Artificial intelligence has become one of the most discussed technologies in marine insurance. Conversations often focus on underwriting models, risk prediction, cyber threats and sanctions monitoring. Yet according to Ty Zamkow, co-founder and CEO of Nolana, the most immediate opportunity may lie somewhere far less glamorous: claims handling.

Nolana focuses on agentic AI for insurance operations, helping insurers process information faster and make more consistent decisions without removing humans from the loop. Its modules cover first notice of loss intake, claims triage, data enrichment, document processing, coverage matching, and customisable workflows designed to fit how each insurer already operates.

"We're an agentic AI platform built for the specific requirements and nuances of the insurance industry," says Zamkow. "We help insurers shorten cycle time and drive efficiencies in claims processing."

Zamkow says the real transformation is not about replacing underwriters or claims handlers with machines. Instead, it is about improving how insurers drive efficiencies that allow them to improve their loss ratio and grow their companies, in a way that traditional insurance systems and processes – some decades old -- were not designed to handle.

While much of the industry's attention has centred on how AI might transform underwriting, Zamkow argues that the greatest value today comes from helping insurers process claims faster, more consistently and with greater accuracy.

"There are real opportunities for AI adoption in underwriting,” he says. “But we find that some of the biggest gains are actually in claims operations."

For an industry facing growing volumes of data, increasing regulatory scrutiny and pressure to improve efficiency, that distinction matters.

Marine insurers have spent decades building systems and workflows designed for a slower-moving world. Today, claims teams are expected to manage complex international incidents involving multiple stakeholders, extensive documentation and growing expectations around response times.

Artificial intelligence, Zamkow believes, offers an opportunity not to replace claims professionals but to remove many of the inefficiencies that consume their time.

Shipping Risks Are Becoming More Complex

Marine insurance is operating in an increasingly challenging environment.

Geopolitical tensions, sanctions compliance, cyber threats, climate volatility and changing trade patterns are all creating new forms of exposure for shipowners and insurers alike.

At the same time, vessels generate more operational data than ever before.

AIS information, satellite imagery, cargo data, engine performance metrics and voyage information provide insurers with unprecedented visibility into maritime operations.

Historically, much of this information was either unavailable or difficult to access.

Today, the challenge is no longer in having available data, but rather  in correctly ingesting and processing it.

"AI allows us process it at scale very quickly without the need to spend weeks chasing it," says Zamkow.

Yet despite the excitement surrounding artificial intelligence, he believes adoption remains at an early stage.

"People talk about AI, but it's not widely used in the actual day-to-day work."

The gap between potential and practical implementation remains significant.

Insurance workflows were designed around manual processes, human review and relatively static information sources. Integrating AI effectively requires more than new software. It requires changes to operational processes, governance and organisational culture.

For many insurers, claims handling is proving to be the logical place to start.

The Claims Bottleneck

Claims management remains one of the most labour-intensive functions within insurance.

A typical marine claim can involve emails, survey reports, cargo manifests, bills of lading, weather data, photographs, policy documents, adjuster reports and correspondence between multiple parties.

Each document must be reviewed, categorised and assessed before a claim can progress.

The result is a process that consumes enormous amounts of time.

According to Zamkow, claims handlers often spend up to 40% of their working day performing administrative tasks rather than applying professional judgement.

That creates inefficiencies throughout the claims lifecycle.

Highly qualified claims professionals frequently find themselves manually reviewing documents, chasing missing information and re-entering data across multiple systems.

These activities add little value but consume substantial resources.

"The real challenge isn't a lack of expertise," says Zamkow. "It's that skilled people spend too much time doing work that technology can now assist with."

This is where AI is beginning to make a measurable difference.

Reinventing First Notice of Loss

One of the most important stages of any claim is the First Notice of Loss (FNOL).

It is also one of the most inconsistent.

Claims notifications arrive through multiple channels and in varying formats. Some contain detailed information, while others provide only a basic description of the incident.

