Secure Model Context Protocol (MCP) Implementation with Azure and Local Servers
Introduction The Model Context Protocol (MCP) enables AI systems to interact with external data sources and tools through a standardized interface. While powerful, MCP can introduce security risks in enterprise environments. This tutorial shows you how to implement MCP securely using local servers, Azure OpenAI with APIM, and proper authentication. Understanding MCP's Security Risks There are a couple of key security concerns to consider before implementing MCP: Data Exfiltration: External MCP servers could expose sensitive data. Unauthorized Access: Third-party services become potential security risks. Loss of Control: Unknown how external services handle your data. Compliance Issues: Difficulty meeting regulatory requirements with external dependencies. The solution? Keep everything local and controlled. Secure Architecture Before we dive into implementation, let's take a look at the overall architecture of our secure MCP solution: This architecture consists of three key components working together: Local MCP Server - Your custom tools run entirely within your local environment, reducing external exposure risks. Azure OpenAI + APIM Gateway - All AI requests are routed through Azure API Management with Microsoft Entra ID authentication, providing enterprise-grade security controls and compliance. Authenticated Proxy - A lightweight proxy service handles token management and request forwarding, ensuring seamless integration. One of the key benefits of this architecture is that no API key is required. Traditional implementations often require storing OpenAI API keys in configuration files, environment variables, or secrets management systems, creating potential security vulnerabilities. This approach uses Azure Managed Identity for backend authentication and Azure CLI credentials for client authentication, meaning no sensitive API keys are ever stored, logged, or exposed in your codebase. For more security, APIM and Azure OpenAI resources can be configured with IP restrictions or network rules to only accept traffic from certain sources. These configurations are available for most Azure resources and provide an additional layer of network-level security. This security-forward approach gives you the full power of MCP's tool integration capabilities while keeping your implementation completely under your control. How to Implement MCP Securely 1. Local MCP Server Implementation Building the MCP Server Let's start by creating a simple MCP server in .NET Core. 1. Create a web application dotnet new web -n McpServer 2.Add MCP packages dotnet add package ModelContextProtocol --prerelease dotnet add package ModelContextProtocol.AspNetCore --prerelease 3. Configure Program.cs var builder = WebApplication.CreateBuilder(args); builder.Services.AddMcpServer() .WithHttpTransport() .WithToolsFromAssembly(); var app = builder.Build(); app.MapMcp(); app.Run(); WithToolsFromAssembly() automatically discovers and registers tools from the current assembly. Look into the C# SDK for other ways to register tools for your use case. 4. Define Tools Now, we can define some tools that our MCP server can expose. here is a simple example for tools that echo input back to the client: using ModelContextProtocol.Server; using System.ComponentModel; namespace Tools; [McpServerToolType] public static class EchoTool { [McpServerTool] [Description("Echoes the input text back to the client in all capital letters.")] public static string EchoCaps(string input) { return new string(input.ToUpperInvariant()); } [McpServerTool] [Description("Echoes the input text back to the client in reverse.")] public static string ReverseEcho(string input) { return new string(input.Reverse().ToArray()); } } Key components of MCP tools are the McpServerToolType class decorator indicating that this class contains MCP tools, and the McpServerTool method decorator with a description that explains what the tool does. Alternative: STDIO Transport If you want to use STDIO transport instead of SSE (implemented here), check out this guide: Build a Model Context Protocol (MCP) Server in C# 2. Create a MCP Client with Cline Now that we have our MCP server set up with tools, we need a client that can discover and invoke these tools. For this implementation, we'll use Cline as our MCP client, configured to work through our secure Azure infrastructure. 1. Install Cline VS Code Extension Install the Cline extension in VS Code. 2. Deploy secure Azure OpenAI Endpoint with APIM Instead of connecting Cline directly to external AI services (which could expose the secure implementation to external bad actors), we will route through Azure API Management (APIM) for enterprise security. With this implementation, all requests go through Microsoft Entra ID and we use managed identity for all authentications. Quick Setup: Deploy the Azure OpenAI with APIM solution. Ensure your Azure OpenAI resources are configured to allow your APIM's managed identity to make calls. The APIM policy below uses managed identity authentication to connect to Azure OpenAI backends. Refer to the Azure OpenAI documentation on managed identity authentication for detailed setup instructions. 