Microsoft Foundry and AI services
Microsoft Foundry is Azure's platform for models, agents and AI tools under one resource, with tracing, evaluation and governance. It covers Foundry Models (including Azure OpenAI), Foundry Agent Service, Foundry Tools (formerly Azure AI Services), Foundry IQ and Azure Machine Learning.
6 technology profiles in this area
Overview
Microsoft Foundry, formerly Azure AI Studio and then Azure AI Foundry, puts agents, models and tools under one Azure resource with shared access control, networking and policy. Microsoft states the catalog holds more than 10,000 models from Microsoft, OpenAI, Anthropic, Meta and others. What used to be Azure AI Services is now called Foundry Tools.
The new Foundry portal is generally available for core scenarios, but several parts remain in preview, including agent memory, monitoring and portal access to Foundry IQ. Standalone Azure OpenAI resources and hub-based projects stay in the Foundry (classic) portal until they are upgraded.
There is no platform fee: you pay for the individual services you use. Foundry Local, a separate runtime, runs models on a user's own device with no Azure subscription.
If a prebuilt API such as Speech, Translator or Document Intelligence solves the problem, Microsoft's architecture guide says to consider Foundry Tools before building a custom model.12345
Problems it addresses
Choosing among many models6
The catalog spans models sold by Azure and partner models. Agents in Foundry can switch models without changes to agent code, which keeps early choices reversible.
Grounding answers in company data5
Foundry IQ is a managed knowledge layer, built on Azure AI Search, that returns permission-aware answers with citations from sources such as Blob Storage, OneLake, Azure SQL and SharePoint.
Knowing what is production-ready2
GA covers core scenarios only. Microsoft lists memory, monitoring and agent guardrails as preview and advises against treating preview features as production dependencies.
Moving from Azure OpenAI1
Existing Azure OpenAI resources can be upgraded to a Foundry resource while keeping the endpoint, API keys and state.
Retired AI APIs7
Anomaly Detector, Content Moderator, Language Understanding (LUIS), Metrics Advisor, Personalizer and QnA Maker are retired and should not be used for new apps.
A typical workflow
Make a first model call1
Create a Foundry project and send a prompt from the SDK in Python, C#, JavaScript or Java.
Build a prompt agent6
Set instructions, a model and tools in the portal or SDK; Foundry hosts and runs the agent with no containers to manage.
Add knowledge and tools6
Attach a toolbox with web search, file search, code interpreter or MCP servers, and ground answers on an Azure AI Search index.
Evaluate and trace2
Use tracing, evaluations and red teaming, which are generally available, before anyone outside the team uses the agent.
Publish6
Promote the agent to a stable endpoint and share it in Microsoft Teams and Microsoft Copilot.
Technologies
- Cloud service
Azure AI Search
Azure AI Search is a managed search and retrieval service that connects company content to apps, agents and language models. It supports full-text, vector, hybrid and multimodal queries and an agentic retrieval engine that powers Foundry IQ knowledge bases. It was formerly Azure Cognitive Search.
- Data
- AI
- Security
- Cloud service
Azure Machine Learning
Azure Machine Learning is Azure's managed service for training, deploying and operating custom machine learning models. It covers notebooks, AutoML, distributed GPU training, pipelines, MLflow tracking, model registries and managed online and batch endpoints, with hybrid options through Azure Arc.
- AI
- Data
- Infrastructure
- Cloud service
Microsoft Foundry Agent Service
Foundry Agent Service is the managed runtime in Microsoft Foundry for building, hosting and scaling AI agents. Teams configure prompt agents with no code, run voice agents, or deploy their own agent code as hosted containers, with Entra identity, tools, tracing and publishing to Teams and Microsoft Copilot.
- Agents
- AI
- Security
- Cloud service
Microsoft Foundry Models
Foundry Models is the model catalog and hosting service inside Microsoft Foundry. It offers models sold and supported by Azure, including all Azure OpenAI models, plus a much larger set from partners and the community, deployable as serverless APIs or on managed compute.
- AI
- Infrastructure
- Cloud service
Foundry Tools
Foundry Tools is Microsoft's family of prebuilt AI APIs for speech, translation, language, document and content understanding, vision, search and content safety. Developers call them through REST APIs and SDKs without training a model. It was formerly called Azure AI Services and, before that, Cognitive Services.
- AI
- Developer
- Platform
Microsoft Foundry
Microsoft Foundry is Azure's platform for building generative AI apps and agents: a model catalog, an agent runtime, prebuilt AI tools, knowledge retrieval and governance under one Azure resource with shared access control, networking and policy. It was formerly Azure AI Studio, then Azure AI Foundry.
- AI
- Agents
- Developer
- Security
No technology in this area matches. Clear the filter to see all.
Use cases
Documented implementations
Air India
In productionAir India: AI.g customer service agent on Azure OpenAI
about 40,000 (2026), up from an average of 30,000 (2024) (measured)
Travel and hospitality, India
C.H. Robinson
In productionC.H. Robinson: AI agents for freight quotes, orders and tracking
from hours to an average of 32 seconds per quote (measured)
Logistics and transportation, United States (global operations)
Chow Tai Fook Jewellery Group
In productionChow Tai Fook: 400+ AI agents for jewelry retail staff
over 400 agents (measured)
Retail and consumer goods, Hong Kong (story dateline); store network across several markets
Direct pojišťovna (part of the Direct group)
In productionDirect pojišťovna: AI agents for windshield claims
15 minutes to about 2 minutes (87% less) (measured)
Financial services, Czechia
Premier League
In productionPremier League: an AI fan companion built on Microsoft Foundry
up approximately 20% (measured)
Media and entertainment, United Kingdom (global audience)
Woven by Toyota, Inc.
PilotWoven by Toyota: AI agents that fix MISRA code violations
81.5% (measured)
Automotive, Japan
UK Power Networks
In productionUK Power Networks: Copilot from pilot to company-wide use
96% (measured)
Energy and utilities, United Kingdom (London, South East and East of England)
Make-A-Wish America (national office and chapters)
In productionMake-A-Wish America: unified data, Copilot and an IT help bot
roughly two-thirds of staff (measured)
Nonprofit, United States
EY
ScalingEY: Dynamics 365 Sales with agents, and Copilot across the firm
85,000 in 18 months; goal 100,000 (measured)
Professional services, Global
Agent Platform Chooser
Where to build your AI agent in the Microsoft stack, or whether you need one at all.
Next steps
Check whether a Foundry Tools API already covers your task before training or fine-tuning anything.
If you run standalone Azure OpenAI resources, plan the upgrade to a Foundry resource.
Restrict preview features in production projects until your team has reviewed them.
Use the Agent Platform Chooser to decide between prompt agents, hosted agents and Copilot Studio.
Connected areas
Sources
- What is Microsoft Foundry? (opens in a new tab)
- Microsoft Foundry portal general availability overview (opens in a new tab)
- Microsoft Foundry product page (opens in a new tab)
- What is Foundry Local? (opens in a new tab)
- AI and machine learning products (Azure Architecture Center) (opens in a new tab)
- What is Microsoft Foundry Agent Service? (opens in a new tab)
- What are Foundry Tools? (opens in a new tab)
Thank you. Your correction was sent.
The editors check it against the sources. If you left an email address, they may reply about it.