Zoom unveils AI governance tools for businesses users

Zoom has introduced a set of AI security and governance tools for organisations, aimed at giving administrators more oversight of AI activity across the platform.

The new features include centralised AI management, custom AI guardrails, sensitivity label integration with Microsoft Purview, compliance archiving for AI-generated content, and deepfake risk detection for video and audio.

The move comes as companies face growing questions about what AI systems can access, what they can generate, and how sensitive information is handled when AI agents interact with meeting content, customer data, and internal systems.

Among the planned additions, Centralised AI Management is intended to give administrators visibility into AI agents operating within an organisation, along with account-level policies, audit controls, and management tools for user-created agents. The feature is expected in December 2026.

Custom AI guardrails are designed to control what information appears in AI-generated outputs. Organisations will be able to set rules based on internal policy, including requirements for human review before certain client-facing content is shared and the exclusion of some channels from AI processing. That feature is expected in January 2027.

Data controls

Zoom has also added sensitivity label integration with Microsoft Purview, which is available now. The integration is designed to carry data classifications from Microsoft systems into Zoom Meetings and Whiteboard, allowing organisations to apply more consistent data-handling policies across documents and collaboration tools.

This addresses a common problem for businesses using multiple software systems, where classification labels and access controls do not always transfer with content as it moves between platforms. In regulated sectors, such gaps can create compliance and record-keeping challenges.

Another planned feature, Compliance Manager for My Notes and Zoom Canvas, is intended to archive AI-generated summaries, notes, and documents. It is designed to help companies that need to retain and supervise business communications and is expected to be available in January 2027.

Threat detection

Zoom is also moving into synthetic media detection with a deepfake risk detection tool. The feature will analyse video and audio signals in real time to identify content that may have been altered outside Zoom and flag possible deepfakes with a confidence indicator.

The tool is being positioned for use in recruitment, customer interactions, and meetings, where false identity or impersonation could present operational and security risks. It is expected in December 2026.

Businesses have been under increasing pressure to put controls around AI use as generative systems become embedded in communications software, workflow tools, and customer service operations. That has made governance, auditability, and content controls more important for information security and compliance teams.

For software providers, the challenge is no longer limited to adding AI features to products. They are also being asked to show how those tools are governed, how outputs are supervised, and how data protections are maintained when AI is used across internal and external communications.

Russell Dicker, Chief Product Officer at Zoom, described the update as part of that broader shift in customer requirements.

"AI agents are becoming essential collaborators in how work gets done, but organisations need the same level of control and visibility over AI that they have over any other business-critical system," said Russell Dicker, Chief Product Officer at Zoom.

"New AI governance capabilities help teams more confidently use AI to accelerate work, while enabling management of what AI can access, what it can generate, and how their data is protected," Dicker said.