Merge pull request #53688 from ivorb/genai-agents-updates

Genai agents updates

Author: v-regandowner

learn-pr/wwl-data-ai/develop-ai-agents-azure-vs-code/includes/11-summary.md

@@ -1,8 +1,8 @@
 You've learned how to develop AI agents using Microsoft Foundry Agent Service through both the Foundry portal and Visual Studio Code. These flexible development approaches enable you to create intelligent automation that handles complex tasks, accesses real-time data, and integrates seamlessly with your applications.
 
-Throughout this module, you discovered what AI agents are and why they're valuable for automating workflows and enhancing decision-making. You explored the key features of Microsoft Foundry Agent Service, including automatic tool calling, secure data management, and enterprise-grade security. You learned to build agents using both the visual Foundry portal and the developer-focused VS Code extension, understanding when each approach fits best.
+Throughout this module, you discovered what AI agents are and why they're valuable for automating workflows and enhancing decision-making. You explored the key features of Microsoft Foundry Agent Service, including the two primary agent types — declarative agents (configured through visual designers and YAML) and hosted agents (created and deployed through code). You learned to build agents using both the visual Foundry portal and the developer-focused VS Code extension, understanding when each approach fits best.
 
-You configured agents with clear instructions that define their behavior and personality. You extended agent capabilities by adding tools like Code Interpreter, File Search, and Bing Search, transforming simple chat interfaces into sophisticated automation systems. You tested agents using integrated playgrounds, deployed them to production environments, and generated integration code to connect agents with your applications.
+You configured agents with clear instructions that define their behavior and personality. You extended agent capabilities using the tool catalog, which provides built-in tools like Code Interpreter, File Search, Bing Web Search, Azure AI Search, and many more, transforming simple chat interfaces into sophisticated automation systems. You tested agents using integrated playgrounds, deployed them to production environments, and generated integration code to connect agents with your applications.
 
 With these skills, you can create AI agents that automate routine tasks, provide intelligent assistance, and integrate with your existing systems. Whether you're building customer service automation, developer productivity tools, or specialized research assistants, Microsoft Foundry Agent Service provides the foundation for reliable, scalable AI solutions.
 
@@ -14,14 +14,3 @@ Continue building your AI agent expertise with these resources:
 - [Microsoft Foundry documentation](/azure/ai-foundry/) - Explore detailed guides and API references
 - [GitHub Copilot fundamentals](/training/paths/copilot/) - Learn about AI agent applications in software development
 
-## Clean up resources
-
-If you created Azure resources for exercises in this module and no longer need them:
-
-1. Navigate to the Azure portal at https://portal.azure.com
-1. Locate your resource group containing the AI Foundry resources
-1. Select **Delete resource group**
-1. Confirm deletion by typing the resource group name
-1. Select **Delete** to remove all resources
-
-Deleting unused resources prevents unnecessary charges to your Azure subscription.

learn-pr/wwl-data-ai/develop-ai-agents-azure-vs-code/includes/2-understand-ai-agents-foundry.md

@@ -77,22 +77,37 @@ Microsoft Foundry Agent Service is a fully managed service designed to empower d
 
 Previously, creating agent-like experiences required significant coding effort using standard APIs. Microsoft Foundry Agent Service handles the complexity through a streamlined interface, enabling you to build agents via the Foundry portal or in your own applications with fewer than 50 lines of code.
 
+### Agent types
+
+Microsoft Foundry supports two primary types of agents:
+
+**Declarative agents** - Agents defined through configuration rather than code. Declarative agents come in two forms:
+
+- **Prompt-based agents** - A single agent configured with a model, instructions, tools, and prompts. This is the most common type and the focus of this module.
+- **Workflow agents** - Multi-agent orchestrations defined in YAML, enabling complex scenarios where multiple agents collaborate to complete tasks.
+
+**Hosted agents** - Containerized agents that are created and deployed in code, then hosted by the Foundry platform. Hosted agents give you full control over agent logic and execution while the platform manages infrastructure.
+
+Understanding these agent types helps you choose the right approach for your scenarios. This module focuses primarily on declarative prompt-based agents, which provide the most accessible path to getting started.
+
 ### Key features of Microsoft Foundry Agent Service
 
 The service offers several powerful capabilities:
 
 **Automatic tool calling** - The service handles the entire tool-calling lifecycle, including running the model, invoking tools, and returning results. This eliminates complex integration code.
 
