Making insurance claims as fast, accurate, and fair as possible

In the quest to make the claims process as fast, accurate, and fair as possible, insurers are adopting several new technologies and approaches—many of which fall under the broad umbrella of “claims automation”. While automating the many manual tasks and sub-processes that take a claim from first notice of loss (FNOL) to settlement is a worthy goal, doing so with intelligence is key.

While it would be easy to automate claims by simply paying all claims, this would be extremely expensive. Refusing all claims would be equally effective, but equally problematic. With both approaches, the insurer would go out of business. Intelligence is therefore about striking a balance between the costs of handling claims, the claims loss, and the fairness with which you are treating your customers in order to have an effective and fair claims handling system.

Data is key to achieving this balance, leaving two challenges for insurers: how to leverage the data you already have, as it is often in silos, and how augment internal data with external data sources.

These are the thoughts of Marcel Gordon, vice president of product at Shift Technology. He will be speaking on Day 3 of Intelligent Insurer’s Claims Innovation USA Event about how to “Improve Your Operational Efficiency for High-impact Business Outcomes” in a sponsored presentation on “Decision Automation and Optimisation: How Improved Decision-Making Can Transform The Claims Process”.

Gordon spoke to Intelligent Insurer ahead of the event.

What fundamental changes are needed to improve the claims experience?

We know that policyholders need a modern and better customer experience—not just digital, but across any channel or touchpoint. The whole experience needs to be re-imagined to be more seamless and transparent and reduce the number of hoops a claimant needs to jump through.

From the insurers’ side, claims handling and business processes are linked together. You can’t separate subjects such as fraud from the claims process, so insurers are thinking about how their entire claims process works and how various technologies can drive efficiency and reduce costs.

Which new technologies are being deployed by insurers to make the claims process as fast, accurate, and fair as possible?

Artificial intelligence (AI) is having a real impact on claims handling, and it’s having some follow-on effects on other parts of the insurance tech stack. Insurers are understanding they need to upgrade their core systems and leverage the cloud. If you are trapped on a legacy system, the chances of using the cloud and new technology are limited. So, insurers are upgrading their core systems and moving to the cloud.

How prevalent is claims automation, and how can it be used to assess claims fairness and accuracy?

Many insurers are already automating some claims, and virtually all of them are thinking about doing it. The key is doing it intelligently to positively impact fairness and accuracy, not the other way around. The AI can be trained with these goals in mind and this is a big area of focus for us.

To what extent can claims automation improve the speed of claims processing, while perhaps detecting fraudulent claims?

The purpose of claims automation is to improve the speed of claims processing. The long-term objective is to automatically resolve every aspect of every claim. To do that, you have to build a human-level AI, a system that is as capable as a human being. This is a future state; it’s not here today.

Until the entire system can meet this accuracy goal, we need to break down the claims process into specific problems that we can solve accurately with the technology we have. This is the approach that is feasible today. It may not automate all of every claim, but it should be able to speed up the process overall while improving the outcome for the insurer and the policyholder.

Fraud is a great example as we can very accurately assess whether the fraud team needs to look at a claim today and are doing so with insurers globally.

What are the typical manual tasks and sub-processes that can be automated?

The way I’d frame this is that we are looking at any point of the process where people need to make a decision. The best way to save time is to not have people spending time making decisions that the AI can make with great confidence. One example is subrogation in claims—passing on liability to somebody else. Other examples are policy coverage and document analysis. We can figure out whether claims need to be reviewed by a person in order to increase the speed of claims handling.

How does claims automation make every process from FNOL to settlement more efficient, fast, and reliable?

It’s about accelerating decision-making, and the processes that are necessary to support decision-making, but which add time. On some cases we can save minutes or hours, or days and weeks.

We split it into two parts, which are quite different. One is process automation, automatically moving a claim from stage to stage. The second part, that we put a lot of focus on, is the decision-making part—what happens inside a stage.

These two parts are quite different. The decision-making part is where we need the AI to examine the claims to know what to do with them. Should you deal with a claim, or pass it on? How can you do your job better?

What is your advice for insurers wanting to use new technologies?

Think about your systems: core systems have to be very stable, but you don’t want them to block what you want to do. You need application programming interfaces.

Next, think about your objectives, as you need to be clear about what you want to achieve, and you need to measure your efforts, otherwise you won’t know that you are achieving your goals.

Third, you need to identify a clear scope. Claims automation is a huge space, and you risk boiling the ocean if you try to tackle it all at once.

How do you see claims automation going over the next five years?

We are seeing real progress in the number of claims that are not touched by the insurer and we are going to see more AI tools in the claims process. Today, if you are a claims handler, you are getting a document as part of the claim that you open with the same PDF reader that I use to read my payslip.

It’s the wrong tool for the job. If you give someone more targeted and intelligent tools you can enable them to work better. This will make the claims handling process more effective.

What would you like delegates to take away from your presentation?

I would like delegates to understand that automating claims handling is key to meeting customer expectations, and that dedicated AI solutions can focus your teams on the right claims and enable better decisions, faster.

Marcel Gordon, vice president of product at Shift Technology, is speaking on Day 3 at Intelligent Insurer’s Claims Innovation USA Virtual Event (July 13–15). The event brings together 60+ expert speakers on claims transformation and is free to attend for insurers and brokers/agents.  Register now to access 15+ hours of content live and on demand.