Product
Loose Lips Sink Ships
Enterprise PII redaction, in the browser, before anything leaves.
A Chrome extension that stops personal information reaching AI tools and other web services. Detection runs entirely in the browser, so zero PII leaves the machine, with 17 Australian PII patterns and a classification gate for PROTECTED content.
Capabilities
Here is what Loose Lips Sink Ships does.
No personal information leaves the browser
Detection and redaction run locally. Nothing is sent to a server for analysis, so the tool itself never becomes the leak it exists to prevent.
We detect 17 Australian data patterns
Built for the Australian context, covering TFNs, Medicare numbers, ABNs, passport and licence formats and more, which a US-centric pattern set will miss.
It blocks PROTECTED content outright
Recognises and blocks security-classified content before it can be pasted into a tool that is not accredited to receive it.
Policy follows identity through Azure AD
Ties policy to identity, so different groups can hold different rules without maintaining a separate user directory.
You deploy it the way you deploy everything else
Deployable via Group Policy, Microsoft Intune or Google Workspace, the channels your SOE already uses, with no per-machine handling.
It applies to every site your staff visit
Applies to every browser destination, which is what makes it a control over your data instead of a feature inside one vendor's product.
The problem it solves
The moment staff have access to generative AI, the risk stops being theoretical. Someone will paste a case file into a chat window to get a summary, with entirely good intentions and no sense that they have just moved personal information across a boundary.
You can respond to this with policy alone, and many organisations do. Policy is necessary and it is not sufficient. It depends on every person recognising, every time, that the text on their clipboard contains something it should not.
Loose Lips Sink Ships puts the control at the point where the mistake would otherwise happen.
How it works
The extension inspects content as it is about to leave the browser and identifies personal information against pattern sets built for Australian data. Where it finds something, it redacts or blocks according to the policy applied to that user.
The critical design decision is that all of this happens locally. A redaction tool that ships your text to a cloud service for analysis has not removed the exposure. It has relocated it, and added a vendor to your data-flow diagram. Nothing leaves the machine.
The classification gate handles the higher-stakes case: content marked PROTECTED is stopped outright, because partially redacted classified material is still classified material.
Deployment
Built for how agencies actually roll software out. Push it through Group Policy, Intune or Google Workspace, tie policy to Azure AD groups, and manage it centrally. There is no per-user setup step, because a control that depends on individual installation is a control that covers the people who were already careful.
Where it sits in a governance model
This is a technical control that makes a governance position enforceable. An acceptable-use policy that says "do not put personal information into generative AI tools" is far more defensible when it is paired with something that prevents it. It also lets you say yes to AI tooling you would otherwise have to refuse.
Next step
We can show you Loose Lips Sink Ships running against your own material.
A demonstration using your documents and your terminology tells you more in an hour than a scripted walkthrough does.