Tinsights

Applying AI and automation in claims processing

Published:

Yanna Winter from Generali outlines her talk on AI and automation to be given at Digital Claims 2022

 

Advert promoting the TINtech 2022 real world event


Transcript

Applying AI and automation in claims processing

Jeremy: So delighted to see you again. Yanna, and thanks for taking part in digital claims next month. I thought we'd catch up before the event and just to get a feel for what you're going to be talking about on the day because you're going to be giving us a bit of a case study around automation in claims. Perhaps you could give us a flavor of the topics and themes you'll be covering.

Yanna: I would like to pick in the area of digital case one specific example. This is where we're using artificial intelligence and smart automation specifically for bordereau management. And I can explain a little bit more, but for us, it is a very, very important subject and it has been a very good use case for one of our early uses of AI.

Jeremy: And how did you identify that process? And in the claims journey, how did you decide that's the one to automate?

Yanna: The process of bordereau management is complex by it's nature. We're receiving bordereaus from hundreds of partners. It's a complex ecosystem. Each document uses a different format, different data definitions, different meaning, different even channels through which we receive it. It is highly manual processing all of those bordereaus before we can bring them to our common format, when we can make sense of it. So wouldn't it be a great idea to automate to put intelligence, to overcome the laborious, repeatable tasks get them to a machine, automate it and bring to the humans only the data in the situations where the human used to make the decisio.

Jeremy: Right. And how much of that is intelligence and how much of that is automation, is that quite a lot of fire IA that process?

Yanna: I would split it 70/30 because the automation is 30. Specifically, we're talking, for example, some of the scanning with if we can put that under the automation, the specifically the robotic process automation, the RPA, the good old fashioned, if you want screen scraping, we're using them once the results are ready from the bordereau management and we want to enter them in various systems. So that element of the process entering into our systems is done with a good old fashioned automation because it is a repeatable process. It's still it can be further trained. But I would not suggest that that has the intelligence element. The artificial intelligence is in the first seven percent. This is where we're analyzing understanding the documents, understanding the key points of the key data in the documents, structuring it, also matching the data glossary of between hundreds of partners and us with potentially dozens of different definitions of the same term is quite a good case for artificial intelligence. Where we used a lot of natural intelligence before.

Jeremy: I'm really looking forward to it here in the case study at the event. So finally, before I let you go, is there one lesson that you could share with us? To do with automation that you would like to impart to the audience?

Yanna: Yes. Two lessons. One is and we all know that lesson. In I.T., there's no such thing as a magic wand and we prove it to ourselves every year with the next cool technology. But even with AI and automation, it is still not a magic wand. But what is really, really exciting about A.I., about automation is that that common model, because to be able to translate many to one too many definitions, data, et cetera, is quite a complex problem. And I remember many years ago people talking about natural language processing, which at the moment is taken for granted. At the time, it was a complex problem. Now, not even natural language processing, but let's call it bordereau processing. Really, the lesson is, wouldn't it be wonderful to have more industry standards? Maybe it's one of the next things that we can look forward to as industry to resolve it with maybe it's a waste of energy across our industry to be resolved in bordereau management, in various places with various formats, various definitions. Again, you can automate, we can apply artificial intelligence, but it would be great if we can work on industry standards as well.

Jeremy: Great. Thank you very much Yanna. Look forward to seeing you at the conference and hearing the case study in full, and thank you for joining us today.

Advert promoting the TINtech 2022 real world event

TIN youtalk weBlog Icon

youTalk
Weblog

HDI Global and HDI Global Specialty’s UK and Ireland branches will, for the first time, be aligned under a common leadership structur

What is your role at Harbour Underwriting?I am the business operations analyst for Harbour Underwriting.

0