Microsoft Foundry: Navigating the New AI Development Paradigm

Microsoft has undergone a significant transformation in its AI development ecosystem, evolving from Azure AI Studio to Azure AI Foundry and now to Microsoft Foundry. This evolution extends beyond mere rebranding, representing a fundamental architectural shift that impacts how organizations approach AI development, deployment, and governance.

Architectural Transformation

The most significant change is the coexistence of two distinct portal experiences at the same URL (ai.azure.com), accessible through a simple toggle in the banner. This dual-portal approach allows organizations to choose between:

• Foundry (Classic): The original experience supporting multiple resource types including Azure OpenAI, hub-based projects, and legacy Azure Machine Learning workloads
• Foundry (New): A streamlined, agent-first experience specifically designed for building and managing multi-agent applications

This architectural pivot reflects Microsoft's strategic direction toward agentic AI systems rather than simple chatbot interfaces. The new platform represents a generational leap in capabilities, particularly in multi-agent orchestration, long-term memory, and enterprise knowledge integration.

Resource Model Evolution

The resource architecture has been significantly simplified in the new experience:

Classic Approach

Hub-based projects operated under the Microsoft.MachineLearningServices provider with a complex hierarchy:

• Azure AI Hub (parent resource)
• Required sibling resources: Storage Account, Key Vault (mandatory)
• Projects as child resources

New Approach

Microsoft Foundry consolidates under the Microsoft.CognitiveServices provider:

• Foundry Resource (Microsoft.CognitiveServices/account, kind: AIServices)
• Foundry Projects (Microsoft.CognitiveServices/account/project)
• No mandatory sibling resources by default

This consolidation enables unified governance, networking, and policy enforcement across Azure's AI services, including Speech, Vision, and Language services. The shift also provides clearer separation between control plane (IT administration) and data plane (developer activities).

Agent Capabilities: The Core Differentiator

The most significant advancement in Microsoft Foundry is the Agent Service, which has evolved from preview to general availability with substantial enhancements:

Multi-Agent Orchestration

The new portal introduces sophisticated workflow capabilities:

• Sequential workflows with defined agent handoffs
• Group chat patterns with dynamic control passing
• Human-in-the-loop interactions for approval steps
• Visual YAML editor with real-time synchronization
• Power Fx integration for conditional logic and data transformations
• Version control for all workflow changes

Agent Types

Microsoft Foundry supports three distinct agent types:

1. Prompt Agents: Declaratively defined LLM-backed agents
2. Hosted Agents: Containerized agents running custom code
3. Workflow Agents: YAML-based orchestrations of multiple agents

Memory and Knowledge Integration

New capabilities in the Foundry experience include:

• Long-term Memory: Persistent user profile memory and chat summaries
• Foundry IQ: A managed, multi-source knowledge base with enterprise-grade permission awareness
• Tool Catalog: Access to 1,400+ tools including Azure AI Search, Code Interpreter, and custom endpoints

Migration Considerations

Organizations face strategic decisions when transitioning between portal experiences:

What Transfers Automatically

• Model deployments
• Data files
• Fine-tuned models
• Assistants
• Vector stores

What Requires Recreation

• Preview agent state (threads, messages, files)
• Open-source model deployments
• Hub project access
• Connections for tools and data sources

When to Migrate

Organizations should consider moving to the new experience when:

• Building agentic applications requiring GA Agent Service
• Needing multi-agent orchestration capabilities
• Requiring Tool Catalog, Memory, or Foundry IQ features
• Planning to publish agents to Microsoft 365/Teams
• Starting greenfield projects with simplified governance

Organizations should maintain the classic experience when:

• Dependent on Prompt Flow for orchestration
• Deploying open-source models on managed compute
• Requiring Azure Machine Learning features (AutoML, ML Pipelines)
• Needing fully isolated end-to-end networking in the portal
• Having significant hub-based project investments

Business Impact

The shift to Microsoft Foundry represents a strategic realignment in how organizations approach AI development:

Governance and Security

The new model provides clearer separation between IT control and developer autonomy. IT administrators can establish governance at the resource level while developers maintain isolated project workspaces. This separation of concerns reduces administrative overhead while maintaining security compliance.

Development Velocity

The expanded tool catalog and multi-agent orchestration capabilities significantly accelerate development cycles. Organizations can build more complex AI systems with greater efficiency, reducing the time from concept to production deployment.

Ecosystem Integration

Microsoft Foundry's integration with Microsoft 365, Teams, and Azure services creates a cohesive enterprise AI ecosystem. The ability to publish agents directly to these platforms enables seamless adoption across organizational workflows.

Future-Proofing

The architecture and capabilities of Microsoft Foundry position organizations for emerging AI trends. The focus on multi-agent systems, long-term memory, and enterprise knowledge integration aligns with the evolution of AI from simple chatbots to sophisticated autonomous assistants.

Strategic Recommendations

For organizations navigating this transition:

1. Assess Current Workloads: Inventory existing Azure AI resources to determine migration urgency
2. Evaluate New Capabilities: Identify features in the new portal that could provide competitive advantage
3. Plan Dual-Portal Strategy: Prepare for a period where both experiences coexist
4. Invest in Training: Development teams will need upskilling on the new agent paradigms
5. Governance Alignment: Update policies and procedures to reflect the new resource model

Microsoft Foundry represents not just a platform update but a fundamental shift in how organizations approach AI development. The dual-existence of classic and new portals provides a transition path while enabling organizations to adopt cutting-edge capabilities at their own pace. As Microsoft continues to enhance the new experience, organizations should develop strategies to leverage these advancements while maintaining operational continuity during the transition period.

For more detailed information on Microsoft Foundry capabilities and migration paths, refer to the official documentation and architecture overview.