01
AI tools and providers
Recognizes calls to dozens of AI providers and tools, including ChatGPT, Claude, Gemini, Microsoft Copilot, Perplexity, Grok, DeepSeek, Mistral, and AI app builders like Lovable, Bolt, v0, and Replit.
Product · Shadow AI Browser Extension
See the AI your employees are already using.
Most AI use at work happens in the browser, on tools nobody approved. The Shadow AI Browser Extension gives you real-time visibility into browser-based AI across your managed devices: which AI tools are being used, which apps offer AI, which AI extensions are installed, and when files are uploaded to AI. It records that AI was used, never what was typed.
What it detects
Shadow AI comes in through more doors than you think.
Blocking ChatGPT does not solve shadow AI. AI is built into the SaaS apps your teams use every day, installed as browser extensions, and reachable through dozens of providers. The extension watches all of those doors.
02
AI built into everyday apps
Maps the AI features embedded in the apps your teams already use, such as Microsoft 365 Copilot, Notion AI, Slack AI, and AI support agents from Intercom, Zendesk, Drift, and others.
03
AI browser extensions
Finds AI extensions installed in the browser, such as AI writing assistants and ChatGPT sidebars. These inject AI into every page and can send page content to their own servers.
04
File uploads to AI
Flags when a file is uploaded to an AI service, with a rough size. It records that data left the building, never the file itself.
05
Explainable risk scoring
Every AI tool gets a transparent 0 to 100 risk score based on account type, whether the provider may train on your data, your policy decision, and whether uploads were seen. Your CISO can read exactly why a tool scored the way it did.
06
Your policy, applied centrally
Classify each AI tool as sanctioned, unsanctioned, or blocked, and push that policy to every device. Add your own internal AI tools so they are recognized too.
Evidence you can defend
It tells you how sure it is.
Not every signal means the same thing. The extension separates "AI was used" from "this page offers AI," so you act on evidence, not guesses.
Near-certain
AI was accessed
A request went to a known AI provider. This is the strongest signal, down to the moment it happened.
Surface only
The page offers AI
A known AI widget or feature is present on the page. Useful for mapping where AI is available, but not proof that anyone used it.
Weak, supplementary
Possible AI response
Text arrived in the word-by-word pattern AI responses tend to have. Supporting context only, never presented as proof.
Privacy by design
Visibility without surveillance.
The extension is built so it cannot read what employees type into AI tools or what those tools send back. That is not a setting that can be quietly switched on. It is how the extension is built.
That protects your employees, and it protects you from the legal exposure that comes with capturing conversation content.
- ✓Records: that AI was accessed, which provider, and the site, with query strings stripped at the point of capture.
- ✓Records: that a file was uploaded to AI, and its approximate size.
- ✓Anonymous by default: reporting shows organization-level trends without identifying individual employees. Your organization can choose to report by user instead.
- ×Never records: prompts, responses, form fields, or the contents of uploaded files.
Company-wide deployment
Rolled out the way IT already manages browsers.
The extension runs in Google Chrome and Microsoft Edge. IT pushes it, along with your organization's AI policy, to managed devices using the tools you already have. Reporting is anonymous by default, or can be set to show activity by user, and can stay on each device or feed your own collector or SIEM.
Google Chrome
Microsoft Edge
Group Policy
Microsoft Intune
Jamf
Google Workspace
Bring it to your organization
Let's talk about installing it company wide.
Tell us a little about your organization and we will reach out to walk through how the extension works, how it is deployed to managed devices, and how it fits with your AI governance program.
Shadow AI visibility pairs naturally with Governance 1st, where your AI policies, risk register, and training live.