models with Data Retention is subject to acceptance of Anthropic's [Commercial Terms of Service](https://www.anthropic.com/legal/commercial-terms) and Anthropic's [Data Protection Addendum](https://www.anthropic.com/legal/data-processing-addendum). The tenant admin must also choose which users (all or specific users or groups) may access and use these models.\n\nWhen users within your organization are granted access to Preview models with Data Retention, users will also be able to identify in the product experience that these models require data retention by Anthropic.\n\nWhen you enable use of Anthropic Preview models with Data Retention, Anthropic retains data as explained in Anthropic's commercial data retention policy: [Data retention for Mythos-class models](https://support.claude.com/en/articles/15425996-data-retention-practices-for-mythos-class-models) This means that Anthropic (not Microsoft) stores most inputs and outputs for up to 30 days before deleting them. If Anthropic's trust and safety classifiers identify potential violations of Anthropic's Usage Policy, Anthropic may retain content (inputs and outputs) for up to 2 years and trust and safety classification scores for up to 7 years. Anthropic doesn't use retained data for model training without your express permission. For more information, see [API and data retention - Claude API Docs](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).\n\n## Additional controls for Copilot Studio and Power Platform in the Power Platform Admin Center\n\nOnce enabled in the Microsoft 365 admin center, additional admin controls are available in the Microsoft Power Platform admin center (PPAC) to allow Anthropic to be used in Copilot Studio and Power Platform. For more information, see [Allow external large language models (LLMs) for generative responses](/en-us/power-platform/admin/allow-llm-generative-responses).\n\n## Disable connection to Anthropic's models\n\nIf your organization is in a region that has Anthropic as a subprocessor set to **On** by default, you can opt-out to restrict Anthropic models in the Microsoft 365 admin center. You must be a member of the global administrator role to perform this task. For more information, see [About admin roles](/en-us/entra/identity/role-based-access-control/permissions-reference#global-administrator) and [Global administrator](/en-us/entra/identity/role-based-access-control/permissions-reference).\n\n- Go to the Microsoft 365 admin center and select\n**Copilot**->** Settings**. - On the\n**User access** page, select**AI providers operating as Microsoft subprocessors**. - On the\n**AI providers operating as Microsoft subprocessors** page, under**Available subprocessors for your organization**, select** Disable Anthropic as a Microsoft subprocessor**.\n\nOnce you disable Anthropic as an AI subprocessor, users won't have the option to use Anthropic's AI models. You can choose to enable Anthropic models at a later date if desired.\n\n## Deprecation of legacy Anthropic admin toggle\n\nThe legacy Anthropic toggle to opt in to Anthropic's separate commercial terms and data processing agreement has been deprecated and replaced by this new Anthropic as a subprocessor admin toggle.\n\n- December 8, 2025: New administrator toggle for Anthropic models appears in the Microsoft 365 admin center. For most commercial cloud customers, it will be enabled\n**On** by default (except customers in EU/EFTA and UK will be**Off** by default).[Some customer and region exclusions apply](#opt-in-regions-and-exclusions). - January 7, 2026: Anthropic as a Microsoft subprocessor becomes enabled in your organization. The legacy admin toggle to opt-in to Anthropic's commercial terms and data processing agreement is deprecated.\n- If you previously opted in to the legacy Anthropic toggle, and you're in a region where the new toggle is set to\n**off** by default, you need to opt in to the new subprocessor toggle to use Anthropic's models.\n\n- If you previously opted in to the legacy Anthropic toggle, and you're in a region where the new toggle is set to", "url": "https://wpnews.pro/news/anthropic-models-in-microsoft-online-services", "canonical_source": "https://learn.microsoft.com/en-us/microsoft-365/copilot/connect-to-ai-subprocessor", "published_at": "2026-06-14 17:47:04+00:00", "updated_at": "2026-06-14 18:12:06.078120+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-products", "ai-policy", "ai-safety", "ai-infrastructure"], "entities": ["Microsoft", "Anthropic", "Microsoft 365 Copilot", "Copilot Studio", "Power Platform", "Microsoft Online Services", "EU Data Boundary"], "alternates": {"html": "https://wpnews.pro/news/anthropic-models-in-microsoft-online-services", "markdown": "https://wpnews.pro/news/anthropic-models-in-microsoft-online-services.md", "text": "https://wpnews.pro/news/anthropic-models-in-microsoft-online-services.txt", "jsonld": "https://wpnews.pro/news/anthropic-models-in-microsoft-online-services.jsonld"}}