About
Grapine is an independent AI studio. We run a small number of consulting engagements for businesses implementing AI, and we build focused AI products alongside that work.
25 years
in software, deep in AI
1
granted patent
Microsoft + Accenture
production deployments
5+
products shipped
The background
The full history is a software career with AI running through the back half of it. Not reading about AI or advising on it, shipping production AI into large organisations and living with the consequences when it worked and when it didn't.
That includes senior AI roles at Microsoft and Accenture, systems deployed inside Fortune 500 organisations, and a granted AI patent.
The useful part of that history isn't the logos. It's month six. It's the systems that demoed beautifully and drifted quietly. The costs that looked fine at pilot volume and didn't at production volume. The tools nobody used because they were built for the process on paper rather than the process people actually followed.
Most AI advice today is written by people who have never had to keep a system running after launch. Ours is shaped almost entirely by that part.
What we believe
Most AI projects don't fail on technology. They fail because nobody agreed what the thing was supposed to do. The brief was vague, scope moved, and six months later something technically functional exists that nobody uses. A brief you approve before we start is the cheapest protection available.
The build price is the number finance tracks. The real cost includes running it, maintaining it, reviewing its output, and occasionally being wrong. Implementations chosen on the lowest quote are reliably the most expensive ones in the end.
Our audit and our build are deliberately different engagements. The audit report is written to be handed to any developer, including one who isn't us. If we're the ones who find the opportunity and the only ones who can act on it, you can't trust the finding.
We turn down work. When the process underneath is broken, when the volume doesn't justify the build, when the judgement genuinely doesn't transfer. Saying so early is worth more than a signed contract.
Every deliverable we produce is meant to lead to action. A report you can build from. Software your team uses. Nothing whose main function is to be presented.
Products and consulting
The products aren't a side business. Each one was built for our own work first, to solve a problem we hit ourselves: context costs, competitive signal, outreach volume. Each one has its own origin story on the product page.
Building our own products keeps the consulting honest. We're subject to the same constraints we advise on: token economics, evaluation coverage, quality drift, the cost of maintenance. It's difficult to give advice you don't have to live with when you're living with it in parallel.
See what we've builtHow we work
We take on a limited number of consulting engagements at a time. That's a constraint, and we'd rather state it than pretend otherwise. It means we occasionally say no to work that would be a fine fit, and it means the engagements we do take get senior attention throughout rather than a senior pitch and a junior delivery.
A small, deliberate studio. Working with businesses in the US, the UK, and India. Most engagements run remotely; on-site is available where the work genuinely benefits from it.
Find out for yourself
Free discovery session. You walk us through your business, we tell you honestly where AI helps and where it doesn't.
Or just write to us, we read everything.