I find your highest-leverage move and build the proof of it.
You already know AI will matter for your business. The hard part is knowing where it actually pays, getting your people to run with it, and tuning out the noise. So I work like your corner, not a contractor: we spar, I analyse and find the opening, then build the proof, and your team comes out able to throw the punch themselves.
Book a leverage sessionA 30-minute sparring session, no pitch. If we're not the right fit, you'll know by the end, or I'll tell you.
Roy Bernheim
Founded & built
$0 to $8MAI-systems-driven consumer brand, 120 countries
Advised & shipped
Data strategy to AI productionEnterprises and forward-leaning SMBs
Span
Strategy to systemanalytical first, builder second
The problem
The hard part isn't AI. It's knowing where it pays.
You already know AI will matter. The technology is real and the opportunity is real. What's missing is a clear read on where it actually moves the needle in your business, and the judgment to ignore the 90% that doesn't.
And if you've already started, you've felt the other half of it: tools bought, pilots half-finished, output nobody fully trusts. Motion, but not leverage.
The leverage was never in the tools. It's in knowing which move to make first, which to leave alone, and getting your people to actually run with it. That's the call I help you make, and then build.
How I work
Top-down. Find the one move, prove it, then make it stick.
Goal
Start from the underlying commercial goal, not the tool. What actually needs to move.
Leverage
Drill in surgically to the single highest-impact, lowest-effort zone. The one move.
Proof
Build the working proof of it, in weeks. Something real you can put in front of your team.
Adopt
Get your people running it, then roadmap and automate what works, so it compounds without depending on me.
How we work together
It starts with a conversation.
Every engagement begins the same way: I find your highest-leverage move and prove it works before you commit to anything bigger. You carry less risk, and you see something real fast.
The Leverage Session
Free · 30 minA sparring session on your business and where AI actually pays, and where it is quietly costing you. You leave with a clear read on your highest-leverage move, whether or not we end up working together. No pitch.
The Leverage Diagnostic
A short, fixed-scope paid engagement, usually a few weeks. Whether you arrive with a hunch or a blank page, I find the single highest-impact move and build a working proof of it: a real, working system, not a slide. You keep everything I build. If we continue, it counts toward your first month.
Embedded build
A monthly engagement where I run the roadmap with you: the strategy, the build, and getting your team to actually use it. Highest-leverage move first, then the next. Systems that compound, not one-off fixes.
Advisory
For owners, CEOs, and operators who want me in the room on the big AI calls. Selective, a handful of clients at a time, by application.
The lever, then the proof
Where I found it, and what I built.
Wanted AI to matter to a core part of its business, not just sit in a strategy doc. I led the AI strategy, prioritized the use cases, and built the highest-value one into a working decision-support tool the team now uses to plan with evidence instead of intuition. Strategy to a live tool, in one engagement.
Was running outbound entirely by hand. I diagnosed outbound as their single highest-leverage zone and shipped a Slack-native AI agent that clarifies the ideal client, finds and enriches contacts, writes tailored outreach, and runs the whole follow-up sequence end to end. A sales motion that runs itself instead of eating the team's week.
Was invisible in AI search, the place its customers now ask before they buy. I restructured its science content into something the AI engines actually cite, and built a loop that tests and improves it continuously. Roughly a 30% lift in AI-search visibility in the first three months.
Treated customer support as a cost to minimize. I saw a revenue channel. I built an AI agent grounded in the brand's catalogue, stock, and promotions that answers questions and guides the buyer toward a purchase. Support started converting engaged visitors at 5x the prior rate, and a cost center turned into a revenue line.
Background
Over a decade turning data into commercial systems, the last five years building with AI, for global brands and the owner-led companies that move fast.
University College London · London School of Economics · Columbia Business School · TEDx speaker · Featured in Forbes and Entrepreneur.com
Full background on LinkedIn →Why I bridge both worlds
Analytical first. Builder second.
At the outset of my career I worked on the analytical side of business, tech, and finance, finding what actually moves the numbers across commercial strategy, brand, and data. Later, I co-founded and scaled a consumer brand from nothing to $8M across 120 countries, and learned what moves a business before I ever wrote a line of production code.
Then I co-founded and led Decommerce, building the systems that put some of the world's leading brands' data to work with AI. Now I do it independently, at companies from Coop and Breitling to forward-leaning SMBs.
Think Angelo Dundee, the man in Muhammad Ali's corner. Dundee never threw a punch. He read the fight, found the opening, said the one thing between rounds that turned it, and got the best out of his fighter without ever taking the spotlight. That's the corner's job, and it's mine: we spar on your business until the real opening shows up, and I stay in your corner through the messy part, getting your people to actually use what we build, because the hardest part of AI was never the model, it's the humans around it. Where I go one step further than Dundee: I also build the thing, the working proof, not just the game plan. You come out sharper and able to throw the punch yourselves, not waiting on me.
Strategy is cheap when it can't execute. Building is cheap when it doesn't know what matters. And both are useless if your people won't touch the result. The hard part is doing all three at once: finding the move, building it, and getting your team to run with it. You're hiring the judgment that connects all three, and a vetted bench when scope needs more hands.
Questions
Before you book.
What happens on the leverage session?
Thirty minutes on your business and where AI fits. You walk away knowing where your biggest AI opportunity is, and where it is not. No pitch. If we are not a fit, I will tell you, and probably point you somewhere better.
My team already uses AI. Why you?
Most teams have AI doing things. Fewer have it doing the right things. I'm not another tool. I'm the judgment that decides which move actually pays, the hands that build it, and the one who gets your people to run with it.
So, are you a developer?
Yes, I write code and I've shipped AI tools that clients run in production. On bigger builds, though, I bring in vetted specialists and developers from a network I've built over a decade in the field, so each part gets the most dedicated expertise, and you keep me on the levers that actually move the needle, not buried in the plumbing.
What does this cost?
It depends on the shape of the work, so we settle it on the call rather than guessing here. The first paid step is deliberately small and you keep everything I build, so you can judge me on a real result before you spend more. I price for commitment, not to be cheap: a price that costs you nothing rarely changes anything.
How fast do I see something real?
The first engagement is built around a fast, visible proof. Usually weeks, not quarters. You don't wait months to find out whether it works.
I've been burned by AI consultants before.
Usually that happens because they hand off the build, or never own the outcome. I do both the thinking and the building, and I stay accountable for what ships. If the proof doesn't land, you find out early and cheaply. And I build systems to keep running, not demos that look great at launch and rot a month later.
Are you a one-person shop? Can you handle our scale?
I keep a deliberately small client list so each one gets senior attention, and I bring in a vetted bench of specialists I have shipped with before when scope needs it. The systems I've shipped run at some of Europe's largest companies.
How much of my team's time does this take?
Less than you would expect. The Diagnostic runs mostly on my side. I need a few focused conversations and access to the right data, not standing meetings. The point is to give your people time back, not take more of it.
Who do you work with?
Companies led by their owners, CEOs, or operators, roughly EUR 10 to 50 million in revenue, where someone can make a decision without a procurement maze.
Not ready to talk? Borrow what I've tested.
On LinkedIn I share the AI tools and systems I have actually put into production: what is worth your time, what is slop, and where the real leverage is hiding. Follow along.