-**Securely managed data** - Conversation states are securely managed with `conversations`, removing the need for manual state management.
+**Securely managed data** - Conversation states are securely managed through the Responses API, removing the need for manual state management.
 
-**Out-of-the-box tools** - Built-in tools support file retrieval, code interpretation, and interaction with data sources like Bing, Azure AI Search, and Azure Functions.
+**Extensive tool catalog** - A rich set of built-in and community tools extends agent capabilities beyond text generation, including code execution, file search, web search, and integrations with Azure services and external APIs.
 
 **Model selection** - Choose from [various AI models](/azure/ai-foundry/agents/concepts/model-region-support?tabs=global-standard#available-models) to match your performance and cost requirements.
 
-**Enterprise-grade security** - The service ensures data privacy and compliance with secure data handling and keyless authentication.
+**Enterprise-grade security** - The service ensures data privacy and compliance with secure data handling, keyless authentication, and built-in content safety filters.
 
 **Customizable storage solutions** - Use platform-managed storage or bring your own Azure Blob storage for full visibility and control.
 
+**Observability and tracing** - Built-in monitoring capabilities help you track agent behavior, debug issues, and optimize performance in production.
+
 These features provide a streamlined and secure way to build and deploy AI agents compared to developing with the Inference API directly.
 
 > [!NOTE]

learn-pr/wwl-data-ai/develop-ai-agents-azure-vs-code/includes/3-explore-development-approaches.md

@@ -22,21 +22,27 @@ The Microsoft Foundry extension for Visual Studio Code brings enterprise-grade A
 
 ### Key capabilities of the VS Code extension
 
-The extension provides comprehensive agent development features:
+The extension organizes its features into three main sections:
 
-**Agent discovery and management** - Browse, create, and manage AI agents within your Microsoft Foundry projects directly from VS Code.
+**Resources** - Browse and manage your Foundry project assets directly from VS Code, including:
+- **Deployed models** - View and manage model deployments
+- **Declarative agents** - View and configure prompt-based and workflow agents
+- **Hosted agents** - View and manage containerized, code-deployed agents
+- **Connections** - Manage connections to external services
+- **Vector stores** - Organize document collections for File Search
 
-**Visual agent designer** - Use an intuitive interface to configure agent instructions, tools, and capabilities without leaving your editor.
+**Tools** - Access development and testing capabilities:
+- **Model Catalog** - Browse and deploy models from the catalog
+- **Model Playground** - Experiment with models directly
+- **Agent Playgrounds** - Test agents using remote or local playgrounds
+- **Local Visualizer** - Debug and visualize agent behavior locally
+- **Deploy Hosted Agents** - Deploy containerized agents to production
 
-**Integrated testing** - Test agents in real-time using the built-in playground without switching contexts or opening additional applications.
+**Help and Feedback** - Access documentation and support resources.
 
-**Code generation** - Generate sample integration code to connect agents with your applications, accelerating the path from prototype to production.
+The extension also provides a visual **Agent Designer** for configuring agent properties, integrated **code generation** for application integration, and direct **YAML configuration** editing for precise control.
 
-**Deployment pipeline** - Deploy agents directly to Microsoft Foundry for production use through streamlined commands.
-
-**YAML configuration** - Access and edit agent configuration files directly, providing both visual and code-based workflows.
-
-![Screenshot of an agent details in the Microsoft Foundry VS Code extension.](../media/vs-code-agent.png)
+![Screenshot of the Microsoft Foundry VS Code extension interface showing the Resources, Tools, and Help and Feedback sections.](../media/vs-code-agent.png)
 
 ### When to use Visual Studio Code
 
@@ -55,8 +61,8 @@ The extension installs directly from the Visual Studio Code Marketplace and conn
 Regardless of your chosen approach, agent development follows a consistent pattern:
 
 1. **Connect** to your Microsoft Foundry project
-1. **Create** an AI agent with a descriptive name and purpose
-1. **Configure** agent instructions defining its behavior and capabilities
+1. **Create** an AI agent in the Foundry portal with a descriptive name and purpose
+1. **Configure** agent instructions defining its behavior and capabilities (in the portal or VS Code)
 1. **Add tools** to extend what the agent can do
 1. **Test** the agent using integrated playgrounds
 1. **Iterate** on the design based on test results
@@ -72,7 +78,7 @@ Both development approaches require the same underlying Azure resources. To deve
 - **Microsoft Foundry project** - Organizes your agents, models, and related assets in one place
 - **Model deployments** - Deployed AI models (such as GPT-4.1 or Claude Sonnet 4.6) that power your agents
 
