Maritime software has forced commercial teams to act like clerks, moving data between disconnected systems. AI is starting to change that, writes Will Robinson, the founder of HelmOS, shifting the battle from who has the best database to who controls the surface where decisions are made.
No broker or commercial manager thinks in terms of tabs, fields and forms. It’s always positions, cargoes and risks. Yet so many smart shipping dealmakers are being forced to work like clerks because their support systems aren’t up to the job. Open one system. Enter one fact. Get one result. Copy that to another system. Five times per email, fifty emails per day. That’s a weak and fragmented way to work; it slows response times, loses information between systems, and gives people a poor foundation for making decisions.
How did we get here? Historically, software vendors have each built three things: a database, a backend and a frontend. The database holds the information. The backend connects to the database and runs calculations; together they form a data layer. The frontend is the surface the user sees. Each software package has a different function, from voyage calculations to market information, and cross-system communication is difficult. The most common way to integrate these systems is to have a human copy and paste it all together.
AI changes this in two critical ways. First, AI-assisted coding has drastically shortened the build time for new features and connections. A feature which might have needed months now takes weeks, perhaps even days to build. This makes it much easier to connect data from different sources reliably and quickly. Second, AI is great for reading and sorting data taken from unstructured sources such as emails, WhatsApp messages and web pages. Now users can get instant feedback when an email comes in, or ask a question in one place and get an answer from five trusted sources.
This does not mean the underlying data is being commoditised; quite the opposite. As data gathering becomes more automated, bad data travels further and faster; trustworthy data becomes yet more valuable. Data providers have taken note and consolidated. Kpler bought MarineTraffic, FleetMon and Spire Maritime, then secured $1bn of growth equity investment. Veson added Q88, VesselsValue, and Oceanbolt while Signal Ocean acquired AXSMarine. Unique, hard-to-generate data is being treated as the scarce asset it is and being positioned as strategic infrastructure for commercial teams.
Now that these consolidated data layers exist, what will users actually interact with each day? Probably not a separate surface for each data source. Instead of acting as clerks passing data between carefully curated silos, it’s more likely we will spend time on a surface that works for us. It will understand what we are trying to achieve and connect us to the data layers we need in that moment.
This surface cannot just be some chatbot. It needs to truly understand the workflow, what context is needed for a decision, where to find it, and how to show its sources clearly. It needs to place its logic in repeatable deterministic code, leaving an audit trail so its responses can be verified and trusted.
Recent feature launches show this is already happening. The Veson Platform offers a suite of products with different functions, from mail to vessel data to voyage management, in one place. Sedna’s integrations with IMOS and GeoServe are similar, with voyage and port DA data shown directly inside the user’s workflow. While these started from different angles, both show a rush towards ownership of the surface.
In reality, the next few years in maritime software may not be a contest between incumbents and startups, or between AI and non-AI solutions. It may be a fork in the road between those who wish to own the most trusted data layers and those who wish to provide the surfaces users interact with every day. The data layers will compete on depth, accuracy and uniqueness while the surfaces will compete to be the most connected, intuitive and efficient. The ultimate winners will be the users: better data layers below and a better surface above for faster, more informed commercial decisions.
