Imagine an AI agent capable of seamlessly combining knowledge retrieval, action-oriented functionality, and advanced analytics to deliver transformative solutions for businesses and individuals alike. Microsoft Azure AI Foundry Agents make this vision a reality. With cutting-edge tools like Azure AI Search, File Search, Microsoft Fabric, Deep Research, custom function calling, MCP integration, and the Code Interpreter, developers now have the ability to craft agents that not only understand complex information but also act on it effectively.

In this blog, we’ll explore how Knowledge Tools and Action Tools work together to create hyper-intelligent agents, revolutionizing industries from healthcare to finance. Whether you’re a developer, a business leader, or an AI enthusiast, this post will give you a comprehensive look at the possibilities these tools unlock. Let’s dive into the future of AI innovation. 

Knowledge Tools: The Foundation of Intelligence

Knowledge Tools empower agents by giving them the ability to access, understand, and organize complex information. These tools form the "brain" of your AI, ensuring it can retrieve relevant data to make well-informed decisions.

1. Azure AI Search: The Brain of Your Agent

Azure AI Search is the cornerstone of any intelligent AI system. It enables agents to retrieve relevant information from enterprise-specific knowledge bases in real-time, ensuring the answers they provide are both accurate and contextually relevant.

Key Features:

• Natural language search capabilities, allowing users to ask questions conversationally.
• Hybrid search capabilities, supporting keyword, vector, and semantic search modes, including combinations like hybrid (keyword + vector) and hybrid semantic (keyword + vector + semantic).
• Multi-language SDK support available through Python, C#, JavaScript, and REST API for flexible integration.

Use Case: 

A customer support Azure AI Foundry Agent can instantly pull up product documentation or troubleshooting solutions based on user queries, reducing average resolution times and enhancing customer satisfaction dramatically.

2. File Search: Unlocking Hidden Insights

File Search is an essential tool for the search and retrieval of information from uploaded documents stored in various file formats. From PDFs to spreadsheets, this tool ensures no insight is lost amidst vast repositories.

Key Features:

• Retrieval-Augmented Generation (RAG) best practices in place out of the box to parse, chunk, and embed files as well as optimize the user query for search.
• Vector store integration allowing for semantic and contextual search capabilities.
• Secure management of sensitive data where uploaded files are stored in your connected Azure Blob Storage account and vector stores get created using your connected Azure AI Search resource.

Use Case: 

A legal assistant Azure AI Foundry Agent that can sift through uploaded case files to find precedents, critical clauses, or specific statutes in seconds transforming how legal research is conducted.

3. Microsoft Fabric: Weaving Data into Knowledge

Microsoft Fabric acts as the data backbone for Azure AI Foundry Agents. By unifying disparate data sources in Microsoft Fabric, agents can query structured enterprise data—such as lakehouses, warehouses, and Power BI semantic models—using natural language, transforming complex datasets into actionable insights through chat.

Key Features:

• Identity passthrough authorization ensures secure, role-based access to data using the end user’s identity.
• Transforms enterprise data into a Q&A system, allowing users to interact with data via chat.
• Generates enhanced data visualization over complex datasets.

Use Case:

A financial services company uses Azure AI Foundry Agents integrated with Microsoft Fabric to analyze customer transaction data. When a user asks: “What were the top 5 spending categories for Q2 2025?” the agent securely queries the Fabric data agent, accesses the relevant warehouse, and returns a ranked list of categories—complete with supporting data—within seconds.

Action Tools: Turning Knowledge into Power

Action Tools empower agents to act intelligently on the knowledge they possess, turning insights into impactful solutions across various applications.

1. Deep Research: The Sherlock Holmes of AI

Deep Research allows agents to dive deeper into real-time data from the public web, uncovering patterns, correlations, and actionable insights that go beyond surface-level understanding.

Key Features:

• Uses OpenAI’s advanced research model (o3-deep-research) to plan, analyze, and synthesize information from across the web.
• Integrates with the Grounding with Bing Search tool to ensure results are based on high-quality, up-to-date sources.
• Fully documents the answer, source citations, and model's reasoning path, including any clarifications requested during the session.

