◗ AI for leverage · humans for the magic
You make the music. I cut the groove.
The tech. The launch. The ops. The books. The vision stays yours. Now go make the thing.
A tech, GTM, and operations partner for brands that make real things. Startup to scale, amplified by AI. I took Google Support into the LLM era at five billion users, with a team. Then I did it alone at Gatefold, all four seats. One build at a time, first real thing shipped in as little as a quarter.
One build at a time. If the fit's crook, you'll know by week two.
The model
An accelerator is a simple idea. A short fixed window, a very small number of companies, and someone all the way in it with you. Then it ends, and you're somewhere you wouldn't have reached alone in the time.
The version you know adds a cohort, a demo day, and a programme. This one strips the theatre. What I invest is leverage — the tech, the launch, the ops, and the books, run by one person moving like a floor of them.
No associates, no account managers, no committee. The reason isn't romance about working alone. It's that every one of them exists to sell you more of itself, and not one of them ships a thing.
You stay the one who decides. My job is to be the best support system you've got, not to tell you about your own business. The taste is yours. The speed is mine.
An accelerator with no stake in whether you accelerate is just a vendor with better branding.
The consulting model is built to bill, not to ship. So it got flipped on its head.
The part you'll feel first
Ask an agency a question and the answer takes a week. Not because the thinking takes a week. Because the thinking has to be socialised. Somebody builds a case, somebody books a call, somebody asks what the partner reckons.
None of that exists here. Nobody to convince, no case to build, nothing written for an internal audience that will never read it. You ask, and the answer comes back the same day with the reasoning attached.
Which also applies to whether we should be working together at all. You'll know by week two, not month six, and I'd rather lose the quarter than spend it being a bad fit politely.
Speed isn't the absence of rigour. It's the absence of an approval queue.
Why the seam
A groove is an analogue encoding of a wave. A feeling, cut into a physical thing, so a stranger can feel it again decades later. That's the whole trick, and it's the reason I don't think digital is the point. Digital only earns its keep when it makes something you can feel.
◗ Side A · analogue
◗ Side B · digital
◗ What I build when I'm the client
Every app I've shipped for myself is a digital product about an analogue pleasure.
gatefold.fm — a shelf, a daily pick, and a clerk with opinions. Vinyl, CDs, tapes.
Poker Clock — a blind timer for people sitting in a room together.
Chiefs Trivia — fandom, which is the least digital thing there is.
None of that was the plan. It's just what I reach for, and each one shipped in weeks rather than quarters. If your business makes something real, you probably want an operator who thinks the real part is the point.
◗ Dogfood · portfolio company no. 1
The record-collector app. Idea to App Store approval in thirty days on zero budget, launched on Reddit, and it hasn't stopped shipping since — an AI record-shop clerk, CarPlay, a turntable-side kiosk mode, and a Shopify channel for record shops now in review. AI in the build and AI in the product. My company, my money, every seat filled by me. Before I ran this on your business, I ran it on mine. Browse it live. It's the pitch, working.
One portfolio company so far, and the seat is open. That is the model, not a gap in it.
“You can feel the continued joy he’s channeling into this with each new update.” — APP STORE REVIEW · ★★★★★
// a consultant would draw you a 2×2 right here. thanks Accenture, they beat it into me.
Most operators have done one of these four. The rare part isn't the breadth, it's the crossing: people who ship 0→1 usually haven't run scale, and people who've run scale usually can't start from nothing. Almost nobody has done both on the product side and in the books.
Idea to the App Store in 30 days, solo. One codebase across iOS, Android, and web, plus a Shopify channel in review.
Catalogue, pricing, a working register, and cost of sales on a crate of records. A month end where there wasn't one.
Accelerated Google Support into the LLM era — my team carried it from insight to shipped product, on a surface serving 5B+ users. 6× chatbot accuracy. Then Claude and MCP across 17 engineering teams.
The recommendation engine and the Fulfilled-by-Amazon integration, the company's #1 OKR. $1.5B in revenue growth across EMEA.
◗ What actually changes
The ceiling is rarely demand. It's that every decision routes through you. What to reorder, whether to discount, what to tell the customer asking where their parcel is. Each one takes a minute and there are three hundred a week, so revenue scales with your hours and growth stops feeling like winning.
The normal fix is hiring, and for a business your size it's a trap. You need the revenue before you can afford the hire and the capacity before you can get the revenue. Then the hire needs managing, which you can't do, because you're the bottleneck.
Process built around AI puts the capacity in before the payroll. The parts of the week that shouldn't need a human stop needing one, the throughput goes up, and the new money pays for whatever's next.
And it's the same work that makes the business worth something without you standing in it. Owner-dependence is the first thing any buyer marks down.
◗ Where the money actually is
I ran Google's digital marketing centre of excellence and creative think tank, aligned with YouTube, for the top fifty brands in EMEA. So take this from someone who was paid to argue the opposite.
