Two areas of expertise that rarely exist in the same person — and that are both necessary to do this work properly:
Fifteen Years in IT & Enterprise Architecture
From 1998 to 2013, in France — enterprise deployments on rotation, then ten years as the outsourced IT department for around fifty companies, before selling the firm. The technical foundation for understanding how AI systems parse, structure, and rank product data.
Thirteen Years Operating Amazon FBA
Selling on Amazon since 2013, running my own brands — including a brand I operate today with my own tooling. Now a US-based operator (Miami) focused 100% on US brands, and the author of the quarterly State of AI Recommendations on Amazon report.
"I know what it takes for a listing to be understood, trusted, and recommended by an AI. I have operated in this marketplace since 2013 — I haven't just studied it; I've lived it."
What I do — and why
Amazon's shopping experience is shifting from keyword search to AI recommendation. Alexa for Shopping, Buy for Me, and the AWS Agentic Shopping Assistant all rely on machine-readable, well-structured product data to decide which listings to surface and recommend. Most FBA listings weren't built for that.
My work is helping founder-led brands close that gap — before it becomes a competitive disadvantage. That means AI-ready listing rebuilds, a free Recommendation Share, open tools for sellers, and guides that explain the underlying mechanics without the hype.
I built my first Amazon product in 2013. Over the years I've worked across sourcing, PPC, listings, and brand strategy. The AI turn in search is the most consequential shift I've seen in that time — and also the most actionable if you approach it structurally rather than reactively.
What I build
- Guides — in-depth breakdowns of how AI search actually works on Amazon, how COSMO, Alexa for Shopping, and agentic systems rank and recommend products, and what sellers can do about it.
- Embedded engagements — four to six weeks inside your team, leaving one workflow running end to end in your own environment. Remote, fixed scope, fixed price.
- ABOS / FirmAIn — a proprietary multi-agent AI operating system built to manage Amazon FBA brands with near-zero manual overhead. Specialized agents handle inventory forecasting, pricing elasticity, PPC allocation, portfolio governance, and M&A due diligence — all governed by a human-in-the-loop permission engine that prevents AI overreach. It's internal, but it's also the technical foundation behind Keoxs AIO.
- Keoxs Co-pilot — an open-source PPC management tool for Amazon sellers, available on GitHub.
- Keoxs AIO — a SaaS platform built to optimize Amazon listings specifically for Alexa for Shopping (Amazon's AI recommendation layer). Founded on Amazon's official AI research — it runs a 22-tool, 8-agent pipeline combining Claude, Gemini, and Llama. The public product that makes the FirmAIn architecture accessible at scale. Available at keoxs.com.
Connect
The best way to follow what I'm working on is the newsletter — periodic and worth-it, no filler. You can also find me on LinkedIn.
If you want to talk about your listing or brand, reach out directly.