-When you create a Microsoft Foundry project, the necessary infrastructure is provisioned automatically. As you add capabilities to your agents, such as File Search or custom tools, the service seamlessly integrates any required supporting services behind the scenes. If you choose to extend the capabilities of your agent even further, for example with Foundry IQ, you may need to deploy some addition Azure services.
+When you create a Microsoft Foundry project, the necessary infrastructure is provisioned automatically. As you add capabilities to your agents, such as File Search or custom tools, the service seamlessly integrates any required supporting services behind the scenes. If you choose to extend the capabilities of your agent even further, for example with Foundry IQ, you may need to deploy some additional Azure services.
 
 ### Optional Azure services
 

learn-pr/wwl-data-ai/develop-ai-agents-azure-vs-code/includes/4-build-agent-azure-portal.md

@@ -6,7 +6,7 @@ The Foundry portal streamlines agent creation through an intuitive interface:
 
 1. **Navigate to Microsoft Foundry** at https://ai.azure.com and sign in with your Azure credentials
 1. **Select your project** from the list of available projects, or create a new one
-1. **Navigate to Agents** in the left navigation menu
+1. **Select Build > Agents** in the left navigation menu
 1. **Select Create** to start building a new agent
 1. **Enter agent details**:
    - **Name**: Provide a descriptive name for your agent
@@ -29,15 +29,13 @@ To test your agent, select the Playground tab and start a conversation. The play
 
 ## Adding basic tools
 
-Before deployment, you can enhance your agent with basic capabilities:
+Before deployment, you can enhance your agent with tools from the tool catalog in the **Tools** section of the agent configuration (also accessible via **Build > Tools** in the portal). The catalog organizes tools into three categories:
 
-**Code Interpreter** - Enables the agent to write and execute Python code for calculations, data analysis, and problem-solving.
+- **Configured** - Built-in tools ready to use immediately, such as Code Interpreter and File Search
+- **Catalog** - Additional tools you can add, including Bing Web Search, Azure AI Search, SharePoint, and more
+- **Custom** - Your own tools added through OpenAPI specifications or MCP servers
 
-**File Search** - Allows the agent to search through uploaded documents to answer questions.
-
-**Bing Search** - Connects your agent to real-time internet information.
-
-You add these tools through the Tools section in the agent configuration. Tool capabilities are explored in detail in the *Extend agent capabilities* unit later in this module.
+Tool capabilities and configuration are explored in detail in the *Extend agent capabilities* unit later in this module.
 
 ## Deploying your agent
 

learn-pr/wwl-data-ai/develop-ai-agents-azure-vs-code/includes/5-set-up-vs-code.md

@@ -4,7 +4,9 @@ Setting up Visual Studio Code for AI agent development brings enterprise-grade c
 
 The Microsoft Foundry for Visual Studio Code extension provides direct access to Microsoft Foundry Agent Service capabilities. This extension creates an integrated experience for agent development that combines visual design tools with code-based configuration.
 
-![Screenshot of an agent details in the Microsoft Foundry VS Code extension.](../media/vs-code-agent.png)
+The extension organizes its features into three sections: **Resources** (for managing deployed models, declarative agents, hosted agents, connections, and vector stores), **Tools** (for accessing the model catalog, playgrounds, and deployment features), and **Help and Feedback**.
+
+![Screenshot of the Microsoft Foundry VS Code extension interface showing the Resources, Tools, and Help and Feedback sections.](../media/vs-code-agent.png)
 
 ## Installing and configuring the extension
 
@@ -14,7 +16,7 @@ Setting up the Microsoft Foundry extension takes just a few minutes and requires
 
 1. Open Visual Studio Code on your machine
 1. Select **Extensions** from the left pane, or press <kbd>Ctrl</kbd>+<kbd>Shift</kbd>+<kbd>X</kbd> (Windows/Linux) or <kbd>Cmd</kbd>+<kbd>Shift</kbd>+<kbd>X</kbd> (Mac)
-1. Search for **Microsoft Foundry** in the marketplace search box
+1. Search for **Foundry** in the marketplace search box
 1. Select the **Microsoft Foundry** extension from the results
 1. Select **Install** to add the extension to VS Code
 1. Wait for installation to complete (status appears in the Extensions panel)
@@ -23,20 +25,20 @@ After installation, the Microsoft Foundry icon appears in the VS Code activity b
 
