3 Examples of the Effective Use of GenAI in Auto Subrogation
Introduction
Generative AI (genAI) is an extremely effective capability when applied to subrogation. GenAI analyzes free-form, unstructured text buried in claims notes and loss descriptions to uncover critical key words, phrases and context that identify a viable subrogation opportunity. Combining these claims indicators with local jurisdictions and relevant data sources yields an accurate estimate of liability quickly and consistently.
Below are three auto examples that illustrate the impact of genAI in various scenarios. Each example contains a claim summary that resides in the insurer’s claim system. The claim details are interrogated for subrogation potential using genAI and advanced analytics. An alert score provides an indicator of viability, at the exposure level, and the reasoning is explained with complete transparency.
Driver PIP
In this example, a Driver PIP exposure was identified as a subrogation opportunity in addition to collision. The combination of driver actions and livery driver status were found using genAI.
Claim Summary
Accident while changing lanes: Insured was traveling down 3rd Street when a driver (claimant) who was on the side of the street and had been parked pulled out into the lane and hit insured.
Key indicators found by genAI
- Driver PIP Recoverability: Relevant state law for PIP. One vehicle is used for transportation of person or property for hire (livery)
- Third Party vehicle listed in NYC Vehicle for Hire Database: License Number: 5574537; VIN: 2T3BK1BA3BC118776; Name: Smith, Alexander, H; Company Name: UBER USA, LLC
Collision
Dense, free-form claim description is analyzed by GenAI. Complex state negligence laws are interpreted to estimate liability.
Claim Summary
We were in a vehicle accident at 8:10 this evening while driving our 2017 Mazda CX-5. I exited the parking lot, crossed westbound Erie Blvd and as I was turning east onto Erie, I collided with a blue 2012 Chevy Malibu. The right rear of my car came in contact with the left front fender of the Malibu. The Malibu was going eastbound on Main Street in the right lane. I believe the Malibu was simultaneously getting into the left lane when the collision occurred. Further, I think the Malibu may have been exceeding the speed limit at that time. This area of Main St. was recently re-paved and did not have painted lane markings - so no clear lane indications. Both vehicles are drivable. I immediately parked the car and met with the other driver to exchange information. In addition to the driver, there were two passengers in the other vehicle.
My passenger complained of neck and shoulder pain. No one was injured in the other car. The collision was a glancing blow, not a hard impact.
Key indicators found by genAI
- Insured Vehicle Contact Point: Right rear
- Vehicle 2 Contact Point: Left front
- Loss State and Comparative Negligence Laws: NY - Pure Comp Neg
Method to identify 3P found with GenAI
Claim details indicate police involvement and a police report that identifies responsible, third party information.
Claim Summary
Insured said her car was hit while parked on the street. Driver said he did not have brakes. Driver also hit mailboxes and telephone pole.
Key indicators found by genAI
- Relevant text: Named insured having issues with PD. Having issues getting a copy of report. Asked NI if they can obtain at least the report #. We need to find C’s info.
- Time delay: Adv we would request PR. (4 month delay and claim still open)
Summary
As seen in these examples, GenAI is a game-changer in identifying hard to find subrogation opportunities. The speed and effectiveness of GenAI is a capability insurers should incorporate as part of their subrogation strategy.
Shift Subrogation helps insurers identify millions more in recoveries. AI enables accelerated analysis of claim details in a consistent manner resulting in trusted liability estimates by exposure. The result is a faster, more efficient process to identify opportunities that accelerates investigation activities and improves results. Learn more about Shift Subrogation.
Interested in learning more about Large Language Models?
The continued advancement of reasoning LLMs creates new opportunities to apply GenAI to important insurance use cases such as subrogation. Shift has done extensive research about how specific LLMs perform when applied against specific insurance
tasks. Read more in The State of AI in Insurance (Vol. VI): Claims Decisioning and Liability Determination in Subrogation
.png)
