Hyperscience at Insurance Innovators: Streamlining the Future of Insurance

Lindley Gooden: Hello. Welcome back. I’m Lindley Gooden. Welcome to Insurance Innovators TV. Let’s talk about automation in insurance. There are lots of things to talk about with Chris from Hyperscience. Chris, how are you?

Chris Bloomfield: I’m very good. How are you?

Lindley Gooden: Very good. Well, first of all, tell us a little bit about who you are and what you do.

Chris Bloomfield: Yeah, so I’m Chris Bloomfield and I lead our pre-sales team at Hyperscience, more specifically for our international business. And effectively anything technology-related when it comes to business processes and automation is responsible for that.

Lindley Gooden: Brilliant. So let’s talk about the obstacles, the difficulties, the challenges in insurance right now around automation and document processing. You have a list. Let’s have it.

Chris Bloomfield: Yes. So I think about this from internal pressures and external pressures. What I class as the five Cs. So really the first one being Cost. Every insurer’s looking to drive cost out of business processes, but underlying that is the Complexity. So the complexity of the actual business process coupled out with the complexity of the customer information that’s feeding into that particular, let’s say, claims process there.

And then we’ve got Compliance as well. So you wanna do more with less. We wanna do that in a highly complex environment, but you have your DPO turning around saying, “Well actually, you can’t use this data potentially ’cause it contains PII information.” And then we’ve got the external regulations as well. And then if we think about Customer Experience as well. Customers vote with their feet very quickly. If you’re not getting good customer experience when you have to unfortunately speak with your insurer, then you’re gonna vote with your feet. And that leads to that last C which is the Competition. If you are not delivering world-class service to your customers very quickly, your competition will be.

Lindley Gooden: It really helps to sort of put a hat on that and to work out exactly those corners that need to be tapped. Now you have a solution. I wonder how you are tackling these issues. I mean, there’s five issues there. I’m sure there are more in principle, but how are you starting to work those out?

Chris Bloomfield: Yeah, so it’s really interesting. Um, when we look at technology as it applies to maybe automating business processes and we think more about the insurance claim, when we speak to insurers, we ask them what their priority is. Do you want to automate or straight-through process these particular claims, or do you want the most correct or most accurate data feeding your backend systems in order to make the decision? A hundred percent will respond that they want the most correct data.

What’s very interesting, all technology solutions are focused on automation. And it’s automation, automation, automation. And actually accuracy becomes a byproduct, or the correctness of the data becomes a byproduct. We realized very quickly that when you scratch under the surface of documents that have been automated in the claims process, actually it’s somewhere about sort of 40 to 50% actually are correct in terms of the data. Right. So now we’ve got incorrect data littering your backend systems making bad decisions.

So at Hyperscience, we are reimagining it and kind of flipped it 180. And what we call a Hypercell, which is a turnkey solution that actually predicates everything on correctness of data. So instead of trying to automate documents as they flow through your front door as an insurer—we are not arrogant enough to think that actually we can automate all of that—instead, what we try to do is automate human routine. How would a human actually interact with these documents? So we break them down into discreet machine learning models, but what’s really key is we allow you through what we class as a Correctness Orchestration layer, is to allow you to actually say to our models and our digital workers, “Hey, I want you to be 99% correct in everything that you do.” Or 90% whatever. But you dictate the correctness as it flows through.

And much like a human, our digital workers will be working on a document. And if it can’t do something to your correctness requirement, instead of blindly actually doing the work, pushing it through… hand up. And we introduce a human into loop.

Lindley Gooden: Oh, that’s important then because you still wanna be able to trace, you wanna a paper trail of where the inaccuracies are and so that’s the point you can put your hand up and say, “Right. Okay. Now devote some attention to that as a human.”

Chris Bloomfield: Absolutely. And these are kind of micro, hyper-focused tasks. So for example, let’s say we think about turning pages over. If a stack of papers and say, “Actually I don’t know what this belongs to. Hey human, can you come in and just help me? Because I can’t do this at 99% quality that you are asking me to achieve.” There’s all of these kind of microtasks and this handoff between humans and digital workers that enabled us to hit any kind of target accuracy. Most organizations at least want to get to sort of 98, 99% accuracy with unprecedented levels of automation.

Lindley Gooden: Is that helping the humans? Because you’ve got a focus task. You’re being brought in at a certain point, you know there’s an issue there so you can drill down rather than just processing lots. Is that good for the human?

Chris Bloomfield: Absolutely. Because it enables you to focus. If we think about the old approach, which is “I’m gonna do all of this work, there’s something wrong with this document, hey human go and review it…”

Lindley Gooden: Errors all creeping anyway.

Chris Bloomfield: Errors, I look in, I’ve got limited visual cues. I don’t really know where I’m looking and actually I distrust the whole thing and I end up reviewing the entire document. Instead giving me a hyper-focused task. I’m zoomed in five, 10 seconds and I’m done.

Lindley Gooden: Absolutely. And you know, I think we know the answer to it, but why is data accuracy so important to make this work?

Chris Bloomfield: Well, I think all of the decisions that are made in terms of business process, think about claims. Do I approve? Do I reject? You need the highest quality data. And I always think about an analogy of feeding an Olympic athlete for an argument. Say, let’s take a Usain Bolt. So Usain Bolt, you want to get him Olympic gold and his performance on the day, it’s gonna be predicated by all the preparation that’s taken in place. And that includes nutrition. Now if I fed Usain Bolt fast food for a month leading up to the Olympic Games, what’s performance going to be like?

Lindley Gooden: Fast food slows you down.

Chris Bloomfield: Absolutely. So focus on this automation, automation, automation, but foregoing the accuracy is exactly the same. By kinda slowing down, focus on the accuracy of the content will enable your application to your decision making processes for it to be a lot faster and of the highest quality.

Lindley Gooden: Before we go, just any examples or a short example of how it’s worked and how you’ve sped up that processing, made accuracy much more part of the equation, and helped the clients out?

Chris Bloomfield: Sure, absolutely. So we’ve been working with an insurer out in Ireland, focus on complex cancer claims. Now prior to working with Hyperscience, this was anything up to six months. Highly manual process, lots of information from disparate sources that you need to get in order to make that decision.

Lindley Gooden: Very emotive as well.

Chris Bloomfield: Absolutely. And you think that myself as a claimant, I’m waiting six months and then I get a no. We’re not gonna pay out.

Lindley Gooden: It’s probably the worst time of their life.

Chris Bloomfield: Absolutely. So we’ve been working with the insurer to ensure that we are getting the data out of all of these disparate sources at the highest levels of correctness. We are then combining that with the power of Gen AI to be able to summarize and create dashboards for the insurer to, you know, red, amber, green, what do I need to look at? And we’ve taken this process down now into a handful of weeks.

Lindley Gooden: Great. That’s a real helping hand that the end customer needs. Which is good Chris. We could talk forever, but that’s it for now. Thank you much, much for joining us.

Chris Bloomfield: Well, thank you. Great to have you here.

Lindley Gooden: And for now, that’s it from Insurance Innovators TV, streaming the future of insurance. See you next time.