Send people from shadow AI to the AI you approve
Instead of a dead-end block page, show people why a tool is not approved and give them the approved one in a single click. Every choice is recorded.
A page that changes behaviour
People open an AI tool because they have work to do. The page respects that: it explains, offers the approved tool, and gets out of the way.
Show the risk
The service’s risk and whether it trains on what you send it, in plain words.
Offer the approved tool
One click into Secure AI Chat or the AI tool you approved instead.
Ask for access
People can request the tool; an admin decides in one place.
Continue, when you allow it
Per service you decide whether people may go on, and each choice is recorded.
Your choice, per AI service
Set one action per risk level, then override it for the services that need it.
| Action | Typical use |
|---|---|
| Allow | Approved tools, and low-risk services you are happy with |
| Warn | The page shows the risk and offers the approved tool; people may continue, and that is recorded |
| Block | The page offers the approved tool, with no way through to the service |
Why blocking alone fails
Block an AI tool on the company network and the need does not go away. People use their phones, personal laptops or home connections, and the data goes with them.
A redirect keeps the person on a path you can see, and gives them something better than a dead end.
In your branding and their language
The page carries your logo and speaks the person's language, so it reads as company guidance rather than an error.
See how many people were warned and how many chose not to continue in Shadow AI Control.
Questions
Why redirect instead of blocking AI tools?
A block page ends the conversation, so people move to their phones or personal laptops, where you see nothing. A redirect explains the risk and offers the approved tool right away, so most people take it, and the use you could not see moves somewhere you can.
What does the person see?
A page in your branding and their language that names the service, its risk and whether it trains on what you send it, with a reminder about what not to paste and a one-click way into the approved AI tool.
Can people still use the original tool?
Only if you allow it. You decide per service whether people may continue, and every choice to continue is recorded. For the services you do not allow at all, the page blocks instead.
Can people ask for a tool to be approved?
Yes. They can sign in and ask for access from the page, and an admin decides in one place.