This piece is by Piotr Piękoś, Head of Insurance Practice at Future Processing
The insurance industry is no stranger to technological change. From the rise of computers in the 1980s, the proliferation of the internet in the 1990s and the advent of big data since the turn of the century, insurers have navigated a host of profound changes which many predicated would render claims professionals obsolete. Now, the next revolution is here.
As agentic AI becomes increasingly prevalent within the London market, claims functions are on the frontline. With the ability to optimise workflows and deliver better outcomes for both insurers and policyholders, claims teams across the market are rushing to stay ahead of the curve. At the same time, however, many are encountering operational barriers that are slowing progress.
Shifting the dial
The broader conversation around agentic AI can often feel slightly distant, with a lot of talk about autonomy and not always much clarity about how that applies to day-to-day operations. However, we have seen first hand how the technology is delivering tangible value for claims teams. Its inherent versatility has generated efficiency gains across the full claims lifecycle, from segmentation and triage to evaluation and resolution.
The technology is particularly adept at handling high-volume, low-value claims, which typically carry the greatest administrative burden. At a time when claims are becoming increasingly complex, often cutting across multiple business lines, claims triage is a key area of focus. Agentic AI can sort claims intelligently as soon as they arise, directing complex claims to the most experienced handlers, reducing friction for claims teams and speeding up the process for policyholders.
However, in an industry powered by human relationships, claims teams can make or break the reputation of any insurer. Agentic AI, therefore, cannot be adopted in isolation and insurers must take a holistic view of the full claims lifecycle, identifying gaps where the technology can enhance their service and build trust. In specialty lines, contact between claims teams and brokers is critical, and communication orchestration through agentic AI is an emerging area of significant potential.
Breaking down barriers
While agentic AI presents immense opportunities, operational and cultural barriers are slowing adoption. Many claims professionals worry about maintaining the quality of their service, about having tools imposed on them by people who do not understand their work and, ultimately, about being replaced. However, if agentic AI agents are designed well, genuinely help to deliver better outcomes and are introduced gradually, this resistance tends to soften.
Operational barriers, however, present a far greater challenge. Historically, the insurance sector has not adequately addressed the issue of data harmonisation, with many insurers storing data across range of different systems, in different formats with varying levels of accessibility. Without a uniform data layer that agentic AI can learn from, integration becomes an overly complex, drawn-out process whereby dispersed data sets need to be collated and standardised.
As agentic AI develops at pace, enhancing its capabilities, becoming more autonomous and expanding its range of use cases, trust, control and accountability are now critical. While there are guidelines that provide a degree of structure, such as FCA guidelines and Lloyd’s of London requirements, as well as the NIS2 Directive and the AI Act at a European level, regulators are struggling to keep pace.
What lies ahead
Over the next five years, we will see a significant rise in agentic AI agents being deployed across claims functions. As a result, claims professionals will take a step back from day-to-day admin and execution tasks and instead focus on decision-making, challenging and endorsing the recommendations of agents. Technical proficiency will not be a requirement for claims handlers, as agents will be managed by IT and technology functions.
While AI agents will reshape how claims professionals work, they will not eliminate the need for human involvement. After all, the claims process is much more than a tick-box exercise. In their moment of need, policyholders demand an empathetic service from humans offering expert counsel and reassurance. Freeing up claims handlers to focus on this part of the process is where the real value of agentic AI lies.
Despite the rise of agentic AI, there will always be a role for human claims professionals. The challenge for insurers is identifying which parts of the claims process should be automated, which parts require human judgement and what the foundations of a scalable model look like. Insurers cannot afford to ignore agentic AI, and their challenge is to deploy it with boundaries that are explicit and enforceable.
Be the first to comment