Traditionally, insurers have relied on staff to review incoming notifications, extract relevant information and manually enter it into claims systems.

The process can involve significant duplication.

"The FNOL process today is essentially a data collection problem dressed up as a workflow," Zamkow explains.

An email arrives. A claims professional reviews the message, checks policy details, extracts key information and enters it into one or more systems.

Additional documents are requested. Missing information is chased.

Only then can the claim formally progress.

AI can automate much of this process.

Modern systems can extract information from emails and attachments, structure the data, identify missing information and automatically initiate requests for additional documentation.

Coverage can be matched against policy terms, and workflows can be triggered without requiring manual intervention at every stage.

The objective is not to remove people from the process.

It is to ensure that claims handlers focus their attention where professional expertise genuinely adds value.

The Hidden Cost of Dormant Claims

One of the most revealing observations from Zamkow concerns what he describes as the industry's "static claims problem."

Across many insurance portfolios, a substantial proportion of open claims are effectively dormant.

"They're sitting open, not progressing, costing money to manage, and nobody is actively working them," he says.

In some markets, Zamkow estimates that around half of all open claims portfolios may consist of dormant or static files.

The implications are significant.

Dormant claims require ongoing administration. They consume management time, occupy system resources and contribute to reserve requirements.

Perhaps most importantly, they tie up capital.

Claims that remain unresolved often continue to sit within reserve calculations, limiting financial flexibility and distorting portfolio visibility.

The causes vary.

Missing documentation, delayed responses, unresolved investigations and fragmented communication can all contribute to claims stagnation.

The problem is not necessarily that claims teams are neglecting these files.

Rather, claims handlers frequently lack the time required to review large volumes of open cases systematically.

AI offers a practical solution.

By reviewing claim files continuously, systems can identify why claims have stalled, determine what actions are required and flag opportunities to move cases forward.

"It sounds simple," says Zamkow, "but it's one of the areas where AI can deliver immediate and measurable value."

Recovering Millions Through Better Subrogation

Perhaps nowhere is that value more apparent than subrogation.

Subrogation allows insurers to recover losses from third parties responsible for a claim.

In theory, it represents an important source of financial recovery.

In practice, many opportunities are never pursued.

Once a claim is settled, attention naturally shifts toward new cases. Reviewing historical files to identify recovery opportunities can be time-consuming and resource-intensive.

As a result, potential recoveries often remain undiscovered.

"Subrogation falls off the radar because somebody has to go back through the claim, assess whether there's a recovery opportunity and then initiate a separate process," Zamkow explains.

This makes subrogation particularly well suited to AI-supported analysis.

Claims files already contain much of the required information.

The challenge is identifying relevant indicators consistently.

AI systems can review settled claims, compare circumstances against policy terms and identify cases where third-party recovery may be possible.

Rather than replacing claims professionals, the technology acts as an additional layer of analysis.

Potential opportunities are flagged for human review, ensuring that valid recoveries are not overlooked.

For marine cargo claims in particular, the opportunities can be substantial.

"There is often a clear third-party liability element," Zamkow notes.

The information already exists within the file. The problem is ensuring that it receives sufficient attention.

For insurers seeking to improve profitability, better recovery management may represent one of the most attractive applications of artificial intelligence.

Managing the Document Explosion

Marine claims generate enormous quantities of paperwork.

A single cargo claim can involve hundreds or even thousands of documents.

Survey reports, photographs, manifests, correspondence, invoices and supporting evidence accumulate rapidly.

Managing that information has traditionally required significant manual effort.

Claims professionals must review documents, identify relevant information and determine how it affects the claim.

The process is time-consuming and vulnerable to inconsistency.

"The document chasing and review problem is genuinely significant," says Zamkow.

Artificial intelligence is proving particularly effective in this area.

Modern systems can categorise documents automatically, extract relevant information and highlight inconsistencies or missing data.

Instead of spending hours searching through files, claims handlers can focus directly on the most relevant information.

The result is not simply faster processing.

