APPLIED ML IN ADVERTISING SINCE 2012 · MEDIA BUYING SINCE 2014
I run your paid campaigns with the tools I spent a decade building, and with the judgment none of those tools have.
The machine proposes. A human decides, and signs for it.
This is no longer a hunch, it is what I see account after account. A decision that takes three meetings and two committees arrives too late in a market where the auction reprices every hour.
Technology changed the balance. A tight team with the right tools now runs more tests, reads results sooner and corrects earlier than an organization ten times its size.
That is exactly where I work: with the people who can decide the same day.
Every engagement activates the full capability stack — not a subset, not a package tier.
Growth roadmaps, channel mix, budget phasing, and audience architecture — before a single euro is deployed. Every engagement begins with a strategic brief that aligns media with business model.
Two rare skill sets, learned in parallel, for more than ten years. It is the only thing that explains what I do today.
I cofound Skylads, out of the HEC Paris incubator. We build bidding and optimization algorithms for advertising, at a time when nobody puts the word AI on a slide. I learn what a model can do, and more usefully, what it cannot.
I start buying media myself. Understanding a bidding algorithm is not enough. You have to live with its decisions, with a real budget in front of you. That shift changed how I design tools.
I cofound Ad360 and spend several years building advertising technology: tracking, attribution, automation. I see from the inside why most promises of full automation do not hold.
Thirteen years cofounding AdTech and AI products, and an ad account open every day. I do not choose between building and operating. The combination is the value.
Grande École Master, HEC Paris. Degree in applied mathematics.
Most talk about AI in marketing stops at the word. Here is the exact split, as I apply it on an account.
ALWAYS THE MACHINE
NEVER A MODEL
On the right, there is never enough data to settle it. There is only experience. That is what you pay for, and that is what does not automate.
Three productized offers, powered by my own AI tools. Each one answers a precise moment in the life of an ad account.
From brief to a complete, multi-angle campaign structure, ready to run — in a fraction of the usual time. AI generates, the expert arbitrates.
Before spending a single euro: tracking, pixel & CAPI, product feed, landing pages and creative assets checked and fixed. The account is ready to scale.
A deep, AI-assisted audit of your existing Meta & Google accounts: structure, tracking, creatives, budget allocation — and a prioritized action plan.
I take read access to your accounts and tell you what I see: what is broken in measurement, what is badly structured, what you are paying for without knowing it. You keep the document, whether we work together afterwards or not.
WRITTEN AUDITI run paid acquisition: account structure, budget, creative, bidding, testing. You see the decisions as they happen, each with the reason behind it. The account, the data and the automations stay yours.
MONTHLY ENGAGEMENT · WEEKLY REVIEWWhen the tool you need does not exist, I build it: server side tracking, cross channel attribution, dashboards, creative production at scale, reporting automation. That was my first job.
FIXED SCOPE PROJECT · CODE AND ACCESS HANDED OVERI don't pretend to do everything. I connect my clients to the best operators in each specialty — and I use them myself.
E-commerce growth partner on a 100% performance model: paid by commission on the sales they generate. Zero fixed risk, fully aligned incentives.
bonzaimoon.com ↗
Ad intelligence platform: real-time spend rankings, a database of winning ads and scaling funnels, updated daily. My competitive radar.
whoscale.io ↗
AI-generated static ads built from your customers' real language: 30 visuals in minutes, scored angles, Meta-ready formats. My creative factory.
staticads.app ↗PARTNER DISCLOSURE — I have a commercial relationship with the partners listed above. I recommend them because I use them myself and believe they represent best-in-class capability. All recommendations remain my independent editorial judgment.
THIS WORKS
THIS DOES NOT
I do not publish results I cannot document. The numbers from my accounts are under NDA. I show them live, inside the account, with the period and the volume.
A thirty minute call, no sales deck. Describe your situation, your channels and your main constraint. If I am not the right person, I will say so on the call.
What does hybrid intelligence mean in practice?
It means a clear, explicit split between what the machine handles and what I handle. The machine runs volume: tests, monitoring, scoring, bidding, reporting. I make the decisions the machine cannot: choice of offer, budget arbitration, reading a weak signal, deciding to stop. That split is written out on this page. It is not a metaphor.
How is this different from an agency that uses AI tools?
I built the tools from scratch, starting in 2012, before AI was a selling point. An agency that adopted AI tools in 2023 has a different relationship with them: they use them but do not understand their limits from the inside. I spent more than a decade learning exactly what a model can do and, more usefully, what it cannot.
