Unlocking the Future of Underwriting | Sompo at AIPCon 4 — Transcript

Full transcript

1. 0:05Thank you for having me today.
2. 0:07So I'm Akiko Murakami.
3. 0:09I'm a chief data officer of Sompo Japan Insurance.
4. 0:13So my career starts as a researcher in AI, especially in the natural language processing area and the writer that transitioned into the software engineering and doing to add related things over 20 years.
5. 0:28So I'm familiar with A.I..
6. 0:31So I am so excited to hear that AIP is now available on Foundry.
7. 0:38So today I am honored to be here and share the experience.
8. 0:44So before diving into the contents there, let me introduce the company first.
9. 0:51Sompo Japan insurance is part of a sample holdings group.
10. 0:55The operates in the various sectors there, including the world wide PNC Insurance, life insurance, elder care, business and digital business.
11. 1:06So the Sompo Japan insurance handles the PNC business in Japan.
12. 1:11And the it is the largest company within the group who is the 135 year history.
13. 1:20So at Sompo Japan insurance, leverage the Foundry extensibility for this show, make at sales and sales branches and the writing processes and the claims, the handling process like the disaster response and more.
14. 1:37So today I'd like to talk about the how have been and will continue innovating our underwriting process.
15. 1:46So I would write to explain what's the underwriting process in the insurance biz.
16. 1:53So underwriting in insurance is the process of separating the risk of ensuring the person or asset to and determine the appropriate premium and conditions.
17. 2:07This is a very critical process as it determines the profitability of our business and whether we can continue providing the our good service and peace of mind for all of our customers.
18. 2:21So the Foundry enables our underwriters to make right decisions based on the key information, such as the various resources, such as the internal document, contract terms and the historical information on accidents.
19. 2:38So premium characterization and simulation can be the perform the tier that they consider the various types of scenario.
20. 2:46So many simple tests are already automated and the underwriters can focus on the process where their expertise has the most impact.
21. 2:56So what was the impact?
22. 2:58So over the last three years, we have seen the 60 million U.S.
23. 3:04dollar improvement in profit.
24. 3:06And we expect an additional 100.
25. 3:10The median US dollar over the next three years.
26. 3:14The due to the three main focuses.
27. 3:17The first the we implemented more appropriate pricing based on the date accurate the data supported by a holistic view of the data, various types of data and logic there a working behind the scenes.
28. 3:32Secondly, the sales represented the can the properly explained the reason of the premium, especially for the the barrier is the higher than the last year.
29. 3:48The finally there was a huge improvement in efficiency in the underwriting task of daming the light information at the right time.
30. 4:00Okay. So now we have to understand the impact.
31. 4:04Let's see what this looks like in practice.
32. 4:11So here the demo burned down the writing application that the including the under development functions and the right tiers of the incoming policy requests which are now automatically press the into the all but the companies we want read the into the three patterns and the process attention and be justified basis and the approval rule based approach to prices.
33. 4:36The key differentiator with the IP is now we can provide the guidance complex policy request by putting together all the information at the same time.
34. 4:45First, the AIP analyze the order information on new policy request and identify all the relevant underwriting standards which are the hit by the company.
35. 4:55Additionally, the AP search the order composition between the sales representative and the customer.
36. 5:02So the are previous claims and the industry.
37. 5:07The based on this information, the risk runs can be provided and to the premium pricing algorithm and the the rising the suggested minimum premium.
38. 5:18So considering all the details of the AIP can provide the suggested action plan in this case, the it suggests the rigorous and more current information and the increase in the salesperson suggested premium amount AIP suggested will not always be correct, since that's the humor in the group is required.
39. 5:41The most important part is that the the user is to be captured, the feedback and to improve the feature responses.
40. 5:50So finally, the underwriters can submit their response by providing a justification of the the addition and the full details of the policy request and the responses and will be written up to the share with the agency.
41. 6:06Okay, so let's see.
42. 6:07That concludes my presentation.
43. 6:09So as you see in the in this video, our AI-powered underwriting workflow is enhancing efficiency and the corporations and it improves the decision making for both Cirrus and and the writers.
44. 6:24So Foundry and AIP have significantly improved.
45. 6:29So the efficiency of our process and reduce the manual work and the enabled underwriters to focus on the very compressed cases.
46. 6:41So we are so excited about what about we are already being able to do and even more excited whole what the future of both.
47. 6:52Thank you very much.