As maritime companies rush to deploy artificial intelligence across chartering, operations, compliance and commercial workflows, a new report from Veson Nautical argues that the industry’s real challenge is no longer whether to invest in AI, but whether the underlying foundations are strong enough to support it. The report, Maritime AI Foundation, frames the issue around five questions shipping executives should be asking now, as AI adoption accelerates across the sector. The first is whether AI systems are genuinely maritime-specific. General-purpose AI can recognise patterns and generate text, but shipping requires contextual understanding of voyage economics, laytime clauses, demurrage exposure and operational dependencies. The report argues that systems trained specifically on maritime workflows will ultimately outperform generic AI tools layered onto shipping operations. The second question concerns proprietary operational data. Maritime companies increasingly understand that competitive AI advantage will come less from public large language models and more from the quality of their own structured operational information. If voyage, port call and commercial data are fragmented or poorly captured, AI outputs risk becoming unreliable. Third comes workflow integration. Rather than functioning as standalone assistants, the report argues AI must be embedded directly into commercial and operational processes. Contractual terms, cost assumptions and voyage logic need to flow automatically through systems in real time if AI is to move beyond experimentation into daily execution. The fourth pillar is the system of record itself. Many shipping companies still operate across multiple disconnected platforms and spreadsheets, creating competing “versions of reality” around voyages, contracts and market exposure. AI systems built on inconsistent data foundations risk amplifying confusion rather than improving decisions. The final question focuses on scale and network effects. AI systems improve through exposure to broader operational contexts, meaning platforms with larger client communities and wider shipping datasets may gain a structural advantage over isolated in-house deployments. Those themes echoed strongly during last month’s AI, Digitalisation and the Dry Bulk Workforce session at Geneva Dry, where panellists repeatedly returned to the tension between technological capability and organisational readiness. “Most of us are probably still trying to figure out how we’re going to deploy AI and not yet worried about the governance of AI,” Scott Bergeron of Oldendorff Carriers told delegates attending the Swiss conference. Moderator Cynthia Worley of Sedna delivered an early jolt to the room, noting that the EU AI Act enters full enforcement later this year, carrying fines of up to €35m or 7% of global annual turnover for companies unable to demonstrate governance of their AI processes. A show of hands confirmed that almost nobody present had heard of it before that week. Bergeron drew a pointed analogy with radar to frame AI’s promise and its limits. “There have been plenty of radar-assisted collisions,” he noted. “So it wasn’t the final solution.” His deeper concern was generational. “What happens 10 years from now when there are no more subject matter experts? Who’s going to be around to question the output of AI?” Ingrid Kylstad of Klaveness Digital pushed back on the radar comparison. “I actually think AI is more transformational than the