From Legacy to Learning Systems: AI's Transformation of Health Insurance Claims

Hello and welcome to the Sapiens Insurance 360 podcast. Today I'm going to be your host, Rob McLean. I head up our life, wealth and retirement insurance practice across Amir and APAC. And I'm so pleased that you're out there listening wherever you may be.

And this in the podcast is where we will discuss the latest news trends and issues from across the insurance solutions and technology spectrum. I think as everybody is well aware, insurance claims operations are facing significant pressures to transform and achieve operational efficiencies. For many years, challenges have been faced in terms of delivering change, more efficient processes, as well as improving the customer experience at the point of claim, with many claimants now expecting instant updates, faster payments, and a seamless digital interaction.

In conjunction with that, claims complexity is an area which increasingly drives change as well as customer frustrations. And many of those can be dovetailed with the insurer experience where we see greater fraud as well as complexity in the adjudication management of claims. In today's podcast, we're going to explore how insurers really need to adapt to the latest SaaS and AI initiatives to drive the future claims operations, providing that better customer experience. And in addition, we need to look at how the operational experience is very much tailored around the propositional challenges that insurers face.

To explore these, I've got Tariq Dedi, founder and CEO of Quantiv. Quantiv are an AI-powered platform helping health insurers and life insurers transform their claims operations through data extraction, automation, and advanced analytics. And I'm very pleased to say Tariq has got a smile on his face, not that anybody can see. Looking very pleased to be here with me.

Before founding Quntiv, Tariq led the data science and innovation initiatives within several major financial institutions, including AXA and Natixis, where he focused on applying advanced analytics to insurance and financial services use cases. So Tariq, thanks for joining us today and glad to have you with us. Thanks Rob, happy to be here. Awesome.

Okay, so to begin with I think a good starting point would be very much looking at what challenges did you see in the industry that really drove the need to change and to come up with the idea of Qantive? Yeah, so it's quite It's intertwined with my professional journey in some way. So I started my career at AXA Group, which led me to work with many of our AXA entities across the world. And one thing that stuck me while deep diving in the different claims operations at the time, especially in the health and life business line, is how manual the processes are and how most of the complex decisions that need to be taken are taken solely based on the know-how of the armies of nurses, doctors that are spending their day reading documents and making tons of checks based on their know-how.

And all of this while we are seeing a tremendous increase of the volume of medical claims and this rampant medical services inflation of 10% a year all across the globe. And all of those observations made it clear for me that things needed to change. so that insurers can scale their operations if they want to get a bigger portion of the pie, if we say it this way. And while they deliver, they need to still deliver a better quality of service, which is mostly driven by how fast they can adjudicate the claims they receive and also be able to go beyond being a simple payer, being a partner.

And for that you need resources instead of focusing your resources on tasks that can be a bit of a low value for when you look at them nowadays It allowing them to focus on more advanced services for the members, especially when it comes to health where a lot of members are expecting not just a payer, but somebody that can help them through their patient journey. Excellent. So I think certainly from our experience, we see very similar aspects where claim administration has become a commodity.

The knowledge that resides within claim assessors' heads is very constricted and can create delays, backlogs, as well as increasing the demand on those resources and, of course, the cost that goes with them to the insurers. So definitely automating and streamlining that has not only a positive output for the insurer, but also a better experience for the claimants who are interacting with them. If we look at from the insurer perspective, claims management can often be seen as an expensive overhead.

As part of that, where do you see Qantive really enabling the improvement of the processes and the decision making within the organisations? Yeah, so I see it in two ways. It's first, increasing the output of those claims operations, enabling them to handle a much higher volume of claims faster. And secondly, automation.

And for me, it's always a two-step aspect. But automation is very hard when it comes to dealing with coverage, in our case, coverage of human health. and for me all of this when it comes to increasing the trial putting in place automation it starts with the data you can have all the fancy AI models or operation workflow orchestration tooling if the data flowing is bad the waterfall of decisions that come afterwards is going to break so that's why we need to start with automating first of all data entry still a lot of insurers on the market need to deal with a huge inflow of unstructured data, of documents coming from providers or members that needs to be kind of put into the system.

And a lot have experimented with OCR and everything, but the granularity of the information that is put into the system or by human-driven data entry operations or by OCR-driven operations is still not granular enough and not at the level of quality that is required to be able to do any kind of intelligence and automation afterwards. Then when you solve that step using a modern approach, we need to tackle the other topic, which is all the tech checks that need to be performed from contract intelligence or understanding the coverage, medical intelligence, understanding the different medical aspects of the claim, and finally, the fraud, waste, and abuse intelligence so that you are able to do any kind of automation at some point.

Interesting. I think you raised a good point there, which is from our experience, again, we see a lot of insurers looking at the real use cases and value AI can bring to their organization. We all acknowledge the movement in the industry towards more AI-driven capabilities, but identifying those key areas where AI can add real value is sometimes challenging. And it certainly sounds as though you've pulled out a real use case here, and that's at the foundation of Qantive's proposition.

