An isolated workspace inside a Microsoft Foundry resource for one AI solution, holding its own model deployments, agents, knowledge stores, evaluations and monitoring while sharing the resource's security boundary.
A Foundry project is an isolated workspace inside a Microsoft Foundry resource, created for a single AI solution. It holds that solution's own model deployments, agents, knowledge stores, evaluations and monitoring, while sharing the security boundary of the resource it belongs to. In practice, the project is the unit in which a team builds, tests and operates one AI application, separate from other applications that live under the same resource.
The main purpose of a project is separation of concerns. Because each project carries its own deployments, agents, knowledge stores, evaluations and monitoring, one solution can be developed and observed without mixing its assets with those of another. The shared security boundary is the other half of the design: projects inside the same Microsoft Foundry resource do not each define a separate perimeter, but inherit the one set at the resource level. This matters when deciding how many projects to create and where to place them, since the resource determines the security posture and the project determines the scope of the individual solution.
Typical decisions around a Foundry project include how to group solutions, which model deployments a given project needs, and which knowledge stores its agents should use. Setting up a project is also an early step in Microsoft's Agent Framework scenarios. The episode "Microsoft Agent Framework - Simply Explained" lists the creation of an Azure AI Foundry resource and project as part of the setup, followed by deploying the required models and services and configuring an AzureSearch Service. It also describes connecting the Azure Search resource to the Foundry project in order to configure grounding with Bing Search. In that walkthrough, the project is the place where models, search and grounding are wired together before an agent is built.
Because a project ties together many connected services, problems often show up at the project level. The episode "Fix Common Azure AI Foundry Project Failures" examines what happens when Azure AI Foundry does not behave as expected, especially when the Agent Service or deployments start acting up. It describes Azure AI Foundry as meant to feel smooth and almost invisible, connecting OpenAI models, search, storage and the other moving parts behind an AI application, and notes that things can slip, for example when an agent stops communicating with the components it depends on. The same episode touches on building a strong pipeline and unified data access to avoid data versioning problems and support the performance of AI models, and it points to governance and security gaps, such as undefined roles and accountability, as recurring sources of trouble.
The episode "Fix Silent Failures in Automated Customer Journeys" does not centre on Foundry projects, but it lists "Fix Common Azure AI Foundry Project Failures" among related M365.FM episodes, which places the topic alongside broader discussions of automation reliability in the Microsoft ecosystem.
For readers who need the authoritative procedure for creating and configuring a project, the official Microsoft Learn documentation for creating projects in Foundry is the reference. The M365.FM episodes complement it with practical context: what a project is used for in an agent scenario, which supporting services are configured around it, and where failures and governance gaps tend to appear when several connected components have to work together.
Overview compiled with AI from M365.FM episodes and the facts on this page. You can improve it with the edit button below.
Azure AI Foundry is Microsoft's enterprise platform for building, deploying, and managing AI applications and intelligent agents at scale. Instead of stitching together multiple AI services, developers get a unified environment for working with foundation…
Questions marked with AI were written with AI from the facts on this page.
What is a Foundry project? (AI)
A Foundry project is an isolated workspace inside a Microsoft Foundry resource for one AI solution. It contains its own model deployments, agents, knowledge stores, evaluations and monitoring.
How does a Foundry project relate to the Foundry resource's security? (AI)
A Foundry project shares the security boundary of the Microsoft Foundry resource it sits in. The resource defines the security posture, while the project scopes the individual AI solution.
Where does a Foundry project appear in the M365.FM episodes? (AI)
The episode "Microsoft Agent Framework - Simply Explained" lists creating a Foundry resource and project as a setup step. "Fix Common Azure AI Foundry Project Failures" covers what happens when the Agent Service or deployments in Azure AI Foundry misbehave.
How do I get started with Azure AI Foundry?
Create a Foundry project and build one agent. If you get stuck, the built-in AI chatbot can answer platform questions. · From the episode: Azure AI Foundry - Simply Explained
Short info above, community knowledge below: add practical tips, examples or corrections. New contributions are reviewed first; trusted members publish directly.
Experts and companies for this topic
Nobody has listed this topic yet. Add it to your profile.