Microsoft continues to deliver technology to aid IT and security pros in governing their expanding AI software estates. In an online event this week, the company outlined four pending enterprise AI features, led by data protection technology in Purview for Copilot users and a dashboard for monitoring AI agent usage and performance.
The latest developments, all of which are slated for delivery by the end of March, follow other recent developments geared to enterprise-scale manageability:
- a new Security Dashboard for AI
- a Model Context Protocol server that simplifies access to internal and external security data
- ramped-up threat detection for the Agent 365 Control Plane.
Collectively, this round of new and enhanced products shows Microsoft, like core competitors, responding to corporate AI customers that need robust controls, governance, and security protections as they strive to manage both corporate-sanctioned AI and shadow AI that poses ongoing challenges.
Enterprise-Grade Controls
One of the major enhancements Microsoft is delivering next month is Purview Data Loss Prevention (DLP) for Copilot, which prevents Copilot from responding to prompts, connecting to internal data sources, or performing web searches if a user’s prompt contains sensitive data.
The functionality applies to Copilot Chat, Microsoft 365 Copilot, and agents built in Copilot Studio. It’s available to all users of 365 Copilot and Copilot chat. “It’s basically ensuring that the security features that your organization cares about…apply to the Copilot and the user that is interacting with it. And if something is flagged as sensitive data, then it’s shut down,” said Cam Smith, director of product marketing, Copilot at Microsoft, during this week’s event.
Purview DLP for Copilot will be delivered on the heels of Purview in the Microsoft 365 Admin Center, which became available in January. Purview in the Admin Center ensures secure Copilot usage by providing visibility into oversharing risks – and driving remediations as those risks arise.
AI Agent & Copilot Summit is an AI-first event to define opportunities, impact, and outcomes with Microsoft Copilot and agents. Building on its 2025 success, the 2026 event takes place March 17-19 in San Diego. Get more details.
Purview in the Microsoft 365 Admin Center will help security and IT leaders understand sensitive, protected Copilot conversations and enable them to turn on Purview DLP for Copilot from within the admin center. It will give tech leaders the tools and visibility to complete suggested actions to ensure they’re aligned with AI compliance standards.
A new Agent Dashboard is designed to measure adoption and impact of agents across an organization. It will provide views into performance metrics for all agents including internally built agents, Microsoft-developed agents, and third-party agents that are in use. It will identify the most widely used agents and dive deeper into how they’re used, while helping leaders to understand agent adoption trends across teams and functions. The Agent Dashboard is available worldwide this month.
“One of the things that we want to make sure we do a good job of is extending the ways in which you can see those [Copilot] analytics to the agents ecosystem that your organization is probably exploring today,” Microsoft’s Smith said.
Finally, the company said it is onboarding two “pay as you go” services for Copilot: specifically, the services are Copilot Tuning and the Microsoft 365 Copilot Retrieval API.
Copilot Tuning fine-tunes AI models on tenant-specific data to improve accuracy and domain relevance for tasks such as summarization and document generation.
The retrieval API enables developers to securely retrieve relevant text snippets from Microsoft 365 content stored in platforms including SharePoint and OneDrive in order to ground GenAI responses.
Ask Cloud Wars AI Agent about this analysis
