For partners, directors, practice managers and DPOs
At the moment, each employee makes that call alone.
In a solicitors' practice, an accountancy firm, a GP surgery or a council office, someone has probably pasted a letter into ChatGPT this morning. Nonimo keeps the habit and takes out the risk: names, National Insurance numbers and other client details become labels on the computer itself, before the text leaves. Begin with one team and widen it when you are ready.
The trouble with a ban
Microsoft commissioned Censuswide to poll 2,003 UK employees in October 2025: 71% had turned to consumer AI tools their employer had not signed off, and 51% were still doing so every week (Microsoft's UK newsroom).
Say no, and the deadlines are still there. The meeting summary gets written on somebody's phone, through a personal login, with no policy behind it and nothing logged about what went in. It makes more sense to allow AI and put the safeguard in front of the send button.
How to begin
Whoever runs your IT, in-house or outsourced, sets a managed policy that installs the Nonimo extension in those colleagues' Chrome and Edge browsers. There is no driver, proxy or root certificate involved.
Colleagues who like a keyboard shortcut can add the desktop app for Windows or Mac as well.
After that it simply works: each message bound for ChatGPT or a similar assistant has its client details labelled first. Nobody has a button to remember or a new routine to pick up.
Any computer that has the desktop app can export a signed usage report for a date range you set: a tally of redactions per category. It contains none of the text and does not say which person typed what. Where it came from can be confirmed by anyone at nonimo.ai/verify.
Something concrete for auditors, the DPO or the next board pack.
Once the pilot team has settled in, apply the same policy to every other department.
In practice
Very little, from where they sit. The usual assistants open just as they did yesterday. Client details in a message are covered automatically before it goes; anything swapped is visible, and reversing it takes one click. If your developers have built something of their own on the OpenAI API, Nonimo Gateway applies the same protection on a server you control.
Pick a profession for its own page: the reference numbers on its paperwork, what Nonimo masks before any of it gets to AI, and why.
Before you sign off
There is. Wherever the Nonimo desktop app is installed, it generates a signed usage report over whatever dates you pick, tallying redactions by category. The report holds no wording at all and gives no clue as to who typed what. Whether it is genuine and untouched is something any third party can confirm at nonimo.ai/verify.
The browser does. A forced install pushed by IT as a Chrome or Edge policy sits outside the user’s control: they can neither uninstall it nor pause it, as with any managed extension. The desktop app is optional, for individuals who find it handy; department-wide cover is the extension’s job.
Yes. Your in-house team or your managed service provider deploys it the same way as any other managed browser extension. If nobody looks after your IT, ask and we will recommend someone.
Not really. Their assistant of choice, whether ChatGPT, Claude, Copilot or Gemini, works exactly as now. Swapped-out details are highlighted where they sit, and one click restores the original.
No. The AI tool only ever sees the redacted version, client details stay on the machine where they were typed, and nothing your staff write is sent to Nonimo.