For principals, partners, practice managers and privacy officers
At the moment, each staff member decides for themselves.
Blocking ChatGPT hasn’t kept it out of Australian workplaces. With Nonimo, your team keeps working in Copilot, ChatGPT, Gemini or Claude, and names, TFNs and Medicare numbers become placeholders right there on the PC before a message goes anywhere. Begin with a single team and grow from there.
The trouble with a ban
In the Australian snapshot of the University of Melbourne and KPMG’s 2025 global study on trust in AI, 44% of employees said they had used AI in ways that contravene policies and guidelines (KPMG and the University of Melbourne, 2025).
The deadline doesn’t go away because the answer was no. The client email gets drafted on someone’s phone instead, in a free account the practice has never seen and can’t check. The better trade is a yes, with the guardrails already sitting on each computer before any client’s details leave it.
A small first step
Whoever looks after your IT, in-house or an MSP, pushes the extension out by policy to Chrome and Edge for that team. No network appliance, driver, proxy or root certificate is involved.
Staff who like a keyboard shortcut can add the desktop app for Mac or Windows as well.
After that the extension does its job unprompted, replacing client details with labels in ChatGPT and other chatbots as each message heads out. Nobody needs a new shortcut or a new routine.
On any computer running the desktop app, you can pull a signed usage report covering whichever dates you need: a count of redacted items per category. It carries no text, and nothing that links a redaction to a person. nonimo.ai/verify lets anyone check it’s genuine and unchanged.
Handy when an auditor, the board or your privacy officer comes asking.
Once the pilot team is running smoothly, apply that same policy to every other team you have.
On the ground
Not much at all. They open Copilot, ChatGPT, Gemini or Claude just as they did last week. By the time a message is sent, the client details in it are already covered by the extension, with no prompting. Every redaction is there on screen, and any of them can be reversed in a click. If your developers have built tools of their own on the OpenAI API, the same job is done inside your own servers by Nonimo Gateway.
Each card opens a guide to one line of work: the numbers its paperwork carries, what Nonimo covers up before any AI tool sees it, and why.
Worth a read
How ChatGPT, Claude, Gemini and Copilot sit under the APPs, including APP 8 on sending personal information overseas.
An AI clause for the engagement letters of Australian law and accounting firms, drafted from the regulators’ own documents.
No. Once IT installs it by policy, the person using the browser can’t remove or disable it: that is simply how forced installs behave in Chrome and Edge. The Mac and Windows app is optional, there for anyone who finds it useful. Team-wide protection comes from the extension and its policy.
Yes. It’s a job for your internal IT team or your MSP: the extension goes out through policy, like every other extension they already manage. If nobody looks after your IT, let us know and we’ll put you in touch with someone who can.
Yes. Wherever the Nonimo app is installed, it can generate a signed usage report counting the items redacted per category between two dates you pick. It contains no text and doesn’t say who typed anything. At nonimo.ai/verify, anyone can confirm that Nonimo issued the report and that nobody has tampered with it.
No. People keep opening Copilot, ChatGPT, Gemini or Claude exactly as before. They see every swap highlighted, and one click puts the original back.
No. What reaches the chatbot is the redacted text and nothing else. Client details stay on the machine they were typed on, and none of your staff’s text comes to us.