While 95% of CEOs believe that there would be immediate business benefits from implementing AI, 60% are hesitant because of uncertainty around return on investment, according to research from Munich Re.

In the race to modernize insurance operations, AI and automation are often positioned as the ultimate destination — a seamless, touchless claims journey powered by advanced technologies. But for Patricia Walsh, Claims Transformation Lead at Munich Re Specialty – North America, the reality is far more complex than it may appear. The future of claims is not about eliminating human involvement, but about redesigning the process so that expertise and automation work in tandem.

“When I approach automation, I look at the claims process from start to finish. Imagining a fully automated journey helps highlight where automation makes sense — and where human interaction and expertise remains essential to the claims experience.”

This framing, starting with a fully automated vision, acts as a strategic exercise rather than a literal goal. It allows insurers to map out possibilities before making deliberate decisions about where human involvement remains essential. As Walsh told IB, the claims journey isn’t just an operational process — at its core there is a deeply human aspect. “The customer needs to be involved in the process, and having full automation may not involve them in the way that you or they want,” added Walsh. “With customers, you really want to offer them that white-glove service — that's important. Another key reason the human element cannot be automated is that some business decisions require certain expertise — something that we at Munich Re Specialty pride ourselves on. “Our claims professionals bring years of experience aligned to the lines of business they support, and their expertise allows them to guide any decisions required throughout the claim's lifecycle.”

Ensuring that human touch is always on hand is something that really sets Munich Re Specialty apart here. While the team is always on the lookout for the next exciting piece of technology, it’s never to the detriment of people; rather, Munich Re Specialty leverages AI and automation to aid their experts, not replace them.

This emphasis on expertise highlights a critical distinction: Automation can support decision-making, but it should not replace it. Instead, Walsh advocates for a more nuanced transformation approach — one that begins with the ideal and then intentionally reintroduces human judgment where it matters most.

In practice, this means automation should be applied selectively — targeting high-volume, repeatable tasks while preserving human oversight in areas that require judgment, complexity, or relationship management. Certain lines of business may lend themselves more readily to full automation, but specialty insurance demands a more tailored approach.

“Certain kinds of claims outside of our industry, such as warranty claims, perhaps you could automate that end-to-end,” added Walsh. “That would make sense. But for specialty insurance, we wouldn't look to automate the whole claims process due to the complexity and focus on the human part of the business.”

In looking at opportunities within the Specialty Claims sector, one example is that in broker-driven environments, automation can significantly accelerate First Notice of Loss (FNOL) handling. “Brokers often send FNOLs to multiple carriers simultaneously,” said Walsh. “When standardized forms are used, AI and other ingestion tools can help move that information into the claims process faster.”

“Where elements of the process fall outside direct control, automation can support timely customer engagement through reminders. Third‑party integrations, such as APIs, can further accelerate information exchange compared to manual or paper‑driven processes.”

While automation is transforming workflows, AI is often seen as a cutting-edge solution that will solve most business problems. Here, Walsh cautions against viewing AI as a universal solution and instead as a strategic tool.

“AI is another tool in the toolbox, but it has limits. When you think about a carpenter, a hammer isn't the best way to cut a piece of wood in two. You could smash it and make the wood into two pieces, but a saw might be a better tool. The same goes for AI — we need to approach problems strategically and look at both the cost and the benefit involved based on the solution chosen.”

Where AI really excels is in augmenting human capability, particularly in areas such as communication, data analysis, and triage while also offering significant advantages in identifying patterns and prioritizing work.

“AI can help summarize data and information,” added Walsh. “When we’re reporting, we think about keywords, and AI can help us determine potential indicators that a claim might be flagged for across the organizations such as loss control opportunities, root causes, or complexity.”

One of the most common pitfalls in transformation is adopting new technologies without a clear understanding of underlying business needs. Here, Walsh was keen to point out that successful transformation begins with problem definition not technology selection.

“You don't want process to get in the way of progress,” she told IB. “There definitely needs to be a balance. I think the best way to approach transformation is to look at your business problems, document them, make a list, and see which ones are the most frequent or the most impactful to your business. Solve for those first for the largest impact.”

When it comes to measuring the success of AI and automation in an organization, it becomes more complex. If data and metrics aren’t available for measurement, it can cause a business to pause on solutioning a problem because they cannot measure the business benefit. Here, Walsh suggests that instead of relying purely on abstract metrics, you should look at real examples of transformation outcomes.

According to Walsh, measuring the impact of automation requires looking beyond traditional metrics. “Not everything is easily measurable,” she said. “By using focused case studies —   examining a small group of claims and the time savings achieved — you can validate outcomes and apply those insights across the wider claims portfolio.”

This hands-on evaluation reveals tangible improvements in a way that large-scale metrics often cannot.

“You look at the improvements and can see how long the claims ingestion took, for example — and that gives you a good indication of success, without having to rely solely on data or KPIs.”

For organizations that do require more formal measurement, Walsh suggested additional layers of visibility using the data to draw measures and metrics.

“There are also business process tools that you can put on top of your systems to measure where things are going well and where there are areas for improvement or churn,” she told IB.

Ultimately, however, even the most sophisticated measurement frameworks cannot compensate for the wrong people or culture. Because transformation, Walsh argues, is fundamentally human.

“The people involved in transformation efforts matter,” she told IB. “You need individuals who are curious, who approach problems with a fresh perspective, and who bring strong business understanding. Equally important is removing barriers and being genuinely open to ideas. At Munich Re Specialty, we’re fortunate to have that depth of curiosity and claims expertise, which truly sets our approach apart.”