Overview
Thierry Moubax is co-founder and CEO of ai-compass, a company he started in May 2023, and is based in Majadahonda. He states expertise in artificial intelligence, ChatGPT, marketing strategy, digital marketing and e-commerce. Since June 2010 he has also run BlueCompass.eu, where he works in interim management and consulting, developed educational apps until October 2015 and wrote the book The Online Marketing Blueprint.
Before founding ai-compass he was CMO of Knn Digital Media from October 2022 to June 2023, and Chief Marketing Officer at Allianz Partners from October 2019 to December 2022. From 2015 to 2019 he was Vice President Marketing & Product Europe in the AI division of bpost, and earlier held roles as Sales & Marketing Director at Moovly and Director Marketing & Innovation at Securitas. Between 1997 and 2010 he worked for DHL Express, in business development, market and product roles at DHL Express IBERIA, as Marketing Director at DHL Express España and as Marketing Strategy Director Europe at DHL Express Europe. He began his career at SHELL Belgium as a business analyst and then retail sites manager.
Moubax studied commercial engineering with a marketing speciality at KU Leuven, with periods at Università di Siena and Universidad de Granada, and later took a blockchain programme at MIT Professional Education.
Career history
In the news
- Jensen Huang was asked, in front of a room full of government delegations, what the worst thing is that could happen to a country in this technology shift of AI. His answer: "The worst outcome is that you don't take advantage of it. That you are left behind. That is the single worst outcome." You will now say, yes of course: He runs Nvidia. Fear slows his order book. But listen also what he said next: Airlines once spent their advertising money telling you their plane was safer than the other airline's plane. Nobody bought a ticket
- It's the first of September. Across Belgium children are going back to school this morning, and I still remember those mornings well. Almost everything about the world has changed in thirty years. Most of what happens inside the classroom hasn't. Yes, we put computers in the schools. Then we largely taught children to use a word processor and a spreadsheet. So where is AI on the timetable this year? How are we preparing these kids for the world they'll actually work in? Nearly four years since ChatGPT launched, and still no
- Here is a fifteen-minute exercise worth more than most AI strategies: --> Take the last AI output that disappointed you. --> Write down exactly what you gave it: The instruction, the files, the examples, the standard you set for good. Now, Ask yourself: if a new hire had received exactly that, would you have expected better? Most of the time the honest answer is no. Then do the same thing forwards. --> Screen-record yourself doing one recurring task once, and score each step on three things: * how often it happens? * how long it
- A paper published last week did something quietly useful. Researchers took passive screen recordings of people doing ordinary work. No interviews, no workshops, no self-reporting. Just screenshots and keystrokes. From that alone, the system separated out the distinct tasks each person was actually performing, with 0.974 agreement against ground truth. It rebuilt around 75% of the real steps. Automations built from those task models beat the previous best method by 30% on work they'd never seen before. The paper is called
- Think about the best hire you ever made. Now give that person a desk, no brief, no examples of good work, no access to the files, and nobody to ask. Two weeks later you read their output and decide they're mediocre. That's roughly how most companies are running AI. Between 90 and 95% of companies report no meaningful return on it. Your competitor is running the same models you are, at the same price, released the same week. What differs is what each of you puts in front of them. AI behaves like an employee. I wrote a whole book
- De Tijd bracht het twee keer op een maand tijd. Ben Serrure en Wim De Preter schreven over IT-budgetten die ontsporen en over bedrijven die naar Chinese modellen grijpen om te besparen. Twee goed onderbouwde stukken. Twee keer over echt geld. Twee keer op dezelfde budgetlijn, en daar wringt het. Ik noem het de Lamborghini-fout. Opus 4.1 kostte $15 per miljoen inputtokens en $75 output. Opus 5 kost $5 en $25, voor een model dat dubbel zo hoog scoort op de Artificial Analysis Intelligence Index. Voor mijn boek "The AI Compass" zocht ik
Related profiles
This page shows public professional information only, each fact cited. Is this you? send a correction, or ask for removal within 24 hours, no questions asked.