Microsoft Teams to Add AI Phishing Detection for External Chats, Targeting November Release
A Microsoft 365 roadmap entry says the feature will analyze conversation context to flag phishing and social engineering.
Microsoft is adding AI-powered phishing detection to Teams messaging, according to a Microsoft 365 roadmap entry published October 14 that targets general availability in November 2026.
What the feature does
The roadmap describes a capability that uses conversation context to identify phishing and social engineering attempts, focusing on higher-risk situations, primarily first-contact chats from people outside a user's organization. When suspicious activity is detected, users may see a warning. Microsoft says the capability is used only for security purposes, and that protection may vary by region or scenario.
Rollout details
- Roadmap ID: RM575288
- Status: in development
- Target release: general availability in November 2026
- Platforms: Android, desktop, iOS, Mac and web
- Cloud: worldwide standard multi-tenant
- Admin action: the entry lists none
Because the item is still marked in development, the timing could change.
Context
Teams already warns users about potential spam or phishing from external contacts requesting to chat, and Microsoft has a separate brand impersonation protection for organizations that allow external access. The new feature adds analysis of conversation content rather than relying only on sender signals. Chat platforms have become a common route for social-engineering attacks because messages from outside a company can appear alongside internal conversations.
Microsoft has not published detection accuracy figures or details on how the model handles false positives.