3. Configure APIM Policy After deploying APIM, configure the following policy to enable Azure AD token validation, managed identity authentication, and load balancing across multiple OpenAI backends: <!-- Azure API Management Policy for OpenAI Endpoint --> <!-- Implements Azure AD Token validation, managed identity authentication --> <!-- Supports round-robin load balancing across multiple OpenAI backends --> <!-- Requests with 'gpt-5' in the URL are routed to a single backend --> <!-- The client application ID '04b07795-8ddb-461a-bbee-02f9e1bf7b46' is the official Azure CLI app registration --> <!-- This policy allows requests authenticated by Azure CLI (az login) when the required claims are present --> <policies> <inbound> <!-- IP Allow List Fragment (external fragment for client IP restrictions) --> <include-fragment fragment-id="YourCompany-IPAllowList" /> <!-- Azure AD Token Validation for Azure CLI app ID --> <validate-azure-ad-token tenant-id="YOUR-TENANT-ID-HERE" header-name="Authorization" failed-validation-httpcode="401" failed-validation-error-message="Unauthorized. Access token is missing or invalid."> <client-application-ids> <application-id>04b07795-8ddb-461a-bbee-02f9e1bf7b46</application-id> </client-application-ids> <audiences> <audience>api://YOUR-API-AUDIENCE-ID-HERE</audience> </audiences> <required-claims> <claim name="roles" match="any"> <value>YourApp.User</value> </claim> </required-claims> </validate-azure-ad-token> <!-- Acquire Managed Identity access token for backend authentication --> <authentication-managed-identity resource="https://cognitiveservices.azure.com" output-token-variable-name="managed-id-access-token" ignore-error="false" /> <!-- Set Authorization header for backend using the managed identity token --> <set-header name="Authorization" exists-action="override"> <value>@("Bearer " + (string)context.Variables["managed-id-access-token"])</value> </set-header> <!-- Check if URL contains 'gpt-5' and set backend accordingly --> <choose> <when condition="@(context.Request.Url.Path.ToLower().Contains("gpt-5"))"> <set-variable name="selected-backend-url" value="https://your-region1-oai.openai.azure.com/openai" /> </when> <otherwise> <cache-lookup-value key="backend-counter" variable-name="backend-counter" /> <choose> <when condition="@(context.Variables.ContainsKey("backend-counter") == false)"> <set-variable name="backend-counter" value="@(0)" /> </when> </choose> <set-variable name="current-backend-index" value="@((int)context.Variables["backend-counter"] % 7)" /> <choose> <when condition="@((int)context.Variables["current-backend-index"] == 0)"> <set-variable name="selected-backend-url" value="https://your-region1-oai.openai.azure.com/openai" /> </when> <when condition="@((int)context.Variables["current-backend-index"] == 1)"> <set-variable name="selected-backend-url" value="https://your-region2-oai.openai.azure.com/openai" /> </when> <when condition="@((int)context.Variables["current-backend-index"] == 2)"> <set-variable name="selected-backend-url" value="https://your-region3-oai.openai.azure.com/openai" /> </when> <when condition="@((int)context.Variables["current-backend-index"] == 3)"> <set-variable name="selected-backend-url" value="https://your-region4-oai.openai.azure.com/openai" /> </when> <when condition="@((int)context.Variables["current-backend-index"] == 4)"> <set-variable name="selected-backend-url" value="https://your-region5-oai.openai.azure.com/openai" /> </when> <when condition="@((int)context.Variables["current-backend-index"] == 5)"> <set-variable name="selected-backend-url" value="https://your-region6-oai.openai.azure.com/openai" /> </when> <when condition="@((int)context.Variables["current-backend-index"] == 6)"> <set-variable name="selected-backend-url" value="https://your-region7-oai.openai.azure.com/openai" /> </when> </choose> <set-variable name="next-counter" value="@(((int)context.Variables["backend-counter"] + 1) % 1000)" /> <cache-store-value key="backend-counter" value="@((int)context.Variables["next-counter"])" duration="300" /> </otherwise> </choose> <!-- Always set backend service using selected-backend-url variable --> <set-backend-service base-url="@((string)context.Variables["selected-backend-url"])" /> <!-- Inherit any base policies defined outside this section --> <base /> </inbound> <backend> <base /> </backend> <outbound> <base /> </outbound> <on-error> <base /> </on-error> </policies> This policy creates a secure gateway that validates Azure AD tokens from your local Azure CLI session, then uses APIM's managed identity to authenticate with Azure OpenAI backends, eliminating the need for API keys. It automatically load-balances requests across multiple Azure OpenAI regions using round-robin distribution for optimal performance. 