Insurance Underwriting Gets an Upgrade
AI Adoption, Integration Lets Providers Move Beyond Manual Practices
Key Takeaways:
- Mass volumes of data have become available since the ELD mandate started in 2017.
- All-at-once or gradual? Companies' approach to AI adoption and integration depends on their needs.
- Some businesses use AI to augment their current practices.
[Stay on top of transportation news: .]
Artificial intelligence is becoming increasingly commonplace in modern life and in business, from aggregating search results and summarizing complex documents to acting as a personal assistant through smartphones.
In the trucking industry, insurance policies represent one of the new fields where AI is beginning to have an impact.
Some insurance firms, such as trucking insurance provider Nirvana, are taking an AI-first approach to risk assessment and underwriting.
“There’s something about insurance that is deeply flawed,” said Bandar El-Eita, ’s head of marketing. “In any cohort, you are going to blend the risk.”
When pursuing an insurance policy from a traditional firm, companies are evaluated based on their industry profiles and placed into a cohort, he noted. The cohort then pools the risk and the insurance provider sets the rates accordingly.
·ˇ±ô-·ˇľ±łŮ˛ąĚý
In trucking, that approach means that the safest, best-run motor carriers in a cohort effectively subsidize the group’s higher risk carriers, which have little incentive to improve, El-Eita said.
That’s where Nirvana’s technology comes in. Before setting a rate for a potential trucking client, the insurance company hooks into the fleet’s telematics, electronic logging devices, dashcams and advanced driver assistance systems.
With the implementation of the ELD mandate starting in 2017, onboard technology on trucks has been quietly collecting all sorts of usable data on a fleet’s performance in real time, El-Eita said.
“The obvious question became [if] I’m making these improvements with ELDs and dashcams, how can I leverage this data to save myself some money on insurance?” El-Eita said. “The way we look at it is, while you can save up to 20%, it’s really about finding the right price for your actual risk.”
Steve Miller, vice president of innovation and mobility for insurance brokerage , has worked with innovative insurance products since 2015, when he first insured a company working on an autonomous vehicle.
Since then, he has watched as the conservative, highly regulated industry that he’s in has struggled to adapt to new technology.
˛Ńľ±±ô±ô±đ°ůĚý
He said he sees two tiers of AI implementation and integration in the industry: First, the companies that are all-in on the tech and look to use it wherever possible; and second, companies that are more reserved in their adoption and are poking at the technology to figure out how it best fits their business model.
“The industry is struggling right now to absorb the technology,” and make it do more than just answer simple queries or generate clever slide decks, Miller said.
The common thread of both AI integration approaches, he said, is the fact that both factions realize the treasure trove of data to be gained from a trucking company’s telematics.
“AI is enabling that dataset to be more readily and robustly digested, analyzed and then used from an actuarial perspective. Actuaries take historical information to price future risk. And what we need to do is disrupt that process and accelerate adoption of real-time data,” Miller said.
The difficulty, he said, is when insurance companies want to take that new way of pricing and adapt it to the patchwork of rate requirements and regulatory frameworks across each of the 50 states.
“That’s a big part of what’s slowing things down,” he said.
AI-Driven Efficiency
Other insurance companies are finding ways to use AI to augment and speed up existing business processes.
has been using AI for applications such as claims fraud models and sales opportunities models, said Denise Christophel, director of advanced analytics at the firm.
“We use machine learning for things that are not decisioning models,” she said. “We are not using AI to make decisions. We’re using AI to help inform our associates as they are making the decisions.”
°äłó°ůľ±˛őłŮ´Ç±čłó±đ±ôĚý
The company is using AI to help generate claim summaries and support customer service representatives with real-time recommendations during customer interactions, but none of the interactions customers havee with the company pass through an AI filter.
Future projects are aimed at helping underwriting work more efficiently. Some of the manuals connected with the firm’s process encompass hundreds of pages, and finding a precise answer to a specific question can be time intensive, Christophel said.
