By: 19 January 2023
The art of the possible: how AI can transform the claims valuation process

Mark Hemsted is a partner in the volume motor practice at Clyde & Co

Using AI will bring about great benefits for personal injury insurers as well as the ultimate policyholder

The phrase ‘artificial intelligence’ might conjure up images of a futuristic sci-fi film, but the reality is that AI already is playing a role in making the claims valuation process more efficient for both insurers and their policyholders.

As well as helping to reduce manual processes and overheads, the power of AI to harvest vast amounts of data and use it to identify trends and claims characteristics looks set to transform the way claims are evaluated.

Clyde & Co believes that using AI will bring about great benefits for our personal injury insurer clients as well as the ultimate policyholder.

Used intelligently, AI can reduce insurer expense, reduce fraudulent claims and speed up the claims valuation process.

For some time now, law firms and others involved in claims have been exploring how AI could be used to evaluate claims.

A template approach, whereby an AI platform reads documents but requires the information on those documents to be in the same format and position, can be used to good effect for many areas of business.

For personal injury claims, however, this approach, we believed, was too limiting.

A lot of the work in valuing personal injury claims involves reading and evaluating medical evidence.

And while many doctors and medical professionals use a standardised approach to document simple cases, for more complex reports—such as psychological evaluations—it’s often the case that doctors dictate notes, leading to a non-templated report being produced.

We realised that for AI to be effective in evaluating these complex claims, it needed to be a platform that could actually read the documents.

Such an AI platform needs to use language processing techniques, rather than the templated approach.

We selected a partner that uses optical character recognition (OCR) to process the language in documents and effectively read them as a human would.

But, much like a human, to be able to carry out this task the platform has to be taught to read.

The platform harvests more than 200 pieces of information from each medical report.

We have set up a training program that teaches the platform how to read this information and enables it to learn from both correct and—crucially—incorrect information.

We now use the system to read all medical reports that have a value of £25,000 or below.

Benefits for insurers

This type of cradle-to-grave automation of lower value claims offers huge benefits to insurer clients.

The system enables claims professionals to focus on the more complex, typically higher-value claims where their expertise is most usefully employed.

It means they do not have to spend time processing lower-value claims manually, but instead are freed up to use their skills where they can be of most benefit to the insurer—and policyholder.

Eliminating manual work previously involved in many lower-value, non-complex claims, results in greater efficiency for the insurer and reduces costs.

AI harvests more data than has ever been harvested before and uses that data to make the best possible valuation of the claim. This enables a bespoke offer to be made to both the claimant and their instructing solicitor reflecting the true value of the claim.

Mass harvesting of data allows trends to be identified much more easily and earlier than ever before.

This can result in huge cost savings as well as ensuring that claims are dealt with more expeditiously where needed.

For example, the data can be used to identify the characteristics of a personal injury claimant that can cause a claim to escalate. This enables the insurer to make a decision about whether to get involved earlier in the life of a claim.

AI analysis of this data also can enable fraud to be detected more swiftly because it can identify even the slightest discrepancies between documents.

As well as the obvious cost and efficiency benefits for insurers, we believe that this automatic harvesting of data is truly powerful for insurers and a real step forward in the way claims are evaluated and managed. All the data is, of course, destroyed once a claim has been resolved.

Benefits for policyholders

There are real benefits for the policyholder, too.

The reduction in insurer overheads facilitated by the use of AI means that insurers’ costs are reduced, often resulting in a consequent reduction in premium rates for coverage.

The ability to reduce fraud also has an effect on cost for insurers and, therefore, policyholders, and also means that genuine and valid claims can be prioritised and dealt with more quickly.

The use of data to identify trends and claim characteristics and, consequently, to empower insurers to intervene more swiftly in those claims that might escalate also benefits the policyholder, by ensuring they are compensated quickly and justly.

What next?

AI will never replace humans in the claims process. But we believe it can enhance their role and enable their skills to be put to best use.

It’s important to remember that this system values claims, it does not determine fault—that is a job for human judgment.

At a time when the insurance industry finds itself facing a talent gap, the ability of AI to free-up claims professionals to focus on the more complex claims—the more technical and skilled work—is a huge benefit for insurers.

Looking to the future, we have some clients who are showing an interest in increasing the scope of the claims that are valued by the system.

There’s potential for other areas of claims to be valued in this way, too. It’s possible to imagine credit hire claims or vehicle damage claims being evaluated by AI—we know that we can teach the system to read this type of documentation.

And we have clients that are interested in the benefits AI could bring to helping evaluate larger, more complex personal injury claims.

Developing this system has not happened overnight. And it’s an ongoing process—we have a team of five people whose job it is to teach the system.

This a big investment, of time as well as money, but one which we believe will enable our insurer clients to be fit for the future.

As the industry looks to the challenges ahead of it and the need to be nimble, efficient and to use human talent in the most effective ways possible, there’s no doubt that optical character recognition AI will play a vital role in the claims process.

Image: © PhonlamaiPhoto’s Images via Canva and Clyde & Co