The global maritime industry is navigating a sea of risks, from operational bottlenecks and data blind spots to rising safety concerns. With so many players — from shipowners and charterers to ports, regulators, and insurers — the industry needs smarter tools to stay afloat and ahead.
Among the companies working to address this need is RightShip, a maritime due diligence provider focused on reducing risk across global shipping operations. The company combines data, AI, and experienced human insight to help decision-makers act with greater confidence.
Marlon Grech, Chief Product Officer and Chief Technology Officer of RightShip, discussed the company’s tech stack and data strategy, which underpin its approach to smarter, more resilient maritime operations.
Building blocks
Every day, RightShip ingests data from multiple sources across the maritime ecosystem, resulting in billions of data points. With operations dating back to the early 2000s, RightShip has built a data foundation that enables it to serve various customer needs while evolving its platform’s intelligence and value.
RightShip’s tech stack is fully cloud-native and runs on Microsoft Azure. The company uses a microservice architecture that allows it to scale based on demand, helping maintain reliability and performance across its global user base.
“AI plays a critical role in our ability to unlock unstructured data, much of which still exists in formats like PDFs that were designed for human eyes, not machines. We’ve invested in AI to ingest, interpret, and structure this information, turning previously inaccessible content into usable benchmarks, trends, and insights,” Grech said.
At the heart of RightShip’s platform is a workflow engine that can be configured to reflect each customer’s specific risk profile and operational requirements. Built on top of this are AI agents designed to automate manual tasks such as reviewing class status reports for expired certificates.
“We continue to adapt the platform to reflect the specific demands of maritime risk management. A recent example is an upgrade to our PSC RiskIQ feature, which now includes AI-driven checklist suggestions and supports workflow customisation. This helps customers make quicker, better-informed decisions in complex and fast-changing environments,” he remarked.
Data integration
Data silos are common across industries, and removing these blind spots is often key to uncovering new operational and business insights.
According to Grech, RightShip’s data platform is built on Azure and uses a data lake architecture powered by Databricks to ingest and process large volumes of structured and unstructured maritime data.
To manage this complexity, the company has adopted a data mesh model, giving individual teams responsibility for developing and maintaining domain-specific data products.
“Our lakehouse architecture lets us land all available data in raw form, then progressively refine it through bronze, silver, and gold layers. This structured approach ensures data is cleaned, enriched, and analytics-ready for various business needs,” he explained.
Grech also noted that data governance plays a central role in their approach, with each data product clearly defined, versioned, and subject to ownership and quality controls.
“We use Databricks Unity Catalog to maintain full lineage across the platform, allowing us to trace any data point back to its original source. This level of transparency is important not just for privacy and compliance, but also for building trust and interpretability in the insights we provide,” he said.
Timeliness
In maritime operations, decisions often need to be made quickly. But acting without the necessary context or information introduces significant risk.
To support timely decision-making, RightShip applies a risk-based prioritisation model. The platform highlights the most urgent and relevant insights based on available inputs and continuously updates its recommendations as new data comes in.
“Maritime data is rarely clean, complete, or real-time — and that’s exactly what we’ve designed our platform to handle. We’ve built a system that’s resilient to gaps, delays, and inconsistencies. We don’t wait for perfect data; we enable decisions with the best available information, while clearly flagging any missing or low-confidence inputs so that human experts can step in where needed,” Grech said.
According to Grech, the platform processes high-frequency signals like AIS movements or inspection outcomes on a continuous basis, while slower, manually submitted information is ingested asynchronously and incorporated as it becomes available. This event-driven, modular architecture is designed to keep the system responsive without delaying time-sensitive decisions, particularly in areas like due diligence or vessel vetting.
Trustworthiness
Not all data are equally useful, and organisations need to be able to determine which sources are appropriate for a given scenario. Likewise, not all data points carry the same weight, and that differentiation is by design, Grech emphasised.
“We’ve designed our platform to reflect that reality. We have governance processes to assess the reliability of each source, factoring in origin, completeness, and historical consistency — recognising that not all ports, vessel managers, or regulators follow the same standards. For example, a PSC (port state control) finding from one region may be treated differently than the same one elsewhere, and that’s based on experience, not assumption,” he said.
Grech added that subject matter experts are closely involved in shaping how data is interpreted and weighted, working alongside data scientists trained to minimise bias and maintain analytical integrity.
“We know that if you torture the data long enough, it will tell you what you want — but that’s not how we operate. We’ve built in controls and review steps to ensure data is used fairly, transparently, and in context,” Grech said.
The goal, Grech explained, is to extract useful insights from imperfect information — applying AI to scale and speed up analysis, while keeping human expertise in the loop where it matters most.
“It’s about enabling safer, smarter decisions, today — not someday,” Grech noted.
ML capabilities
Machine learning plays an increasingly important role in RightShip’s risk models, but the company takes a deliberate approach to how these models are used, Grech said.
“Our models are embedded throughout the platform, from classifying incident severity to assessing port state control deficiencies, and they help us interpret complex data at scale. But we don’t treat machine learning as a black box. Every model we deploy is governed, explainable, and intended to support — not replace — human expertise,” he said.
Current applications include classifying the severity of inspection findings, identifying patterns in historical incidents, and extracting recurring themes from unstructured data.
While much of the industry’s data tends to focus on problems or failures, Grech noted that predictive models can unintentionally overemphasise negative outcomes. RightShip is working to counterbalance that.
“We want our models to not only flag risk, but to recognise safety. That means identifying and highlighting operators and vessels that are investing in better practices, preventive actions, and ongoing improvement. It’s a shift from purely punitive to more constructive — creating incentives for the industry to raise standards, not just avoid penalties,” he said.
Interoperability approach
Not all platforms are built for interoperability, and many organisations struggle to connect systems in a way that supports meaningful data exchange. In a sector where trust, data integrity, and model transparency are critical, RightShip takes a “controlled openness” approach to integration.
“We actively encourage our customers to use more of our data within their own environments. Whether it’s a fleet management platform or an internal analytics dashboard, we design our APIs to make integration straightforward, secure, and reliable. We’ve also invested in partnerships with third-party platforms to bring RightShip data closer to where users already work, because real impact happens when insights can move across the ecosystem,” Grech said.
He added that while consistent, well-governed APIs are essential for innovation, this hasn’t always been a strength of the maritime tech sector. RightShip is continuing to explore ways to address that gap.
“We’ve built our platform with modularity, clear interfaces, and API-first design — and we’re always looking for ways to expand that approach in ways that maintain trust while delivering more value to the industry,” he said.
