How commoditized AI will reduce the impact of fraud on your bottom line.

A chat with Paul Fondie, Insurance Industry thought-leader and Chief Product Officer at Programable AI.

Hi Paul. You’ve been talking recently with insurers in the US and UK about the impact of fraud on their businesses. What’s the real situation?

Paul: There’s a lot of concern about the increasing impact that fraud is having on everybody’s businesses. What I’ve been hearing in my conversations with industry leaders is supported by the statistics. All the recent data confirms that insurance claims fraud is rocketing. In the US, it’s grown from $80 billion in 1995 to $308 billion in 2023. Even factoring in inflation, that’s still a figure that is almost double what it was 30 years ago. In the UK, reported cases of opportunistic insurance fraud increased by 61% between 2021 and 2023.

Never has insurance fraud been such a risk to our industry.

Andy: Those figures are alarming. What’s driving this dramatic increase?

Paul: The clear consensus from the industry experts I’ve been talking to is that it’s due to a combination of key factors. Digital technology is increasingly prevalent and accessible, as well as the general public’s ability to use it effectively. If you are intent on committing insurance fraud, it’s not challenging to edit a receipt for a television. Combine this with a shift towards remote and increasingly automated communications with insurers, accelerated by the recent pandemic, and the opportunities to submit fraudulent claims multiply.

This goes some way to explain the evolving array of tactics that fraudsters can employ. But it’s the well-publicised cost-of-living pressures, that much of society is experiencing, that is leading to the high increase in opportunistic claims fraud. It’s a perfect storm and it’s causing serious losses for our industry.

Andy: Don’t a lot of insurance companies have fraud teams and SIU’s to combat insurance fraud?

Paul: ?Yes, the vast majority of insurers do have a Special Investigations Unit, or SIU. Or you may be contracting this work out to a Third Party Administrator, or TPA. A recent PWC survey confirms that the percentage of insurers that have a SIU capability has risen from 87% in 2023 to 99% last year. But there is a fundamental imbalance at play here. Insurers want to accelerate straight-through-processing, to reduce associated costs and provide a streamlined and positive experience for genuine claimants. Unfortunately, the sophistication of this processing provides numerous opportunities for criminals to leverage the system, and the detection methods many insurers are using are not fit for purpose.

What I’m hearing increasingly is that insurers are realising they need to invest more resources and leverage suitably advanced solutions to give themselves fresh advantage and re-balance their fraud detection capabilities.

?

Andy: Okay, so clearly there’s a strong understanding across the insurance industry that they need to fight back. How’s that fight going?

Paul: It’s going okay. There’s no doubt that many suppliers of fraud detection and claims management solutions are leveraging technology advances, such as predictive and probabilistic AI, to counter fraudulent activities. The tools are available to support SIU activities to an extent but I am also hearing there are still big gaps in this line of defense, especially if you are a smaller insurers or some of the mutuals.

I’ve had a bunch of indicative chats at event in the US and UK recently and they’ve highlighted how the majority of software has not been designed with the smaller and/or the mutual insurers in mind.

Incumbent industry solutions provide a capability that can be over-kill in terms of operational complexity, offered at a price-point that only Tier 1 insurers can afford.

Furthermore, there is common industry feedback that there is no single software tool that can detect all types of existing and emerging fraud. This points towards all insurers needing to integrate a suite of solutions into their SIU processes to ensure a coherent and comprehensive fraudulent claim identification and management system. This applies to Tier 1 insurers as well.

There's a final point of interest here. I’ve heard lots of feedback from insurers that are starting to incorporate data analytics software into their fraud detection processes, but many of them have yet to extend an automated approach across their claims management process. This is confirmed in a recent report stating over 50% of insurers still do not use automated processes to refer claims to their investigators.

My perspective on this, in the nicest possible way, is that your manual system is getting in the way of improved fraud detection and stopping you keeping up with the influx of fraudulent claims. I’m not saying you need to remove investigators from the process, rather you need to provide the correct tools to help them be even better.

?

Andy: New software normally means new IT hardware, more employees to run it, and more training. Which all adds up to lots more cost. For insurance organizations struggling to maintain profitability, especially mutuals, does a path to digital transformation really make financial sense?

Paul: That’s a great question and it’s one I’ve been asked a lot recently. Whilst enterprise-level claims management software may work for the largest insurers, it is not designed or priced for more smaller organizations. What they need is affordable software that is simple to use and easy to integrate into their existing processes. A great example is Fraud Graph, that meets those requirements and, fundamentally, commoditizes a Tier 1 fraud detection capability into software ideally positioned for the smaller and/or mutual insurer.

?

Andy: Okay, using Fraud Graph as an example of a digital solution with easy-win benefits for insurers, can you tell me how it achieves that?

Paul: Sure. Fraud Graph plugs into your existing claims data system or takes in CSV files and uses cutting-edge and reliable AI to automatically prioritize the highest risk claims in your pipeline. Requiring minimal training to use, it has an intuitive and user-friendly dashboard that flags claims with the greatest potential risk of fraud and financial impact to the insurer.

?

Imagine knowing each morning the ten most suspicious claims as soon as you turn on your computer? And why they’re suspicious.

?

As we’ve heard from multiple SIU teams, it’s the ‘why’ that is key to delivering value as this tells them where to focus their attention and is a capability presently missing from the systems they are using.

So, yes, Fraud Graph is a great example of delivering to the insurance industry, especially for smaller insurers and mutuals whose requirements are presently unmet, a software tool that will help them fight back against ballooning levels of fraud. Due to its low cost, it will return a genuinely awesome ROI without any major disruption to ongoing operations and resources. This is the simple approach that insurers are increasingly looking for and that the insurance software industry should be providing.

?

Thanks Paul, until the next time.

________________________

If you’d like to learn how you can reduce fraud and boost profitability with Fraud Graph, a user-friendly and inexpensive solution, visit: https://www.fraudgraph.ai/