To be honest today, the AI adoption rate is still quite low when it comes to claims operations, especially on the health side. It a buzzy word that the industry has overused and that every software vendor has sprinkled over their products to the point where there is a bit of disillusionment today if we can say so And nowadays you know with all the agentic craze there is also a bit of the insurers are a bit worried when it comes to that I think they evolved a lot and now they expect ROI from those initiatives.

And they expect also a mix of both AI expertise and domain expertise. They don't want any more these kind of off-the-shelf solutions that are targeting the agro industry, the car industry, and at the same time, the insurance industry. They want people that understand their business and that can apply and that can bring solution, AR-driven solutions to that. So I shared in the previous question, really it starts with solving the data question.

And that's on its own, it's hard. It's hard being able to extract the right information to infer what is the right CPT code to apply to a line item, what is the right ICD code to apply to the diagnosis. And bringing the right tooling to that operation is already a challenge. But what gets us excited at Comtev is trying to reproduce through AI all those checks that are currently done by the claims operators that are often nurses or doctors that are working in the back office.

When deciding what to do with a claim, they need first to be able to check if the claim is covered by the policy. They need to check if the care that has been billed is necessary from a medical standpoint. They need to check if the medical services are coherent with the diagnostic, if the price of the service is coherent with what is observed on that specific market. So many, many checks that live in their heads that are based on their cognition that needs to be reproduced through AI.

And all of those checks require some specialized approach to understand what is the grounding data that enables them to make a decision. And at the end of that, you have to add all the capacities in regards to fraud, waste and abuse when it comes to identifying the anomaly and then explaining it to make it actionable for the SIU teams to be able to recover or stop any kind of dynamic that is going there. So all of this requires tons of AI techniques to be stacked together and to be embedded into a software that is capable, that understands how those operations work and be part of their transformation.

I think it's a good point that you've raised there when we look at the completeness of proposition, certainly from a lot of the capabilities capabilities that may be able to be sourced in the market. Many of them don't provide that full end-to-end capability of the data entry through the administration, the adjudication, but most importantly also flagging and identifying the fraud elements or where human intervention potentially needs to come into play to really call that out in the retrospective looking at the performance through the life cycle of the claim management process.

So it certainly sounds from a Contiff perspective as though you've brought that all together in that one cohesive offering. If I was to pull a crystal ball out now, and certainly fortune telling is a unique art in our industry, where do you see the future if we were to look five years ahead from now from a claims operating model and a technology and legacy standpoint? I believe that's a great question. I believe claims is going to see some heavy changes in the years to come.

Status quo is not an option anymore. You cannot just throw money at the problem. Now insurers understand where the impact can be. They have experimented enough and they are looking for technology that can blend, as I said, AI expertise with domain expertise.

So I see a bigger chunk of the volume of claims being automated especially the ones that we believe that were too complex that were too sensitive we are going to see a huge jump in terms of trust in AI models when it comes to tackling those specific claims that we believed are difficult to handle through an automation through an automation setup. But I see also more augmented style of claims management. The software is not only about automating, it's about augmenting the capacity of some specialized operators that are working in the claims operation.

So it's by bringing them this kind of 360 view of the claim, allowing them to study complex patient journeys, highly sensitive wines, VIP kind of claims in a more thorough way, in an enhanced way. So that to, again, come back to that trout fruit level, you have these very specialized operators, very knowledgeable experts that just need to be enhanced in their ability to handle more claims per day than they are doing today because of a lack of tooling. Excellent. So it sounds like a very much a strategy aligned to insurers' agentic frameworks whilst recognising that need for human interaction and then supporting the human interaction element of the processes, which we all acknowledge will still be there in the future.

I think I am slowly running out of time, but maybe we have time for one last question. For any of the insurers out there that are beginning their AI journey, specifically around their health claims operations, what would you see or recommend as the first step they could take and what their strategy should really look like? So one of the mistakes that we see often is that insurers acquire a new tool, AI-driven tool, before thinking how it's going to transform their operation. So for me, always the first step is having a clear transformation plan of what they want their claims operation to look like after implementing new tooling and bringing some AI capabilities.

And of course, whatever that transformation plan is going to be, the first step is to make sure to start with the data, making sure that the data that you are getting that you are starting with to enable that workflow is at the right level from the get-go. If not, it's going to just create issues afterwards. Excellent. And Tariq, thank you ever so much for sharing all your insights today.

I think your perspective on transforming the claims process, our risk exposure and innovating in the claims space is very insightful for certainly us and hopefully also the listeners that are out there. You clearly show that you're at the forefront of the technology and technical shift in the industry in this area. and certainly from a Sapiens perspective, we're very proud to be partnering as an ecosystem partner with Quantive. To all our listeners out there, thank you ever so much for joining us as we've deep dived into the future perspective of insurance claims.

And if any aspects of today's conversation really resonated with you, we'd love to hear your thoughts. Please feel free to connect with us on social media, share your feedback on this podcast, even if you don't necessarily disagree or agree with some of the points and subscribe to the podcast so you can hear any of the new and exciting episodes that are still to come. Until next time, this is Sapiens Insurance 360 and thank you for listening. Thank you.