April 11, 2025: WitnessAI Release v2.0

Network device integrations have been updated in WitnessAI v2.0. These include Netskope SWG, Palo Alto Networks NGFW, and Zscaler Internet Access.

Devices configured for WitnessAI v1.5 will need a minor update to support v2.0. Some new features will not activate until your devices are updated.

Existing v1.5 network device integrations will continue to work without interruption in v2.0.

See the network integration guides in the Integrations menu for details.

Have questions? Our support team will be happy to assist.

Breaking Changes

No breaking changes.

New Features

Expanded Model Support

Added over 500 new discoverable models, now totaling over 3200.

Over 200 total API-Integrated Models.

See the full list in Appendix: Model Support

Role-Based Access Control (RBAC)

Role-Based Access Control with VIP Tagging adds seven new Roles, providing granular access for both features and and permissions for the console. Enhanced data privacy and security is provided, including special handling for designated VIPs.

Critical for larger orgs with differentiated access needs. Managed under Settings > Access Control.

Witness Anywhere

“Starbucks” Use Case and Endpoint Visibility – Enables visibility into unmanaged device access via public networks (e.g., coffee shops), helping organizations monitor and manage shadow AI activity.

Witness Anywhere is now Generally Available for Windows, and in Beta for MacOS.

Applications supported on Windows:

Office365 Word, Excel, & PowerPoint apps, both native and browser

Vscode

Google gemini (gemini.google.com/app)

MS Copilot (browser based)

ChatGPT

Applications supported on MacOS (Beta):

Office365 Word, Excel, & PowerPoint apps, browser only

Vscode

Google gemini (gemini.google.com/app)

MS Copilot (browser based)

ChatGPT

Safari not currently supported

Valuable for security and compliance teams. Available automatically for customers with Witness Anywhere enabled.

Organizational Behavior Guardrail (Beta)

Organizational Behavior Guardrail (Beta) is now available in product, for modeling and detecting usage trends and anomalies in AI traffic. Six additional Alerts have been added to the original use case.

The full list is:

Burnout Risk

Potentially Departing Employee

Workplace Conflict

Freelance / Side Work

Malicious Intent

Knowledge Hoarding

Violation of Company Policies

Essential for security and compliance teams.

Harmful Response Prevention (Beta)

Harmful Response Prevention (Beta) is WitnessAI’s response analysis and control Guardrail. The purpose of this Guardrail is to analyze the responses back from AI models to user prompts, and then detect, and optionally prevent these responses from being sent to the user.

These harmful responses are evaluated in three broad categories:

Harm to Self

Harm to Others

Illegal activity

When this Guardrail detects these activities, it provides the option to Allow, Warn, or Block the response from the AI model, and optionally include a customizable message to the User.

Navigation

New icons throughout the app, and renaming of some major sections:

Dashboard is now called Home. The menu icon and label have been changed to match.

New Analytics, Lists, and Docs sections have been added

User Menu now includes a Theme toggle (Light/Dark), User Role, and WitnessAI version number. The “Chat” label has been changed to “Secure AI Portal”

Settings Redesign – Formerly Preferences, the new Settings page is full-screen with left-hand secondary navigation.

Policy Layout Refresh – The Policies screen now includes a new Global Block Policy locked at the top, and a Global AI Policy locked at the bottom.

Catalog UI Enhancements – Placeholder icons are now shown for apps missing logos.

New Proxy Block Screen – A new block notification page is rolling out for users blocked by proxy.

Home

Home is the new title for the former Dashboard. The menu icon and label have been updated to match. You can now one-click directly from the alerts timeline to the Conversations page, filtered to the time range you clicked on.

Enables one-click navigation from Home to a filtered view of Conversations.

Analytics

The new Analytics section provides detailed reporting and trend analysis on all AI usage across your org:

Behavioral Activity

Data Protection

Organizational Behavior

Risk Analysis

Model Protection

Harmful Response

Witness Anywhere

SOC Dashboard

Benefits both analysts and decision-makers. Provides insights into usage, trends, and policy impact across the org.

Advanced Policy

Still referred to as a Policy in the Console, our new Advanced Policy framework Introduces flexible and granular policy definition capabilities, and the ability to consolidate multiple previous policies into a single related policy. Combined with a new user interface, the Advanced Policy framework enables more precise control over AI usage and streamlines policy management.

Ideal for administrators managing complex environments. Available in the Policy section.

Lists Framework

Lists enable easier administration of Providers, Apps, and Models. Adds support for dynamic allowlists and blocklists to streamline policy management across organizations.

Sanctioned Providers

Sanctioned Apps

Blocked Apps

Private Models

Public Models

Helps security teams centralize control. Managed within the Policies tab. Blocked Apps are enforced at the proxy layer (via Global Block Policy). Other lists can be used as destinations in policies.

Documentation Integration

A new Docs icon is now available in the main navigation menu, and users can now search documentation directly from the magnifying glass at the top right of the Console.

Improves user guidance without leaving the product.

Google Gemini

WitnessAI now detects and reports usage of Google’s Gemini AI applications, including browser-based and native application (thick client) integrations.

Helps organizations monitor AI engagement across Google’s ecosystem. Enabled automatically as part of core detection coverage.

Windows 11 Copilot

Adds visibility into usage of Microsoft’s native Windows 11 Copilot features, including in-app suggestions and OS-level AI tasks.

Ideal for IT and security teams managing Windows environments. No configuration required.

Office365 Copilot

Monitor and surface usage of Microsoft 365 Copilot across Word and Excel.

Supports compliance and data governance in productivity tools. Detection is available out of the box.

Netskope SWG Integration

Connects WitnessAI with Netskope’s secure web gateway to monitor and enforce AI policies at the proxy level.

Ideal for organizations already using Netskope.

Known Issues

This section will be updated as issues are detected.

Supported Applications & LLMs

The list of supported Applications & LLMs has been moved to https://docs.witness.ai/v2-0/applications

Contact your account team if you have questions or need support with this release.