Nick Saeger, assistant vice president of products and pricing for transportation and specialty at Sentry, said the company has put its focus on AI applications that allow people to focus on higher-level interactions and projects while computers handle the mundane.
ł§˛ą±đ˛µ±đ°ůĚý
“These are things that allow them to take more time to think through. On the claims side, it’s [allowing them] to communicate with claimants, with attorneys, with our customers, and do the things that need to be done to progress the claim and get it to its resolution,” Saegar said. “On the underwriting side, it allows them to not be leafing through or digitally leafing through an underwriting manual to find a particular thing. We can find it more quickly and now we can get at underwriting the actual account itself and it’s spending more time getting to the right price. That’s why I love it from a business perspective.”
Peter Berg, principal and practice leader at , oversees the company’s final-mile transportation products.
µţ±đ°ů˛µĚý
“That sector is ripe for opportunity because that final-mile component makes up [about] 40% of the cost of the supply chain,” Berg said. “There’s a ton of technology in terms of how you automate, how do you scale and how do you take out some of the overhead that is associated with that model, which is very much human labor driven.”
Berg said he’s increasingly seeing the insurance companies his firm brokers use AI to drill down into data to offer more accurate pricing and create better risk assessments.
Reducing Risk
AI may also come into play when managing risk and limiting liability.
“Fleets are subject to nuclear verdicts. That’s just a reality. So are insurance carriers because they’re the ones that are on the hooks paying for it,” Hub International’s Miller said.
Increasingly, fleets have been deploying telematics systems, advanced driver assistance systems, and inward- and outward-facing cameras. All of those systems generate data.
Mark Hill and Danielle Villegas of PCS Software discuss their AI engine, Cortex, designed specifically to level the playing field for midsized carriers. Tune in above or by going to .ĚýĚý
“Insurance companies are starting to get better at integrating and taking that data and using it to price to the degree that they can,” Miller said.
The real-time data collection and interpretation also can allow for real-time driver coaching and faster identification of driver safety risks, he said.
Nirvana’s El-Eita said using AI to speed along the process after a claim is filed can help mitigate claims expenses as well.
“If you can resolve a claim faster, you can get everyone back on the road faster,” he said.
Plus, being able to provide hard data from the driver’s trip experience leading up to the claim event has proved valuable in helping ward off some lawsuits that might have tied up the company in legal proceedings.
“How did they drive? Were they crazy people behind the wheel or were they driving with all diligence and care? And that can really protect the fleets from more fraudulent factors,” El-Eita said.
Want more news? Listen to today's daily briefing above or go here for more info
TrueNorth’s Berg says he sees insurance companies moving toward requiring fleets to embrace AI technology and advanced telematics.
He predicted that AI-backed cameras and telematics will become almost a requirement because the added cost and risk of running without them will become too great.
As with any new technology, AI is subject to a wide range of differing regulatory pressures that vary from state to state. Some states have no guidance for private companies on AI usage and integration in insurance policies and processes, while other states are setting up limits and frameworks for implementation.
Most states that do have regulation follow the examples set by California and Colorado. Colorado began by placing limits on how AI can be used for life insurance policies, demanding that technology not be used to discriminate against people based on their race. The state is considering whether the regulations should be expanded to include other types of insurance, including property and liability.
California’s Insurance Commissioner issued Bulletin 2022-5, stating that companies using AI must ensure their algorithms and procedures avoid both conscious and unconscious bias or discrimination.
The state’s regulations must be followed when insurance companies are “marketing, rating, underwriting, processing claims, or investigating suspected fraud relating to any insurance transaction that impacts California residents, businesses and policyholders.”
The state also warns companies against using unrelated factors, such as social media usage, purchase history and other datasets to make arbitrary decisions on risk and premiums.
“I think that’s a good canary for the rest of the country,” Hub International’s Miller said.
