Microsoft expands native AI and data controls
Microsoft released Copilot Studio updates in October 2026 that allow makers to discover and use shared plugins from a central registry. This integration provides a consistent way to extend agents across the Microsoft AI ecosystem. Microsoft also updated Copilot Notebooks with new audio overview formats like Critique and Debate. These updates expand audio overview context to more content types. Microsoft Copilot in OneNote can now create, edit, and refine content across various workflows. In November 2026, Microsoft Edge will start rolling out support for viewing sensitivity labels on Microsoft Information Protection protected PDFs. This feature helps users understand data handling requirements without leaving the browser. This update closes a visibility gap that previously left users unaware of a document’s protection status at the point of use. Microsoft Purview also improved permission checks for users accessing the portal and added Graph API support for archive mailboxes. These changes ensure uninterrupted retention and allow for scalable archive growth beyond the 1.5 TB limit. The Purview permission audit log improvements help administrators troubleshoot access issues and monitor authorization outcomes.
Third-party security responds to AI complexity
Proofpoint is releasing new M365 security capabilities in Q4 2026 to help organizations manage AI risks. The 365 Permission Manager makes access and sharing structures across SharePoint, OneDrive, and Microsoft Teams visible. This tool helps organizations manage the complex permissions that arise with every new team, project, or external partner. The new Tenant Risk Exposure feature provides administrators with a visual overview of permission and sharing risks across M365 tenants, including anonymous access, external users, organization-wide permissions, public groups, inactive guests, and orphaned accounts. Security teams use Explainability in Email Live Tracking to understand why a system identifies an email as suspicious. This feature allows them to investigate rare false positives or potentially incorrect decisions more efficiently. The AI-powered Security Awareness Service uses AI.MY to act as an Awareness Coach. This service contacts employees after e-learning sessions and asks three questions tailored to their awareness level to provide immediate feedback. These results update the user’s knowledge profile and enable organizations to measure learning progress. The AI Cyber AssistantAgent in the Control Panel provides conversational product assistance for partners and administrators during onboarding and configuration. This agent also helps partners identify the solutions that best match individual customer needs. Abnormal Security uses behavioral AI to detect business email compromise and account takeovers. This platform integrates with Microsoft 365 via API to analyze communication patterns and user behavior. Abnormal Security focuses on real-time phishing attack prevention and catches socially engineered attacks that bypass conventional tools. You should pick Abnormal Security if your organization needs AI-driven protection against advanced email threats.
Pricing and platform comparison
| Feature | Abnormal Security | Proofpoint |
|---|---|---|
| Starting Price | $150/mo | $200/mo |
| User Rating | 3.9 | 3.9 |
| Primary AI Focus | Behavioral AI | Human-centric security |
| Deployment | API-based | Integrated platform |
| Target Audience | Mid-size to large enterprise | Mid-size to enterprise |
Abnormal Security starts at $150 per month, while Proofpoint starts at $200 per month. Abnormal Security targets mid-size to large enterprises through API-based behavioral AI. Proofpoint targets mid-size to enterprise organizations with an integrated platform for threat detection and data loss prevention. Both companies hold a 3.9 star user rating. Proofpoint serves over 10,000 large enterprises and 80 of the Fortune 100. Proofpoint is the better choice for large organizations that require integrated data loss prevention and human centric security. Does the expansion of Microsoft’s native AI capabilities reduce the need for third-party behavioral tools?