It is greater consistency across claims portfolios and improved visibility into complex cases.

Improving Adjuster Coordination

Loss adjusters play a critical role in marine claims.

However, coordinating their involvement often introduces additional administrative burdens.

Appointments must be arranged, reports tracked and findings incorporated into claim files.

Although these tasks are relatively straightforward, they consume valuable time.

AI can streamline much of this activity.

Routine communications can be automated, progress monitored and documentation integrated directly into claims systems.

This allows claims teams to spend less time coordinating logistics and more time assessing outcomes.

When multiplied across hundreds or thousands of claims, even modest efficiency gains can have a significant impact.

Why Data Quality Matters

Despite the enthusiasm surrounding AI, Zamkow repeatedly returns to a fundamental issue: data quality.

Artificial intelligence is only as reliable as the information it receives.

This challenge is particularly visible in areas such as sanctions compliance and shadow fleet monitoring.

AI may be capable of analysing vessel movements, ownership records and operational patterns at extraordinary speed.

However, if the underlying information is inaccurate or manipulated, the conclusions may also be flawed.

"The quality of data that you feed the model is critical," he says.

This reality has important implications for marine insurers.

As organisations become increasingly dependent on data-driven decision-making, the provenance and reliability of information become strategic concerns.

Satellite imagery, independently verified records and cross-referenced datasets may become increasingly important as insurers seek greater confidence in AI-supported analysis.

Technology can accelerate decision-making, but it cannot eliminate the need for verification.

Cyber Risk and the Connected Ship

The growing connectivity of modern vessels presents another challenge.

Bridge systems, machinery, communications platforms and operational technologies are increasingly linked through digital networks.

This connectivity improves efficiency and visibility.

It also expands cyber exposure.

"In the past, not a lot of it was connected," Zamkow says. "Today, much more is connected."

For insurers, cyber risk can no longer be viewed solely as an IT issue.

Operational technology, onboard systems and third-party integrations all influence exposure.

While AI can assist in identifying anomalies and vulnerabilities, meaningful risk assessment still depends on access to reliable information.

The challenge once again returns to data quality and visibility.

Human Expertise Still Matters

Despite AI's capabilities, Zamkow rejects the notion that claims handling or underwriting will become fully automated.

Regulators increasingly demand explainability and accountability in decision-making.

More importantly, insurance remains a business built on professional judgement.

"You cannot write a policy based on an output from an AI model that people cannot explain," he says.

The same principle applies to claims.

AI can review documents, identify patterns and recommend actions.

But decisions involving coverage, liability and settlement still require experienced professionals.

For Zamkow, the future is not about replacing claims handlers.

It is about making them more effective.

"AI can help sort through a lot of data very quickly and provide recommendations," he says. "But at the end of the day, you still need highly qualified professionals."

The role evolves rather than disappears.

Claims professionals spend less time on administration and more time applying expertise.

A Pragmatic Future

The marine insurance sector has heard countless predictions about technological disruption.

Many have failed to materialise.

Zamkow's outlook is notably pragmatic.

He does not foresee a future dominated by fully autonomous underwriting or claims departments.

Instead, he expects gradual adoption focused on areas where value is easiest to demonstrate.

Claims operations sit at the centre of that transformation.

First Notice of Loss, document processing, dormant claims management, subrogation recovery and adjuster coordination all represent areas where AI is already delivering tangible benefits.

For insurers under pressure to improve efficiency, reduce costs and accelerate claims resolution, these applications may prove far more significant than headline-grabbing discussions about artificial intelligence replacing humans.

The technology itself is increasingly capable.

The real challenge lies elsewhere.

Success depends on trusted data, effective governance, collaboration across the insurance ecosystem and a clear understanding of where human expertise remains essential.

In marine insurance, AI may ultimately transform many functions.

But according to Zamkow, its most immediate impact is likely to be far simpler.

Helping insurers close claims faster, recover more money and allow skilled professionals to focus on the work that matters most.