In Insurance, the Policy Review is one of the most valuable areas for adopting practical AI solutions. This critical activity is owned by the Underwriting function and is an important aspect of very specific steps in the typical insurance workflow. In detail:

1. When the carrier receives a new submission from the broker. In this scenario, underwriters must ensure that coverage and exclusions are in line with the written documents (i.e binder or instruction vs policy)
2. At renewal time, when brokers send a summary of the renewal and the underwriter must verify this against what was expected.
3. For international programs, to ensure that, in the process of localizing a policy or a contract, the changes required do not introduce critical, risk profile impacting changes to the terms of the global policies.
4. In response to unexpected and rare events, as many carriers have experienced with COVID. In this scenario, underwriters might be required to analyze all existing policies to evaluate the potential exposures to claims related to these rare events.

While we focused mainly on carriers, brokers, for example when they may receive a binder from the carrier and need to make sure that binder and policy are aligned, face similar requirements.

Policy review is not straightforward. The review process typically involves multiple team members, it happens more than once and could take anywhere from 2 to 6 hours or longer for certain lines of business-like Reinsurance. This makes Policy Review probably the first and most valuable target area for AI-driven automation.

When we think about the value of automation in policy review, there are different aspects that contribute, often in a compounded way, to create value.

The first is, obviously, to mitigate the carrier’s exposure to risk if an event not intended to be covered in the policy is either included in the coverage or worded in a way that doesn’t comply with the carrier’s risk profile. Automating the policy review  can ensure that 100% of policies are reviewed before being executed, even at times when resources are scarce, i.e. during renewal peaks, and it  can ensure that the policy is reviewed based on standard and consistent criteria. While this may seem obvious, in the traditional, pre-AI world, policies are often reviewed by multiple people at different times, and those reviews might therefore vary.

The second way that automation creates value is by increasing capacity, possibly in an unlimited way. By nature, the Underwriting workload is seasonal. There are peaks, at renewal time for example, that limit the resources that can be allocated in managing new submissions. Because the speed in which a carrier responds to new submission is an important factor in winning the new contract, delays or a lack of response, can result in losing business that could be avoided through automation.

The third driver of value is the one that is most often associated with automation: Efficiency. In underwriting, “efficiency” means significant cost optimization in terms of less time dedicated to the review of each policy without compromising the quality and accuracy of the review.

While the value of automation in underwriting is undisputable, the adoption of AI in this area is still in the early phase. Carriers are showing an appetite for the AI technologies that drive automation because they understand that any improvement in activities at the beginning of the Insurance value chain can generate a positive impact across the organization.

However, to enable automation (and fully maximize the value of AI-driven solutions), carriers have to think about process design from the very beginning. In Insurance, and underwriting specifically, the amount of manual work required even today is astounding. While this means that the value automation can add is significant, many processes are designed around manual tasks, which prevents automation from coming in.

Going back to the example of automating policy review, while AI solutions that ensure the deep understanding of the content are not 100% accurate, neither are humans. That’s why many processes include human review, the so-called “human in the loop”, at the end for reviewing the outcome of AI-driven comparison. This is a persuasive use case for those in the underwriting community—a position that has traditionally been at the center of value creation in insurance—who are still skeptical of the value of AI.

However, if insurers cannot overcome this resistance to innovation, the alternative is being left behind and unable to compete. For those who are still unsure, there are lots of use cases that not only demonstrate the value, but are also not complex implementations. Underwriting is the perfect area to take innovation to the next step, where the full potential of technologies like AI-based automation can be realized.

Luca Scagliarini
EVP Strategy & Business Development

Pamela Negosanti
Global VP of Insurance

L'articolo Let A.I. review the policy review proviene da Expert System.