4. Create Azure APIM proxy for Cline This FastAPI-based proxy forwards OpenAI-compatible API requests from Cline through APIM using Azure AD authentication via Azure CLI credentials, eliminating the need to store or manage OpenAI API keys. Prerequisites: Python 3.8 or higher Azure CLI (ensure az login has been run at least once) Ensure the user running the proxy script has appropriate Azure AD roles and permissions. This script uses Azure CLI credentials to obtain bearer tokens. Your user account must have the correct roles assigned and access to the target API audience configured in the APIM policy above. Quick setup for the proxy: Create this requirements.txt: fastapi uvicorn requests azure-identity Create this Python script for the proxy source code azure_proxy.py: import os import requests from fastapi import FastAPI, Request from fastapi.responses import StreamingResponse import uvicorn from azure.identity import AzureCliCredential # CONFIGURATION APIM_BASE_URL = <APIM BASE URL HERE> AZURE_SCOPE = <AZURE SCOPE HERE> PORT = int(os.environ.get("PORT", 8080)) app = FastAPI() credential = AzureCliCredential() # Use a single requests.Session for connection pooling from requests.adapters import HTTPAdapter session = requests.Session() session.mount("https://", HTTPAdapter(pool_connections=100, pool_maxsize=100)) import time _cached_token = None _cached_expiry = 0 def get_bearer_token(scope: str) -> str: """Get an access token using AzureCliCredential, caching until expiry is within 30 seconds.""" global _cached_token, _cached_expiry now = int(time.time()) if _cached_token and (_cached_expiry - now > 30): return _cached_token try: token_obj = credential.get_token(scope) _cached_token = token_obj.token _cached_expiry = token_obj.expires_on return _cached_token except Exception as e: raise RuntimeError(f"Could not get Azure access token: {e}") @app.api_route("/{path:path}", methods=["GET", "POST", "PUT", "PATCH", "DELETE", "OPTIONS"]) async def proxy(request: Request, path: str): # Assemble the destination URL (preserve trailing slash logic) dest_url = f"{APIM_BASE_URL.rstrip('/')}/{path}".rstrip("/") if request.url.query: dest_url += "?" + request.url.query # Get the Bearer token bearer_token = get_bearer_token(AZURE_SCOPE) # Prepare headers (copy all, overwrite Authorization) headers = dict(request.headers) headers["Authorization"] = f"Bearer {bearer_token}" headers.pop("host", None) # Read body body = await request.body() # Send the request to APIM using the pooled session resp = session.request( method=request.method, url=dest_url, headers=headers, data=body if body else None, stream=True, ) # Stream the response back to the client return StreamingResponse( resp.raw, status_code=resp.status_code, headers={k: v for k, v in resp.headers.items() if k.lower() != "transfer-encoding"}, ) if __name__ == "__main__": # Bind the app to 127.0.0.1 to avoid any Firewall updates uvicorn.run(app, host="127.0.0.1", port=PORT) Run the setup: pip install -r requirements.txt az login # Authenticate with Azure python azure_proxy.py Configure Cline to use the proxy: Using the OpenAI Compatible API Provider: Base URL: http://localhost:8080 API Key: <any random string> Model ID: <your Azure OpenAI deployment name> API Version: <your Azure OpenAI deployment version> The API key field is required by Cline but unused in our implementation - any random string works since authentication happens via Azure AD. 5. Configure Cline to listen to your MCP Server Now that we have both our MCP server running and Cline configured with secure OpenAI access, the final step is connecting them together. To enable Cline to discover and use your custom tools, navigate to your installed MCP servers on Cline, select Configure MCP Servers, and add in the configuration for your server: { "mcpServers": { "mcp-tools": { "autoApprove": [ "EchoCaps", "ReverseEcho", ], "disabled": false, "timeout": 60, "type": "sse", "url": "http://<your localhost url>/sse" } } } Now, you can use Cline's chat interface to interact with your secure MCP tools. Try asking Cline to use your custom tools - for example, "Can you echo 'Hello World' in capital letters?" and watch as it calls your local MCP server through the infrastructure you've built. Conclusion There you have it: A secure implementation of MCP that can be tailored to your specific use case. This approach gives you the power of MCP while maintaining enterprise security. You get: AI capabilities through secure Azure infrastructure. Custom tools that never leave your environment. Standard MCP interface for easy integration. Complete control over your data and tools. The key is keeping MCP servers local while routing AI requests through your secure Azure infrastructure. This way, you gain MCP's benefits without compromising security. Disclaimer While this tutorial provides a secure foundation for MCP implementation, organizations are responsible for configuring their Azure resources according to their specific security requirements and compliance standards. Ensure proper review of network rules, access policies, and authentication configurations before deploying to production environments. Resources MCP SDKs and Tools: MCP C# SDK MCP Python SDK Cline SDK Cline User Guide Azure OpenAI with APIM Azure API Management Network Security: Azure API Management - restrict caller IPs Azure API Management with an Azure virtual network Set up inbound private endpoint for Azure API Management Azure OpenAI and AI Services Network Security: Configure Virtual Networks for Azure AI services Securing Azure OpenAI inside a virtual network with private endpoints Add an Azure OpenAI network security perimeter az cognitiveservices account network-ruleIntroducing developer solutions for Microsoft Sentinel platform
