According to Jimmy Spears, head of automotive at InsurTech Tractable, “Speed is everything” in auto claims adjusting.

“Nothing gets cheaper over time,” he said during a recent interview, addressing a question about whether the computer vision company, which uses artificial intelligence to visually assess car damage, might be able to help auto insurers lower costs in inflationary times.

Spears and Tractable’s Co-Founder and CEO Alex Dalyac spoke to Carrier Management about Tractable’s recent successes in partnering with U.S. carriers to speed auto claims handling processes and about innovations in the works for property insurance at a time when auto insurers were in the midst of reporting second-quarter combined ratios above 100 and sizable premium rate hikes driven by soaring parts prices and escalating repairer costs. (Related article, “Tractable’s Next Move: Property Claims Assessment Without Damage Photos,” featured in CM’s third-quarter magazine with other articles on claims and supply chain impacts.)

Could AI-driven estimates of repair and labor costs derived from photos of crumpled hoods and dented fenders help carriers at a time when nothing else seems to be working?

“Speed, accuracy and service—that’s really where your advantages come in as a carrier,” Spears said. “The sooner you can visually assess if a vehicle, No. 1, hopefully from the point of accident, is either repairable or non-repairable, that’s generally one of the most effective decisions that you make when it comes to having anything to do with the indemnity portion of a policy,” he said, referring to triaging decisions that Tractable’s AI can handle.

Next, with AI not only deciding if the vehicle is repairable but also writing the appraisal, the car can get to a shop and parts can be ordered immediately “because nothing gets cheaper.”

“It’s our AI that can make the humans available much more when they’re needed.”

Jimmy Spears, Tractable

“There’s a lot of issues today. You’ve seen the logistical issues with parts coming from everywhere” and supply chain delays. “Getting those parts orders in and then getting those parts orders fulfilled really helps with that cycle time. So, you’re not sitting in a situation where, but for one day, had you ordered that part earlier, your customer could have had it.”

Beyond speed, accuracy matters. A series of inaccurate decisions can mean periods of uncertainty for your policyholder. “That’s how you lose that business,” Spears said, concluding that a high-service engagement for a customer comes as a result of getting the speed and accuracy pieces right.

Focusing more specifically on the inflation question, Spears illustrated how process delays can make claims more expensive to settle and damage customer relationships with an example. “There’s nothing worse for a customer than to put it down the repairable path, and due to you being slow at getting an appraisal, finding the parts, doing what you need to do, now that vehicle may be in an upside-down situation…You have to tell this person, we now have to take it to a total-loss situation when in their mind they have already decided they’re going to have the car repaired. They haven’t even gone to look at other vehicles that are in the market…”

“And vehicles are not getting cheaper. The used cars are even at a higher amount of money,” Spears said.

Dalyac honed in on the accuracy piece of the speed-accuracy-service formula for claims handling. “Thanks to the visual assessment, the AI’s able to predict: Should you replace that part, or can you repair it? Should it be three hours of labor to repair that quarter panel or two?”

“By ensuring that it’s accurate, each time you’re preventing an unnecessary replacement or an unnecessary hour of time,” Dalyac said. “The impact of that is even higher because of the inflation.”

“Increasing the accuracy by one repair operation or one hour is just that much greater when those prices and those rates are higher,” he said.

“Increasing the accuracy by one repair operation or one hour is just that much greater when those prices and those rates are higher.”

Alex Dalyac, Tractable

Asked whether Tractable could help beyond getting assessments done quickly and accurately—specifically whether the company has access to more repairers or more parts than everybody else, Spears referenced a relationship that Tractable has with alternative parts supplier LKQ. Last year, the two companies announced that LKQ would use Tractable’s AI to accelerate and optimize the recycling of the salvage vehicles that LKQ procures in North America—essentially, using computer vision technology to assess specific damage on each vehicle and determine which parts could be recycled and reused.

