AI in shipping: from automation to strategic integration

Artificial intelligence is becoming more deeply embedded in both boardroom strategy and daily operations across shipping, and Island Oil Holdings chief technology officer Gregoris Gregoriou believes its real value lies in integration rather than experimentation.

Speaking in an interview with Cyprus 4.0, Gregoriou outlined how Island Oil is applying AI and automation across the business, from trading and compliance to internal workflows and finance, while stressing that the company sees technology as a way to enhance human capabilities, not replace them.

Asked what artificial intelligence means in a maritime context, Gregoriou described shipping as “a highly dynamic and fast-moving industry,” adding that bunkering operates at an even more intense pace, in “an extremely volatile environment where pricing fluctuates rapidly and commercial decisions often need to be made within minutes.”

To illustrate the point, he referred to the speed of trading decisions.

In Island Oil’s operations, he said, the time between receiving a fuel request and submitting a competitive quotation can be as little as ten minutes. Within that narrow window, traders must simultaneously assess market prices, vessel location, port conditions, supplier availability, credit exposure and historical data.

Gregoriou said “this is where artificial intelligence becomes truly valuable.”

He described AI in a shipping and bunkering group as an intelligent assistant. Rather than manually running multiple reports or moving across different systems under time pressure, he said traders can interact with AI tools, for example through an internal chatbot, and instantly receive consolidated insights.

As he put it, traders can ask targeted questions about historical pricing trends, customer exposure, margin analysis or risk indicators and receive immediate, data-driven responses.

Still, he stressed that AI does not replace experience.

“AI does not replace the trader’s judgment or experience; rather, it enhances it,” he said, adding that it allows people to make faster, more informed decisions, reducing reaction time while increasing accuracy in a competitive and time-sensitive market.

Turning to the broader purpose of AI, Gregoriou said Island Oil primarily sees it as a way to improve efficiency and compliance, with competitiveness emerging as a result.

“For our organisation, AI is primarily a tool to strengthen efficiency and compliance. Competitiveness, in our view, is the natural outcome of getting those two elements right,” he said.

He first linked AI to efficiency in fast-moving trading environments, where timing and clarity are essential.

The company, he said, uses AI to provide traders with immediate decision-making insights, allowing them to act quickly and confidently. At the same time, repetitive procedures are being automated so employees can focus on work that requires judgment, experience and critical thinking.

“Technology should remove friction, not replace expertise,” Gregoriou said.

He then turned to compliance, noting that operating in more than nine countries means dealing with multiple regulatory frameworks as well as strict internal governance policies. AI-driven systems, he said, help monitor transactions, documentation and workflows in a more consistent and proactive way.

That, he explained, reduces risk and ensures compliance is embedded in everyday operations rather than treated as a separate control function. When operational efficiency improves and compliance is strengthened, competitiveness follows organically, as “a byproduct of discipline and smart execution,” rather than as an isolated goal.

On strategy, Gregoriou said Island Oil never intended to pursue isolated pilots in silos.

“From the very beginning of our AI journey, we made a conscious decision: we did not want to implement isolated, stand-alone AI applications that automate processes in silos,” he said.

Instead, he described the company’s objective as integration, ensuring AI coexists with employees and supports their daily work.

“AI at Island Oil Holdings is not designed to replace people; it is designed to augment them.”

In practice, he said, the group identifies operational processes where AI can work alongside users, strengthening capabilities rather than removing human involvement. He described this as an approach based on augmentation, combining human judgment with machine speed and analytical power.

As one example, Gregoriou said Island Oil is testing various AI models to provide faster insights to traders through an AI chatbot embedded directly within its ERP system. Rather than moving between multiple screens or reports, traders can ask specific questions and receive immediate, consolidated responses within their normal working environment.

As a second example, he said the group is exploring AI assistants capable of reading incoming emails, extracting relevant commercial information and initiating the first step of the trading cycle by capturing customer requirements into the system.

“This reduces manual data entry and shortens reaction time, while the trader remains fully in control of the commercial decision,” Gregoriou said.

He summed up the company’s approach to embedding AI into daily operations as making it “invisible but impactful”, integrated into workflows, supportive of users and aligned with the way teams already work.

Asked when AI stopped being a pilot initiative and became a strategic pillar, Gregoriou said the shift came when adoption began to spread organically across the organisation.

“AI became a strategic pillar for us the moment we saw adoption growing organically across the organisation,” he said.

He explained that the company initially introduced AI through small, carefully selected pilots involving specific user groups, allowing it to test practical use cases, measure impact and gather feedback.

The real turning point, he said, came when employees in other departments started asking to be included in future initiatives.

“That was the turning point,” Gregoriou said.

For him, that was clear evidence that people were not resisting the technology. On the contrary, he said, employees quickly recognised the value AI could bring to daily work and saw that automation reduced repetitive tasks, improved speed and supported better decision-making.

Gregoriou also linked this transition to clear messaging from management. From the outset, he said, the company made clear that AI was not being introduced to replace employees, but to coexist with them as an assistant rather than a substitute. That reassurance, he said, helped build trust.

When employees understand that technology is there to empower rather than threaten them, he added, adoption accelerates naturally. At that point, AI stops being an experiment and becomes part of the company’s strategic direction.

Discussing the Island Oil Agent chatbot, Gregoriou said it was developed as an internal knowledge assistant to help employees navigate company policies and working procedures more efficiently.

As organisations grow, he said, policies become more detailed and updates more frequent, making alignment more difficult, particularly across multiple countries and departments.