Security is being reengineered for the AI era, moving beyond static, rule-bound controls and toward after-the-fact response toward platform-led, machine-speed defense. The challenge is clear: fragmented tools, sprawling signals, and legacy architectures that can’t match the velocity and scale of modern attacks. What’s needed is an AI-ready, data-first foundation - one that turns telemetry into a security graph, standardizes access for agents, and coordinates autonomous actions while keeping humans in command of strategy and high-impact investigations. Security teams already center operations on their SIEM for end-to-end visibility, and we’re advancing that foundation by evolving Microsoft Sentinel into both the SIEM and the platform for agentic defense—connecting analytics and context across ecosystems. And today, we’re introducing new platform capabilities that build on Sentinel data lake: Sentinel graph for deeper insight and context; Sentinel MCP server and tools to make data agent ready; new developer capabilities; and Security Store for effortless discovery and deployment—so protection accelerates to machine speed while analysts do their best work. Today, customers use a breadth of solutions to keep themselves secure. Each solution typically ingests, processes, and stores the security data it needs which means applications maintain identical copies of the same underlying data. This is painful for both customers and partners, who don’t want to build and maintain duplicate infrastructure and create data silos that make it difficult to counter sophisticated attacks. With today’s announcement, we’re directly addressing those challenges by giving partners the ability to create solutions that can reason over the single copy of the security data that each customer has in their Sentinel data lake instance. Partners can create AI solutions that use Sentinel and Security Copilot and distribute them in Microsoft Security Store to reach audiences, grow their revenue, and keep their customers safe. Sentinel already has a rich partner ecosystem with hundreds of SIEM solutions that include connectors, playbooks, and other content types. These new platform capabilities extend those solutions, creating opportunities for partners to address new scenarios and bring those solutions to market quickly since they don’t need to build complex data pipelines or store and process new data sets in their own infrastructure. For example, partners can use Sentinel connectors to bring their own data into the Sentinel data lake. They can create Jupyter notebook jobs in the updated Sentinel Visual Studio Code extension to analyze that data or take advantage of the new Model Context Protocol (MCP) server which makes the data understandable and accessible to AI agents in Security Copilot. With Security Copilot’s new vibe-coding capabilities, partners can create their agent in the same Sentinel Visual Studio Code extension or the environment of their choice. The solution can then be packaged and published to the new Microsoft Security Store, which gives partners an opportunity to expand their audience and grow their revenue while protecting more customers across the ecosystem. These capabilities are being embraced across our ecosystem by mature and emerging partners alike. Services partners such as Accenture and ISVs such as Zscaler and ServiceNow are already creating solutions that leverage the capabilities of the Sentinel platform. Partners have already brought several solutions to market using the integrated capabilities of the Sentinel platform: Illumio. Illumio for Microsoft Sentinel combines Illumio Insights with Microsoft Sentinel data lake and Security Copilot to revolutionize detection and response to cyber threats. It fuses data from Illumio and all the other sources feeding into Sentinel to deliver a unified view of threats, giving SOC analysts, incident responders, and threat hunters visibility and AI-driven breach containment capabilities for lateral traffic threats and attack paths across hybrid and multi-cloud environments. To learn more, visit Illumio for Microsoft Sentinel. OneTrust. OneTrust’s AI-ready governance platform enables 14,000 customers globally – including over half of the Fortune 500 – to accelerate innovation while ensuring responsible data use. Privacy and risk teams know that undiscovered personal data in their digital estate puts their business and customers at risk. OneTrust’s Privacy Risk Agent uses Security Copilot, Purview scan logs, Entra ID data, and Jupyter notebook jobs in the Sentinel data lake to automatically discover personal data, assess risk, and take mitigating actions. To learn more, visit here. Tanium. The Tanium Security Triage Agent accelerates alert triage using real-time endpoint intelligence