“I don’t believe anybody actually purchases any more salvage vehicles than what LKQ does,” Spears told CM. “We really are in a great position to be able to address that inventory if asked by a carrier to make sure that they could have first identification of what parts would work within that appraisal.”

Spears referred to the fact that inflationary pressures are causing carriers to continue to increase premiums, noting that Allstate, for example, saw earned premiums jump 8.6 percent in the second quarter (with rate hikes fueling the increase), but the combined ratio was still 106.9.

“It’s tools like our AI that will help bring down claims-associated expenses [and] then also indemnity.” AI can “really help keep that indemnity very accurate. Pay what you owe by helping those decisions be made properly upfront very fast.”

Training AI

During the interview, Spears and Dalyac also responded to questions about the data sources used to train the AI—specifically whether vehicle damage photos, parts cost data and repair times are specific to individual carriers or whether they are aggregated together.

“It depends on our contractual agreement. But quite a few customers will tend to opt into the pool because they can get such a big AI performance uplift,” Dalyac said.

Spears said labor rates are supplied to Tractable by the carrier. “We don’t invent those,” he said, confirming that they are carrier-specific. “Every customer would certainly supply the type of labor or the amount of money that they’re paying in certain markets.”

“Our AI goes far, far above that, however. That’s really sort of a rudimentary thing. It’s more of a rule. What our AI does is to actually see the damage and then start to calibrate it the way that the repair would be written by the [carrier] customer. ”

“That’s really what our advantage is—our AI is seeing the images. What labor rate gets applied to it, that comes later on within an estimating platform,” Spears said.

“The parts prices we’ll obtain through our partnership with Mitchell International,” Dalyac said, referring to aggregators of parts prices, such as Mitchell, Motor Information Systems and CCC, which Dalyac identified as the dominant solution in the United States. (Editor’s Note: CCC has filed a lawsuit against Tractable alleging that a Tractable employee obtained a license to CCC’s estimating software using a false identity in an effort to access proprietary data. Tractable denies the allegations. Related article, “No Arbitration in CCC’s Trade Secrets Lawsuit Against Competitor, published by Claims Journal)

What’s Next?

Spears and Dalyac also spoke to Carrier Management about Tractable’s expansion from auto insurance claims to property and the prospect of allowing carriers to respond to disasters before getting claims notices from policyholders—with AI flagging damage locations and severity even before damage photos are available. (Related article, Tractable’s Next Move: Property Claims Assessment Without Damage Photos”)

Beyond insurance, Spears is also helping Tractable to find opportunities in the automotive ecosystem, for example, working with sales groups to deliver accurate trade-in values from vehicle photos.

Are there other opportunities within auto insurance—in underwriting instead of claims, or in liability instead of physical damage?

Spears says yes to both propositions. “We believe that our AI, if partnered with some of the right data sources, could really start to take a look at what type of injuries would be associated with an accident,” he said. In addition, he said, “I believe our AI could do a pretty good job at looking at points of impact, although Tractable hasn’t yet been asked to reconstruct an accident.

Asked what’s next for the overall property/casualty industry, and specifically whether there’s a future for human claims adjusters, Spears says the role of AI is simply to bridge gaps during periods of uncertainty. “There’s always going to be a role for people in a people-associated business.”

“When you do need that expert, you want that expert,” he said, contrasting a slight damage situation with the total loss of a home. Working in Florida after Hurricane Andrew, Spears saw cement pads where houses used to be. “That homeowner is going to need a lot more than AI.”

Even in less catastrophic events, the USAA veteran sees a role for human adjusters. “You may have a single parent. It could be a military situation. One’s deployed overseas. And they’ve lost the only vehicle that they have in the family. They need someone to help them. That’s when you need to be able to take care of what you can with AI and then escalate that very quickly to a person…”

“The secret to working with AI [is] knowing when to apply it and knowing where to apply it, and then knowing where to have it really escalate it to the expert that’s going to be there at that time of need.”

“It’s our AI that can make the humans available much more when they’re needed,” he concluded.