Use Case:  

A legal research team needs to monitor evolving regulations across multiple jurisdictions. By integrating the Deep Research tool into their internal dashboard, an Azure AI Foundry Agent can automatically scan legal databases and government websites weekly, summarize changes in data privacy laws, and generate a report with citations saving hours of manual research and ensuring the team stays compliant with the latest legal developments.

2. Custom Function Calling: The Connective Tissue

Function calling is a powerful tool that enables agents to interact with APIs, databases, and external systems, making them not just intelligent but also functional.

Key Features:

• Developers define the logic and parameters of each function, enabling tailored responses and actions.
• Agents can call functions by returning the function name and arguments based on user input and predefined schemas.
• Enables agents to fetch live data (e.g., weather, stock prices) or perform tasks (e.g., send emails, update databases) dynamically.

Use Case:  

A health care assistant agent in Azure AI Foundry is equipped with a custom function called  scheduleAppointment. A clinician asks: “Schedule a follow-up appointment for John Doe next Tuesday at 10 AM.” The agent calls scheduleAppointment(patientId, dateTime) and confirms the booking with the hospital system.

3. MCP: The Toolbox

The MCP (Model-Context Protocol) tool exponentially extends the capabilities of the agent by allowing it to access tools hosted on remote MCP servers. The possibilities are limitless with the tools developers and organizations are making available through MCP.

Key Features:

• MCP servers provide tools and contextual data that can perform tasks extending its capabilities beyond built-in or custom-built functions.
• Supports custom headers, so you can connect to the MCP servers by using the authentication schemas that they require.
• MCP is not limited to Microsoft services—integrate any MCP-compatible tool, including those hosted by third parties.

Use Case:  

A software development team creates an Azure AI Foundry Agent that utilizes the Azure DevOps MCP server and its tools to manage work items, check build statuses, query repositories, handle project management tasks, and more.

4. Code Interpreter: The Developer’s Best Friend

The Code Interpreter simplifies complex tasks in coding, mathematics, and data analysis, enabling agents to write and execute Python code in a secure, sandboxed environment—all triggered by natural language prompts.

Key Features:

• Real-time code execution for immediate results.
• Agents can safely run Python code in isolated sessions, minimizing security risks.
• Iterative problem-solving workflows to refine solutions continuously. If code fails, the agent can automatically revise and retry until successful execution.

Use Case: 

A data analysis Azure AI Foundry Agent with a natural language query and a dataset could clean, process, and visualize the data automatically, saving data scientists hours of manual effort and reducing errors.

Bringing It All Together

The true power of Azure AI Foundry Agents lies in how Knowledge Tools and Action Tools work in harmony. Imagine an AI-powered business consultant:

• Azure AI Search and File Search retrieve relevant market data from established stores of enterprise knowledge.
• Microsoft Fabric integrates insights from multiple departments’ structured data.
• Deep Research identifies trends and opportunities from the web.
• Function Calling interacts with your custom CRM systems to provide personalized recommendations.
• Using MCP, you’re interacting with a stock market data MCP server that gives you easy access to tools providing current and historical company stock market data, cash flow statements, and balance sheets.
• Code Interpreter analyzes financial models and generates actionable strategies based on your goals and data.

The result? A single agent capable of transforming business decision-making processes, offering unparalleled efficiency and intelligence.

More Than Just Tools

Microsoft Azure AI Foundry Agents are more than just tools—they’re the masterminds behind a new era of intelligent innovation. By seamlessly integrating Knowledge Tools with Action Tools, developers can craft AI agents that don’t just respond—they revolutionize. From transforming customer service to advancing healthcare and streamlining data analysis, these agents unlock a world of limitless possibilities.

Take the Next Step:  

Ready to build your own AI agent? Dive into the Azure AI Foundry documentation and start crafting the future today. For more about how Spyglass MTG can help with unleashing the power of Azure AI Foundry Agents, contact us today.