from Tanium. Tanium intends to expand its agent to ingest contextual identity data from Microsoft Entra using Sentinel data lake. Discover how Tanium’s integrations empower IT and security teams to make faster, more informed decisions. Simbian. Simbian’s Threat Hunt Agent makes hunters more effective by automating the process of validating threat hunt hypotheses with AI. Threat hunters provide a hypothesis in natural language, and the Agent queries and analyzes the full breadth of data available in Sentinel data lake to validate the hypothesis and do deep investigation. Simbian's AI SOC Agent investigates and responds to security alerts from Sentinel, Defender, and other alert sources and also uses Sentinel data lake to enhance the depth of investigations. Learn more here. Lumen. Lumen’s Defender℠ Threat Feed for Microsoft Sentinel helps customers correlate known-bad artifacts with activity in their environment. Lumen’s Black Lotus Labs® harnesses unmatched network visibility and machine intelligence to produce high-confidence indicators that can be operationalized at scale for detection and investigation. Currently Lumen’s Defender℠ Threat Feed for Microsoft Sentinel is available as an invite only preview. To request an invite, reach out to the Lumen Defender Threat Feed Sales team. The updated Sentinel Visual Studio Code extension for Microsoft Sentinel The Sentinel Extension for Visual Studio code brings new AI and packaging capabilities on top of existing Jupyter notebook jobs to help developers efficiently create new solutions. Building with AI Impactful AI security solutions need access and understanding of relevant security data to address a scenario. The new Microsoft Sentinel Model Context Protocol (MCP) server makes data in Sentinel data lake AI-discoverable and understandable to agents so they can reason over it to generate powerful new insights. It integrates with the Sentinel VS Code extension so developers can use those tools to explore the data in the lake and have agents use those tools as they do their work. To learn more, read the Microsoft Sentinel MCP server announcement. Microsoft is also releasing MCP tools to make creating AI agents more straightforward. Developers can use Security Copilot’s MCP tools to create agents within either the Sentinel VS Code extension or the environment of their choice. They can also take advantage of the low code agent authoring experience right in the Security Copilot portal. To learn more about the Security Copilot pro code and low code agent authoring experiences visit the Security Copilot blog post on Building your own Security Copilot agents. Jupyter Notebook Jobs Jupyter notebooks jobs are an important part of the Sentinel data lake and were launched at our public preview a couple of months ago. See the documentation here for more details on Jupyter notebooks jobs and how they can be used in a solution. Note that when jobs write to the data lake, agents can use the Sentinel MCP tools to read and act on those results in the same way they’re able to read any data in the data lake. Packaging and Publishing Developers can now package solutions containing notebook jobs and Copilot agents so they can be distributed through the new Microsoft Security Store. With just a few clicks in the Sentinel VS Code extension, a developer can create a package which they can then upload to Security Store. Distribution and revenue opportunities with Security Store Sentinel platform solutions can be packaged and offered through the new Microsoft Security Store, which gives partners new ways to grow revenue and reach customers. Learn more about the ways Microsoft Security Store can help developers reach customers and grow revenue by visiting securitystore.microsoft.com. Getting started Developers can get started building powerful applications that bring together Sentinel data, Jupyter notebook jobs, and Security Copilot today: Become a partner to publish solutions to Microsoft Security Store Onboarding to Sentinel data lake Downloading the Sentinel Visual Studio Code extension Learn about Security Copilot news Learn about Microsoft Security StoreInvestigating M365 Copilot Activity with Sentinel & Defender XDR
As organizations embrace AI-powered tools like Microsoft Copilot, ChatGPT, and other generative assistants, one thing becomes immediately clear: AI is only as trustworthy as the data it can see. These systems are increasingly woven into everyday workstreams, surfacing insights, drafting content, and answering questions based on enterprise data signals. Yet behind the magic lies a new security frontier: making sure AI only accesses the right data, the right way, at the right time. That’s where Data Security Posture Management (DSPM) comes into play. Data Security Posture Management (DSPM) for AI is a Microsoft Purview capability designed to help organizations discover, secure, and apply compliance controls for AI usage across your enterprise. With personalized recommendations, one-click policies help you protect your data and comply with regulatory requirements and get ahead of questions like: Where is my sensitive data stored