I’ve spent years aboard transformation projects, and I’ve learned something that seems obvious but almost nobody applies: you don’t sail a new sea with the same old crew. You put four brilliant technicians in a room, give them a use case and a budget, and three months later you have a beautiful prototype that nobody in the business knows what it’s for. I’ve seen it more than once.
AI isn’t just technology. It’s technology, plus process, plus data, plus business rowing in the same direction. And if one of the four is missing, the boat takes on water, even if it takes a while to notice.
The first is the captain, the business profile. They set the course, raise the real need —not the one that sounds good on a slide— and say clearly what’s needed and what for. Without them, you sail with no destination.
The second is the cartographer, the process profile. They have a structured mindset, the character to question the established and the rare ability to look at how work is done today without falling in love with it. Their job is to redraw the map, analyse the current process and propose how it should look to really work with AI, not to dress up the same old thing with a layer of technology on top.
The third is the engineer, the technical profile. They build the boat: they assemble the solution, decide the architecture, connect the pieces so they withstand the storm. They’re the one who turns the cartographer’s map into something that actually sails.
And the fourth, the boatswain, the one almost always forgotten until it’s too late: the data specialist. They look after the hold, prepare, clean and handle the information that’s going to feed the project. Because you can have the best boat in the world, but if the cargo is rotten, you won’t reach any port.
Four roles. Four different perspectives. One single crew.
And here comes what really makes the difference: the final validation isn’t done by one person alone, it’s done by the whole crew together. The business one checks that they’ve arrived where they wanted to and that the journey delivers the value expected. The process one checks that the defined route works, meets expectations and has no flaws. The technical one checks that the engine performs, with quality and good throughput. And the data one checks that what goes in and what comes out is correct, with no leaks or noise along the way.
When all four give the green light, then yes, then you’ve reached port. Before that, you were just sailing on luck.
And you? How many of these four profiles do you have on your next AI project? Which one is always missing in your organisation?