First · the business
Your real growth drivers
Winning more customers, lifting revenue per customer, launching new products, taking cost out. Usually two or three things genuinely move the number. We find them before we mention AI.
The method, in full
How we work, and where we stop.
Most AI projects fail because they start from the tool. We start from the business, match the right kind of AI to the right kind of work, take each process only as far as it's worth, and hand it back to you. Three ideas carry the whole thing.
01 · We start from the P&L
The business first. Then the AI.
Everything begins with one question, asked in the right order. First: what does the business actually need to do to grow? Only then: where can AI help execute that? Get the order wrong and you buy clever tools that change nothing.
Then · the AI
Where it executes the strategy
Only now do we look at the handful of processes behind those drivers and ask what AI can do for each. The AI serves the strategy. It is never the strategy.
Make all fifty people 10% more productive and the business-level gain rounds to nothing. Point AI at the few processes behind growth and you actually move the number.
The difference between AI that feels busy and AI that pays.
Strategy in practice: for DAZN we turned scattered AI pilots into an operating model and a roadmap tied to strategic priorities — "clear focus with strategic thinking", in the words of Marcus Parnwell, EVP Product.
02 · Match the AI to the work
Four honest answers for any process.
The biggest mistake in the market is treating AI as one thing. It isn't. The kind of work decides the answer, and they run in ascending order of commitment. Most work should stop at one of the first two. The deeper two are earned by a few processes, not handed out by default.
Answer 01 · the honest one
Leave it alone
The first answer, and the one nobody else gives you. Where a person is cheap and a mistake is costly, forcing AI onto it is pure waste. We will tell you what to leave well alone.
Low-volume · high-stakes · cheap to do by hand
Answer 02 · sharpen
Assist the judgement work
The work only experience can do: analysis, proposals, design, decisions, client relationships. AI does the legwork and lays out the options at speed; your best people still decide. Where most knowledge work should stop. We sharpen judgement, we never automate it away.
Proposals · market analysis · creative drafts · advisory
Answer 03 · run it
Automate the repetitive work
High-volume, rules-based, low-judgement work that quietly drains the team. The machine runs it; a person owns the exceptions. Letting it run unsupervised is earned through proven reliability, not switched on by default.
Invoice intake · reconciliations · list-building · data entry
Answer 04 · own it
Build for the core
The few defensible, high-value, repeatable workflows worth owning outright. These become real software, with a codebase, a named owner, and IP that stays yours. Where our in-house engineering team lands. We build the hard part; we never subcontract it.
Bespoke platforms · the company brain · proprietary IP
The skill isn't pushing everything to the deep end. It's knowing, per process, exactly where to stop, and being honest when the answer is "not far at all".
03 · How we run it
A paid diagnostic, then prove it in days.
No twelve-month roadmap before you've seen anything work. We start small, prove the value fast, and only then commit to a bigger build, building alongside the tools you already run, never ripping them out.
01
Diagnose
A short paid diagnostic. We walk the business, find where AI actually pays, and rank the opportunities by value, not by what's easy.
02
Prove
We pick one or two high-value processes and build a working proof in days. Quick wins that matter to the P&L, not pilot theatre. Not every test works; that's why we test.
03
Build & hand over
We build alongside the CRM, ERP and tools you already run, then train a standing AI Champions group, one per function, on your real work, and leave them a short Business AI Manifesto to live by. You own the code, the IP and the capability. We step back.
Two things run in parallel the whole way: the work, and your people. A tool doesn't change a business; a team that owns it does. That's why the hand-over is the point, not the afterthought.
No hype, no pressure
See where AI pays in your business.
Thirty minutes, no pitch. We'll talk through where AI genuinely moves your numbers and, just as usefully, where it doesn't. Even if you don't end up working with us, you'll leave knowing what matters.