Our approach
Governance-led AI adoption, in four stages.
A practical framework that works on any platform — built from years of doing this work inside regulated organisations.
Our approach
Our four principles
Governance-led, not governance-late.
Most consultancies treat governance as a phase you bolt on at the end — after the technology is in, after the use case is picked, after the rollout has started. We treat it as the foundation that everything else stands on.
It’s not a slower way to work. It’s faster. When governance is designed in from the start, AI adoption sticks. People trust it. Risk teams stop pushing back. Decisions get made faster because the trade-offs are explicit instead of hidden.
The work we do here is concrete and practical: classifying data, controlling permissions, setting retention, auditing ownership. Not policy theatre.
Proportionate to your risk and size.
There is no single right answer to “how should we govern AI?” A 500-person professional services firm shouldn’t pay for the controls of a global bank, but it shouldn’t run on the same controls as a 20-person startup either.
We start by understanding your actual risk — what data you hold, who needs it, what’s regulated, what isn’t. Then we design controls that fit. The aim is the right level of oversight, not the maximum level.
Maximum controls slow you down and rarely get used; minimum controls leave real risks exposed.
Proportionality is a judgement call, and it’s the kind of judgement that needs senior consultants to make well.
Senior people, every engagement.
Most consultancies operate a pyramid: senior people sell the work, junior people deliver it. The economics make sense for the consultancy. They don’t always make sense for you.
We don’t run a pyramid. The consultant you meet in the first call is the consultant who scopes the engagement and the consultant who delivers it. When we need more capacity or deeper expertise, we bring in named, trusted partners from our network — not whoever’s available.
This limits our scale. That’s deliberate. It also means you get senior judgement on every conversation, not just the sales pitch.
Platform-fluent, vendor-independent.
We’re a Microsoft AI Cloud Partner and most of our current work is in the Microsoft ecosystem. But we’re not paid by Microsoft. We’re paid by you.
That means we’ll tell you when not to buy something. When Copilot isn’t worth the licence cost. When a simple approach will do. When a non-Microsoft tool is a better fit. And when the right answer is to slow down and fix foundations before adding more AI on top.
Our principles apply wherever AI is being adopted — Microsoft, Google Cloud, AWS, your own platforms. The technology changes; the way to do this work properly doesn’t.
Our process
Four stages from assessment to handover
Most engagements move through four stages. Some clients work with us across all four; others bring us in for one or two.
| Number | Stage | What happens |
|---|---|---|
| 01 | Assess | We map your current data, governance, and platform position. We identify where AI is safe to deploy today and where the risks are. |
| 02 | Align | We agree the right level of control for your risk and maturity. Proportionate — not maximum, not minimum. We document the trade-offs explicitly. |
| 03 | Build | We implement the foundations — data governance, classification, controls — and pilot AI where it’s ready. Senior delivery throughout. |
| 04 | Embed | We support adoption with change, training, and ongoing oversight. We hand over capability that your team can run independently. |
What we don’t do
What we don’t do
Being clear about what we don’t do is part of how we work. It keeps engagements focused and outcomes realistic.
- We don’t do AI strategy theatre. If a workshop won’t change a decision, we won’t sell you one.
- We don’t resell software. Our recommendations aren’t shaped by vendor margins or co-sell incentives.
- We don’t put junior consultants on engagements. Every project is led by a senior practitioner.
- We don’t do open-ended retainers without scope. Every engagement has a clear end state and handover plan.
- We don’t pretend to be expert in everything. Where we’re not the right partner, we’ll say so.
Ready to talk
An initial conversation is the quickest way to see whether our approach fits what you’re trying to do.