By integrating the chatbot directly into Microsoft Teams, he said the group placed the tool where employees already work and communicate every day. Instead of searching through folders, emails or manuals, users can simply ask a question.

Those questions, he said, can range from simple requests, such as how to submit a business travel request, to more complex procedural guidance, including the full workflow from request approval to procurement and purchase of a specific item.

Gregoriou said the impact on workflows has been significant, reducing time spent searching for information, minimising misinterpretation of procedures and ensuring employees are guided by the latest version of company policies.

In his words, it creates “a single, reliable point of reference,” improving productivity, consistency and internal compliance across the group.

On predictive analytics, Gregoriou said the initiative is moving forward steadily, although the company remains in the early stages of full predictive maturity.

At present, he said, Island Oil’s main strength lies in descriptive analytics. Through business intelligence reports, the company provides visibility into historical performance, sales trends, profitability, credit exposure and operational metrics.

That, he explained, is an essential starting point, because reliable predictive analytics can only be built on structured, high-quality historical data.

Gregoriou said Island Oil has already moved into the next stage, testing predictive models designed to anticipate developments, forecast sales behaviour, margin evolution and identify early signals related to credit risk.

The aim, he said, is to give users not only insight into what has happened, but also into what is likely to happen.

“Our longer-term ambition is to move toward prescriptive analytics,” Gregoriou said.

He explained that prescriptive analytics would allow systems not only to predict outcomes but also to recommend corrective or preventive actions, advising users on what adjustments should be made to mitigate anticipated risks or capitalise on emerging opportunities.

While the company has not yet reached that stage, he said it is building the foundations in a methodical and responsible way.

“For us, it is important that predictive capabilities are accurate, explainable, and aligned with sound business judgment,” Gregoriou said.

Turning to robotic process automation (RPA), Gregoriou said Island Oil began with back-office financial reconciliations because RPA is one of the most practical and accessible technologies currently available, offering meaningful benefits to organisations of all sizes.

He explained that RPA is particularly effective in processes that are mature, well-structured and rule-based. Where workflows are clearly defined and repetitive, he said, they no longer require constant human intervention. Instead, software robots can execute them consistently and accurately, allowing employees to focus on exception handling and higher-value tasks.

Financial reconciliation processes, he said, met exactly those criteria. They are critical to the organisation, highly structured, but also time-consuming and repetitive, making them ideal for the first phase of RPA implementation.

He added that starting with reconciliations served two purposes: it delivered measurable efficiency gains and helped build internal trust by demonstrating reliability.

According to Gregoriou, the success of that first phase showed the wider potential of automation, and colleagues in other departments have since started proposing additional, and more complex, processes for RPA implementation.

He described that organic demand as a sign that automation is no longer viewed as an experiment, but as a valuable operational tool.

Asked how employees responded to the introduction of AI tools, Gregoriou said there was some hesitation at first. AI, he said, is often portrayed as disruptive or threatening, and some employees were concerned about what it might mean for their roles.

That concern, he added, was especially evident with RPA. When people saw that time-consuming tasks could be automated by a single software robot, the immediate question was whether automation would replace human involvement.

This, Gregoriou said, is why clear communication from management was so important. From day one, the company emphasised that AI and RPA were being introduced to augment capabilities, not eliminate roles.

The approach, he explained, has been to design processes in which technology and people coexist, combining the speed and consistency of automation with human judgment, experience and critical thinking.

As employees began to experience the benefits firsthand, including reduced repetitive workloads, faster access to information and fewer manual errors, confidence increased. In place of resistance, he said, the company began to see curiosity, with teams exploring new use cases and proposing ideas of their own.

For Gregoriou, digital transformation succeeds only when people trust it, and trust built through transparency and inclusion has been one of the most important factors in the group’s AI adoption journey.

Asked whether AI is reshaping roles across commercial, technical and IT departments, Gregoriou was clear.

“Yes, it does, and to a certain extent, that evolution is inevitable.”

“As organisations increase their adoption of digitalisation, advanced reporting, automation and AI-driven tools, roles naturally begin to evolve,” he said.

In commercial departments, he said, decision-making becomes more data-driven. In technical departments, access to real-time insights enhances preventive maintenance, performance monitoring and operational planning. Human expertise remains central, he added, but is increasingly supported by deeper analytical capability.

However, Gregoriou said the biggest shift is taking place within the IT department, where the role of the IT function itself is undergoing a fundamental transformation.

“Traditionally, in many shipping companies, IT departments were viewed primarily as execution units, responsible for maintaining systems and providing technical support,” Gregoriou said.

Even where companies invested in sophisticated IT systems, he added, IT personnel were not always positioned strategically.

“That model is no longer sustainable for organisations that aim to remain competitive and resilient,” he said.

Adopting emerging technologies, he said, requires greater expertise in systems architecture, cybersecurity, data governance, artificial intelligence and continuous innovation management. As a result, organisations increasingly need larger, more specialised and more strategically positioned IT departments capable of supporting complex digital ecosystems.

“IT must operate as a strategic partner, not just a service provider,” Gregoriou said.

He added that the same shift is visible more broadly across the industry, with technology leaders increasingly participating in executive teams and, in many cases, being represented at board level. Digital capability, he said, is no longer a back-office function but a core pillar of corporate strategy.

Finally, asked to define Island Oil’s digital philosophy in one sentence, Gregoriou said it is to integrate technology seamlessly into daily operations in a way that augments people, strengthens governance and transforms data into timely, responsible decision-making.