 ### Connecting to Azure
 
-Before creating agents, connect the extension to your Azure account and project:
+Before working with agents, connect the extension to your Azure account and project:
 
-1. Select the **Microsoft Foundry** icon in the VS Code activity bar
-1. Select **Sign in to Azure** when prompted
-1. Complete authentication in your browser
-1. Return to VS Code after successful authentication
-1. Select your **Azure subscription** from the dropdown
-1. Choose your **Microsoft Foundry project** or create a new one
+1. Select the **Azure** icon in the VS Code activity bar
+1. In the **Azure Resources** pane, sign in to your Azure account if prompted
+1. Expand your **Azure subscription** in the resource tree
+1. Expand the **Foundry** section to see your projects
+1. Right-click your **Microsoft Foundry project**
+1. Select **Open in Foundry Extension**
 
-The extension now displays your project resources, including existing agents, model deployments, and other assets.
+The extension now displays your project resources in the Microsoft Foundry panel, including existing agents, model deployments, connections, and vector stores.
 
 ## Preparing for agent development
 
-Before creating your first agent in VS Code, ensure you have the necessary resources deployed.
+Before working with agents in VS Code, ensure you have the necessary resources deployed.
 
 ### Deploying a model
 
@@ -52,24 +54,13 @@ Agents require deployed AI models to function. If you don't have a model deploym
    - **Capacity settings**: Configure throughput based on your needs
 1. Select **Deploy** and wait for deployment to complete
 
-The deployed model becomes available in dropdown menus when you create or configure agents.
-
-## Creating your first agent in VS Code
-
-With the extension configured and a model deployed, you're ready to create an agent.
+The deployed model becomes available in dropdown menus when you configure agents.
 
-### Agent creation process
-
-1. In the **Microsoft Foundry** extension, navigate to the **Resources** section
-1. Find the **Agents** subsection
-1. Select the **+** (plus) icon next to Agents
-1. The extension opens two views:
-   - **Agent Designer**: A visual interface for configuration
-   - **YAML file**: The agent's configuration in code format
+## Working with agents in VS Code
 
-![Screenshot of creating a new agent in the Microsoft Foundry VS Code extension.](../media/vs-code-create-new-agent.png)
+Agents are often created in the Foundry portal (as described in the previous unit) and then managed and configured in VS Code through the extension. Once you've created an agent in the portal, it appears automatically in the extension's **Resources** section.
 
-Both views remain synchronized; changes in one automatically update the other. This dual approach lets you work with whichever format feels more natural for your current task.
+Changes to agents in VS Code can be saved directly to Foundry, so you can work with your agent across platforms.
 
 ## Managing multiple agents
 

learn-pr/wwl-data-ai/develop-ai-agents-azure-vs-code/includes/6-configure-manage-agents.md

@@ -1,4 +1,7 @@
-Once you create an agent in Visual Studio Code, the real work begins — configuring its behavior, instructions, and properties to match your requirements. The Microsoft Foundry extension provides comprehensive configuration options through both visual interfaces and direct file editing, giving you flexibility in how you work.
+Once you have a declarative agent (created in the Foundry portal or through the SDK), the real work begins — configuring its behavior, instructions, and properties to match your requirements. The Microsoft Foundry VS Code extension provides comprehensive configuration options through both the visual Agent Designer and direct YAML file editing, giving you flexibility in how you work.
+
+> [!NOTE]
+> The configuration workflow described in this unit applies to **declarative prompt-based agents**. Hosted agents are configured through code, and workflow agents use a different YAML schema for multi-agent orchestration.
 
 ## Configuring agent properties
 
@@ -30,7 +33,7 @@ These settings appear in both the Designer interface and the YAML file, remainin
 
 ## Understanding the agent YAML structure
 
-The YAML file contains all your agent's configuration in a structured, readable format. Understanding this structure helps you make precise changes and work efficiently when the visual interface isn't the best fit.
+The YAML file contains all your declarative agent's configuration in a structured, readable format. Understanding this structure helps you make precise changes and work efficiently when the visual interface isn't the best fit.
 
 ### Complete YAML example