AI works when your data does.
Most AI projects stall on the data, not the model. Wrong grain, no agreed definitions, nothing to measure against.
We build the data systems underneath, the applications and models on top, and we get your people using them.
Email us What we do
What we do
Seven areas. Most engagements start in the first two.
Who this is for
The question is the same at both ends of the market: what's in your data, and what could it be doing?
Business schools and universities overseas, where the data sits in six systems bought a decade apart and no two agree on a number.
Enough data to be worth modelling, not enough team to build the platform and then run it.
Ten people who want AI doing real work this quarter, and someone honest about the parts it can't do yet.
You get the people who write the code
No account manager between you and the engineers. Whoever runs your scoping call is on the build.
-
We read your data first. It tells us more in an afternoon than a requirements document does in a month.
-
First engagement is two to four weeks, fixed price, ending in a written recommendation you keep either way.
-
Everything we build lands in your accounts, documented, so your team can run it without us.
What we say no to
Being clear about the edges saves you a procurement cycle.
-
We don't sell a platform. No licence, no seat count, nothing to migrate onto.
-
We don't start an AI build without a way to measure the output. If you haven't got one, that's the first piece of work.
-
We don't take on more than we can staff. Sometimes the answer is next quarter.
10+
Years of delivered projects
12
Engineers, product and business
Certified, and sceptical about it
We hold certifications in AI product management, model development and training. Useful for procurement, less useful for deciding whether your idea will work.
That decision comes from your data. Sometimes the honest answer is a report rather than a model, and we'll give you that in writing.
Before the first call
If yours isn't here, email it.
-
How does an engagement start?
With a short piece of paid work. Two to four weeks, fixed scope, ending in a written recommendation you own whether you continue or not.
It costs less than a long discovery phase and tells both of us whether we're the right team.
-
We're small. Are we too small?
No. A ten person company often gets more out of this than a large one, because the work sits closer to the person doing it and a change lands the same week.
What matters is whether there's a real task worth changing, not headcount.
-
Do you work with our existing tools?
Yes, and most of our work assumes you're keeping them. Rip and replace is expensive, slow, and rarely the thing that was broken.
-
Why are there no names on this site?
The person you'd want depends on what you're asking, and a photo grid doesn't tell you that. Email us and you'll get the engineer or product manager who'd do the work.
-
What happens to our data while you're working?
Wherever possible it stays in your environment. We work inside your cloud accounts, under access your team grants and can revoke.
Where that isn't possible for a piece of work, we say so before it starts.