Skip to main content
Non-English translations are provided for convenience only. Please consult the EN-US version of this document for the definitive version.
Foundry Agent Service is a fully managed service designed to empower developers to securely build, deploy, and scale high-quality, and extensible AI agents without needing to manage the underlying compute and storage resources. Agent Service integrates models, tools and technology and enables you to extend agents with knowledge from connected sources (such as Bing Search, SharePoint, Fabric, Azure Blob storage, and licensed data) and actions using tools such as Azure Logic Apps, Azure Functions, OpenAPI 3.0 specified tools and Code Interpreter
This article provides details regarding how data provided by you to Agent Service is processed, used, and stored. Please also see the Microsoft Products and Services Data Protection Addendum, which governs data processing by Agent Service (but may not necessarily apply to external tools or services with which Agent Service interacts, which are subject to their own data processing terms).
Your prompts (inputs) and completions (outputs) and your data:
  • are NOT available to other customers.
  • are NOT available to OpenAI, Meta, Cohere, or Mistral.
  • are NOT used to improve OpenAI, Meta, Cohere, or Mistral models.
When you use Agent Service with tools that retrieve data from external sources or services (such as the Grounding with Bing Search tool), the terms (including data processing terms) for those services apply to any data processed by those services. For example, the Grounding with Bing Search tool is subject to separate data collection and privacy terms (see Terms of Use for Grounding with Bing Search in Azure AI Agents Service), and the services are Microsoft-as-controller services and thereby excluded from the Microsoft Products and Services Data Protection Addendum.

​
What data does Agent Service process?

Agent Service processes the following types of data:
  • Prompts and generated content. Prompts are submitted by the user, and content is generated by the service, via the completions, chat completions, images and audio operations of models with which Agent Service interacts.
  • Uploaded data. You can provide your own data for use with certain tools in Agent Service (for example, File Search, Code Interpreter, Azure AI Search) using your own Azure Storage account or a configured data store.
  • External Data. When you use the Grounding with Bing Search tool or tools that support function calling, the service processes and stores the outputs of these tools.
Data for stateful entities. When you use Threads, Messages and Runs, the service will create a data store to persist message history and other content, in accordance with how you configure the feature and in keeping with applicable privacy, security, and compliance commitments. When you use the MCP tool, the data you pass in headers is only available for the current Run and will not persist. Augmented data included with or via prompts. When using data associated with stateful entities, the service retrieves relevant data from your configured data store and augments the prompt to produce generations that are grounded with your data. Prompts may also be augmented with data retrieved from a source included in the prompt itself, such as a URL.

​
How does Agent Service process data?

The diagram below illustrates how your data is processed. This diagram covers several types of processing:
  1. How Agent Service processes your prompts via inferencing to generate content (including when additional data from a designated data source is added to a prompt using File Search, Code Interpreter, or other tools).
  2. How Agent Service stores data in connection with Messages, Threads, and Runs.
  3. How Agent Service processes data ingested into the service from external tools.
A diagram showing how Agent Service processes data.

​
Model inferencing

Agent Service interacts with the configured model inferencing endpoints you designate. Models (base or fine-tuned) process your input prompts and generate output responses which may be further used or processed by your agent. Data is processed for model inferencing in accordance with the terms that apply to the relevant model. Learn more at Data, privacy, and security for Azure OpenAI Service and Data, privacy, and security for use of models through the model catalog in Foundry portal.

​
Data storage for Agent Service features

Some Agent Service features store data in the service. This data is either uploaded by the customer or is automatically stored in connection with certain stateful entities such as Messages, Threads, and Runs. Data stored for Agent Service:
  • Is stored at rest in the Azure OpenAI resource which is created when you configure Agent Service in the customer’s Azure tenant, within the same geography as the Azure OpenAI resource.
  • Can be double encrypted at rest, by default with Microsoft’s AES-256 encryption and optionally with a customer managed key (except preview features may not support customer managed keys).
  • Can be deleted by the customer at any time.

​
Location of data processing

To use Agent Service, you must create an Azure OpenAI Service resource in which Agent Service will be hosted. If you use features of Agent Service that store data at rest, that data will be stored in the geography where your Azure OpenAI resource is located. When Agent Service interacts with models and tools, the location of processing of data will depend on the configuration of the model inferencing endpoints and the hosting location of the tools.

​
Preventing harmful content generation

To reduce the risk of harmful use of Agent Service, the Service includes content filtering support. The outputs processed by Agent Service will be filtered in accordance with any content filtering that has been applied to the model deployment used by your agent.

​
See also