and who has access? Are we protecting data from potential oversharing risks Are we protecting sensitive data references in Copilot and agent responses? How to maintain compliance and governance over data accessed by AI Are we empowering users with AI safely and responsibly, backed by security? In this blog, we will explore how Microsoft Sentinel and Defender XDR can help security teams operationalize DSPM for AI. From capturing Copilot interaction telemetry to building investigations and accelerating response. To learn more about Data Security Posture Management (DSPM) for AI, please visit DSPM for AI M365 Copilot activity in the SOC Getting Started: This advanced hunting table is populated by records from Microsoft Defender for Cloud Apps. If your organization hasn’t deployed the service in Microsoft Defender XDR, queries that use the table aren’t going to work or return any results. To make sure the CloudAppEvents table is populated, make sure to enable Microsoft 365 activities. Follow this article for detailed steps. For more information about how to deploy Defender for Cloud apps in Defender XDR, refer to Deploy supported services. You can perform Advanced Hunting of Microsoft 365 Copilot data through CloudAppEvents. CloudAppEvents is a powerful table in Microsoft Defender XDRs advanced hunting schema that captures user and admin activities across Microsoft Cloud apps. To make sure the CloudAppEvents table is populated, follow the steps mentioned in the article here. The CloudAppEvents table contains enriched logs from all SaaS applications connected to Microsoft Defender for Cloud Apps refer to Apps and Services covered by CloudAppEvents DSPM for AI and CloudAppEvents Activity Explorer is the central investigative hub in Data Security Posture Management (DSPM) for AI. It surfaces granular telemetry about AI interactions, capturing prompts, responses, user identities, and sensitive information types (SITs) provided you’ve the right permissions/policies enabled. Whether the activity originates from Microsoft Copilot, third-party GenAI apps, or custom enterprise agents, Activity Explorer provides the visibility needed to assess risk and take action. Microsoft Purview’s Data Security Posture Management (DSPM) for AI provides visibility and it’s tightly integrated with Microsoft Sentinel and Defender XDR through the CloudAppEvents table. The Flow Explained Event Generation in DSPM for AI Every AI interaction, whether from Copilot, Fabric, or unmanaged apps like DeepSeek is captured in the Microsoft 365 Unified Audit Log. These logs include metadata like user identity, app name, agent name, prompt content, and sensitivity label matches Ingestion into CloudAppEvents The audit data flows into the CloudAppEventstable within Microsoft Defender XDR if you’ve enabled the app connector. Follow this article for detailed steps. This table is part of the advanced hunting schema and includes telemetry for user and object activities across Microsoft 365 and other cloud apps Availability in Microsoft Sentinel Because CloudAppEvents is also exposed in Microsoft Sentinel, customers can query AI-related activities using KQL for threat hunting, incident correlation, and compliance investigations. This enables a unified view across Sentinel and XDR without needing a separate connector What You Can Do with CloudAppEvents Advanced Hunting: Use KQL to search for AI interactions that match specific sensitivity labels, user risk scores, or app types. Incident Investigation: Correlate AI activity with alerts from Office 365. Compliance Audits: Track for activity hunting. Custom Dashboards: Visualize AI usage patterns in Power BI or Sentinel dashboards or Workbooks. Example KQL Query Best Practice: Build KQL queries that filter by Application == "Microsoft 365 Copilot" and ActionType == " Interactwithcopilot" to surface relevant events. For eg., A simple query to get started analyzing the interactions of M365 Copilot CloudAppEvents | where Application in ("Microsoft 365", "Microsoft 365 Copilot Chat") | where ActivityType == "Interactwithcopilot" Known Gaps The CloudAppEvents table, which ingests AI activity from the Microsoft 365 Unified Audit Log, is incredibly useful for activity hunting. It gives you metadata like: Timestamp User identity App and agent name Action type (e.g., AIInteraction) You won’t see the actual prompt or response from the AI interaction and you won’t get DSPM enrichment like, Sensitivity Information Types (SITs), Policy hits. These records only contain message metadata. Navigate to Purview’s DSPM for AI Activity Explorer to review the prompts and responses. While CloudAppEvents is great for identifying patterns and correlating activity across users and apps, it doesn’t give you the full picture needed for deep investigation or compliance auditing. If you need that level of detail, you’ll want to pivot into DSPM for AIs Activity Explorer, where you can inspect the full interaction including prompt, response, and policy context. Acknowledgements: Special Thanks to Martin Gagné, Principal Group Engineering Manager, for reviewing this blog and providing valuable feedback.