The gap between excellent digital marketing and merely adequate is small. The gap between a business that knows its margin and one that doesn't is enormous.
Everyone sells you the first one. Almost nobody will go near the third, which is where most brands your size are quietly bleeding — profitable on paper, no cash in the bank, and growing only makes it worse. AI transformation isn't a stage in this diagram. It underpins all four.
What that buys you · five fronts
The whole route to the customer. Positioning, funnel, lifecycle email, launch, and the number it all has to move.
→ the funnel mapped end to end · flows live in your Mailchimp or Klaviyo · a dashboard that says whether it worked
From a beer-coaster idea to a shipped product, then the version that survives scale. Chatbots that deflect real tickets, recommendation engines that earn their keep, and the go/no-go call when the honest answer is no.
→ a working v1 in real users' hands · evals in CI, not a demo that carks it on Monday · the reasoning written down
The unglamorous half nobody else will go near, and the half that decides whether scale kills you. Know what you own, what it cost, and what it earned. You can't scale a business you can't count.
→ a 13-week cash forecast off your PO calendar · contribution margin by SKU · a month end that lands by the 10th
Identity that leans all the way in, so you don't look like every other AI startup, and true cultural adaptation so it lands in the next market too. Led by a fluent French speaker. BA in French, GTM for EMEA's top 50 brands.
→ the identity system, in files you own · every string adapted, not just translated · one market live, then the next
Not a chatbot, not a workshop. AI without process is a demo. Process without AI is admin. The joining is the entire job, and it's the one I've done twice at scale. The creative calls stay human. So does the taste.
→ use cases ranked by payback, not by hype · the process running in your stack, not a pilot · runbooks, because a process is judgement you get to keep
The seat · one build, 1:1
Someone who understands the vision first, then helps you shape it and ship it. One build at a time, all the way in, as deep as it takes: ecommerce, recommendation engines, GST registration, transfer pricing, the lot.
Twelve weeks to something real. Then as far as it goes.
The two nobody else combines
Someone who can ship the product and close the books. Most people who can do the first have never done the second, and the ones who can do the second have never shipped anything.
How it runs
No standing meetings about meetings. The first real thing ships in as little as a quarter, and then you get the keys. Some of it takes longer than twelve weeks, and scale always does, so it runs a quarter at a time and you renew when the last one earned it. I'd rather launch you and leave than be a line item that never ends.
Two weeks of riffing. I get across the business, the taste, and the real goal. Then we name the one thing worth building, and the number it has to move.
We scope it for real. Funnel, stack, edge cases, the ATO-and-recommendation-engine stuff nobody warned you about. A plan you could hand to anyone. You won't have to.
I build it with you, in the open. AI for leverage, TPM rigour for the finish. You end the quarter with a shipped product, not a deck.
◗ The deal
From €50k a quarter.
The same trade the famous accelerators run, with one difference. They invest and run. I invest and operate. Leverage mostly, capital where it fits. Never hourly, because hourly just pays me to be slow, and we'll sort the rest on the call.
Receipts, not promises
None of it was some solo act of genius. It was orchestration, which is the actual job. Connect the people, the teams, and the pieces into one thing that ships. Imagine the dots I can connect for you.
◗ About me · where all four corners come from
Twenty years of being the bloke who gets it shipped. Physical goods at Zappos, creative GTM in Paris, global ML at Google, AI across seventeen teams at Affirm, and then a company of my own to prove it still works when nobody's holding the budget.
I helped launch Google's Star Trek computer. Now you can have your own.
The Verge “…the most successful by far.” On Project Kennedy, Google’s first cross-product redesign, which I led as TPM. · read ↗The work, in the press
◗ Atelier Le Nous Royal · la galerie
Proof we ship: et que nous sommes drôles. Built for the love of it, then made to actually work. Range, humour, and a finished product. Pick three.
Fan-trivia game for Chiefs Kingdom. Brand obsession, gamified.
A tournament blind-timer that actually looks decent. Structure for chaos.
An app for vinyl/CD/tape nerds. One shelf, a daily pick, a smart-arse AI clerk. Built & shipped solo. iOS · Android · Web.
Turn a crate of records into a priced, scannable catalogue and a real register, folding table to full storefront. Live demo on the page.
A Shopify app for record shops — scan the barcode, ID the exact pressing, and the collector-grade listing writes itself. In App Store review.
◗ Who I partner with
Founders and creatives who've already made the thing and want it to go further. Indie labels, makers, D2C, creative commerce with taste and a story (Gatefold is one of mine). If you build something and want more, better, faster, soul fully intact, we should have a chat.
◗ Book a fit call · one build at a time
The seat's open, or it isn't. One call tells us both. The person you meet is the person who ships. Bring a concept, a total mess, or a maybe.