The First Rule of Maritime AI
AI initiatives succeed when they begin with real operational problems, not technology. A problem-first approach builds trust, drives adoption, and ensures digital investments deliver measurable business value. As one operator put it bluntly: if you don't start with the problems, you're going to be screwed.
One of the most common reasons AI initiatives fail in maritime is surprisingly simple: they start with technology instead of problems.
When organizations begin with questions like 'How can we use AI?' they often end up with solutions in search of a purpose. The technology gets implemented. The dashboards get built. And then nothing changes, because no one has articulated what success actually looks like or connected the tool to decisions that matter.
"If you don't start at the problems right, then you're going to be screwed."
— Nils Israelsson, Head of Operations, Stena Bulk
This is not a theoretical concern. It is the lived experience of operators who have watched digital initiatives stall, of technology teams who have built features no one uses, and of executives who have approved budgets without seeing returns.
The Fear of Missing Out
Part of what drives technology-first thinking is the fear of being left behind. When competitors announce AI initiatives and vendors promise transformation, there is pressure to act—even without clarity on what action to take.
"The first thing you have to do is just looking at what is my business and what does my business need. Especially with those hypes, AI defense, that you act on the fear of missing out. It's good to first think what it can do for you and then act according to a plan to make it happen."
— Felix Jan van den Bos, Independent Digital Transformation Consultant
This does not mean organizations should wait indefinitely. The risk of not doing anything absolutely outweighs the risk of doing something in a controlled way and having it go wrong. But action without direction is waste disguised as progress.
What Problem-First Looks Like in Practice
At Stena Bulk, the approach has been systematic. Rather than asking 'what can AI do?', the operations team has mapped specific problems, assessed their business value and complexity, and prioritized accordingly.
"We took a big step roughly two years ago and invested quite a lot in this development, upskilling ourselves and then also really detailing all the problems that we had. So we put them on a scale—business value and complexity—and then we pointed out, okay, these are the ones that we should actually go for."
— Nils Israelsson, Head of Operations, Stena Bulk
This structured approach serves multiple purposes. It ensures resources flow to high-impact opportunities. It creates clear criteria for evaluating success. And critically, it builds organizational alignment before implementation begins.
The 900 Definitions Problem
A problem-first approach also forces precision about what operators actually need—which varies enormously across the industry.
"There's about 900 different definitions of what performance is. Bunker optimization is a good answer for Stena Bulk, but probably not if you're a container liner and probably not if you're dry bulk or probably not if you are an owner versus an operator or a ship manager."
— Alexandre Lapointe, Chief Product Officer, OrbitMI
This is why generic AI solutions often disappoint. They are built for an average user who does not exist. Effective AI adoption requires specificity: what decisions need to be made, by whom, with what information, under what constraints. Only then does it become clear whether AI is even the right tool—and if so, what kind.
AI Is a Means, Not an End
Perhaps the most important mindset shift is recognizing that AI is one tool among many. It is not a strategy in itself.
"AI is a mean, it's not an end. When upper management says 'we want to do AI', it would be the same as saying 'we want to do software.' What does that mean? You need to go back. What exactly do you want to do? What is the problem you're trying to solve?"
— Alexandre Lapointe, Chief Product Officer, OrbitMI
Sometimes the answer to an operational problem is process improvement. Sometimes it is basic automation. Sometimes it is better integration between existing systems. AI should compete for resources against these alternatives based on merit, not hype.
Organizations that internalize this—that treat AI as a potentially powerful means rather than an end in itself—are the ones building sustainable capability rather than collecting unused tools.
This series is based on a webinar from Digital Ship. Watch the video here
