AI is probably already in your business.
Whatever your policy says, people on your team are likely using AI to get through their work. Maybe with ChatGPT, maybe Gemini, maybe Claude. It's possible your people are pasting client emails into these tools for a quick redraft. Summarising long contracts. Cleaning up customer notes. None of this is reckless. Just useful.
The productivity gains are obvious. But there's a downside.
If a regulator or a client asked what data went where, your organisation might not have a clean answer.
Why that's a problem
When a staff member pastes a client document into ChatGPT, that document goes to a US company. But under UK law, the responsibility for what happened with that data sits with your organisation, not the staff member who pasted it.
Your data (and maybe that of your clients and customers) is somewhere you can't see, in a system you don't control, sitting in logs you can't audit.
This has a name. When people use AI tools your organisation never approved, that's Shadow AI.
"Working with Peter was an absolute pleasure, thanks to his flexibility, excellent communication skills and honesty."
Luca Russo, Web & Social Collaboration Functional Lead, Givaudan SA
The Sovereign AI approach
The good news: none of this means your team has to stop using AI. The risk comes from where the data goes, not from the AI itself.
Sovereign AI is the way around that.
It isn't a product you can buy. It's a property of how an AI system is put together, and it has three parts.
- Data sovereignty. Where your prompts and answers live, and who can touch them.
- Model sovereignty. What the AI was trained on, who owns it, and where it runs.
- Operational sovereignty. Who can change the system, audit it, and switch it off.
Often, a lot of what's sold under the name Sovereign AI only covers the first one.
Here's a full walkthrough that covers all three, and what falls over when one is missing.
What a Sovereign AI build can do
That's the idea. Here's what it looks like once it's built.
The technical part lives in your cloud. What your team sees is closer to the chat tools they're already used to.
Six capabilities cover the shape of what's possible: where your data ends up, whose rules the assistant follows, how it stays grounded in your own knowledge, how it reasons across many of your documents at once, how it connects to the systems you already use, and how it stays portable when models change.
Each capability is illustrated with a short clip from a real build.
How we'd work together
The Discovery. Most engagements start here: a fixed price two week scope of what your team is already doing with AI, what they'd actually want to do with it, what state your data is in, and what's feasible to build first.
You get a written report covering the use cases worth pursuing, the data work involved, and an architecture sketch for a first build. No commitment to go further.
What happens after the Discovery depends on what it finds. Usually it's one or both of two phases.
The data work. More often than not, the data isn't ready. It's spread across systems, stored in formats that are awkward to use, or nobody's quite sure which of it you actually need. Getting it into shape is usually the larger piece of the job, and it's the part that decides whether everything built on top of it holds up.
The Sovereign AI build. Then the tools themselves: the private AI assistants your staff use day to day, built and deployed inside your own infrastructure. This is the part the use cases above show.
"Pete has been amazing to work with. He explains things to us in a way that makes sense in English, not developer-speak. We feel like Pete is an extended part of our team."
Katie Angotti, Programme Lead, Danone
About me
I've spent over 25 years building the data architecture under big software for UK organisations.
The platforms have changed (Perl and Python before the web was the default, Drupal and Laravel through the 2010s, Sovereign AI today) but the work hasn't. It's still about getting messy data out of legacy systems and into shape so something useful can sit on top.
Since 2009 I've done that for over 45 UK organisations. Sovereign AI is just the new layer on top. The data plumbing underneath is still the hard part.
A few you may recognise:
"We have worked with many developers and can confidently say that Pete is by far the best, an exceptional coder and consultant with an impressive skill set."
Kathryn Maxwell, IT Project Manager, Royal Meteorological Society
