AI Helps Fleets Navigate Risk Management
[Stay on top of transportation news: Get TTNews in your inbox.]
Insurance companies and brokers are finding the benefits in artificial intelligence to better assist trucking companies with their risk management needs. The technology has helped by quickly collecting and deploying data, enhancing safety and potentially paving the way to lower insurance costs, technology and insurance experts said.
“Insurance companies are using data and AI technology on the route, on the origin, on the destination, how the driver is driving, to ultimately price insurance,” said Lisa Paul,Hub International’s chief strategy officer, specializing in transportation. She added that commercial auto insurance “has not been profitable doing it the old-fashioned way” because it was based mainly on motor vehicle reports on drivers.
“But that really didn’t tell the insurance companies how those drivers were actually driving,” Paul said, adding that motor vehicle reports are a reflection of how many times you got caught, not what you were actually doing. Hub International offers an AI-supported contract review service that can sift a 50-page agreement in PDF or Word in two minutes and automatically generate a response.
Artificial intelligence has utilized data from electronic logging devices, telematics systems and dashcams to address issues like driving behavior and problematic routes. AI is also beneficial in proposing and revising agreements with shippers.
Jeff Davis, vice president of safety for Napa River Insurance Services, the third-party administrator for Hudson Insurance Group, said AI is already proving its benefits in many areas of trucking safety, especially in dashcam systems.
AI has been an effective tool, for example when used with dashcam technology. (drKokos via Getty Images)
By enabling the processing and management of large volumes of information, AI allows some dashcam systems to distill thousands of data points down to “actionable items for which drivers can more effectively be coached,” Davis said. “These devices are no longer simply event recorders.”
The more advanced systems observe following distance, intersection behavior and speed, scoring drivers on overall habits so they may be addressed before they lead to accidents, Davis said. “I feel we will see this technology become more widespread in identifying preventive/corrective measures that can be undertaken in all areas of performance,” he predicted.
Dashcam technology with AI can be instrumental in reducing risky driving behavior, help to lower the cost of insurance premiums and exonerate drivers, insurance executives and tech vendors said.
“Video evidence takes the guesswork out of the equation when fleets need to defend drivers against fraudulent claims,” said Ingo Wiegand, vice president of product management, safety, at Samsara. “It can also speed up investigations for legitimate claims when fleets can easily pull and share HD video footage as evidence directly with their insurance provider.”
Wiegand added that AI-enabled dashcams provide fleets with insights to build a safety program and culture. “With in-cab alerts and mobile workflows, you can proactively coach drivers in the moment and work toward reducing risk across your fleet,” he said.
The implementation of AI-enhanced systems promises to yield maintenance benefits too. Paul of Hub International noted, “A vehicle that’s had multiple rapid acceleration events and speeding events is going to have a different maintenance requirement” than a vehicle that’s being driven by a more cautious hand, belonging to a driver who has received remedial training.
Davis of Napa River said that for any industry or end user, success with AI is dependent on the quality of data. “It directs the assumptions of any AI-based system,” he said. “This is why there must be constant and ongoing validation of the results from any AI process.”
Chuck Wallace, CEO of High Definition Vehicle Insurance (HDVI), said that his company works with a large set of data sources, including telematics systems, ELDs and camera systems. Data from all those sources “comes in different formats at different intervals,” he said. HDVI built its own technology to field the data, analyze and standardize it, and to cope “when there’s a little blurb or a bobble in the data which there inevitably will be — and then fix that” to maintain high-quality data.
“Different telematics vendors spin off data of varying levels of quality,” observed Keith Halasy, vice president of marketing for HDVI. “From a truck operating standpoint, it’s important to make technology selections that support” AI-assisted systems, Halasy said. Some fleets lack the tech resources to put together an AI program, which must be fed by large volumes of data, he noted. “But when you start getting into the pools of some of the leading telematics providers [that] have a substantial amount of data, you can start to take advantage of that.”
Hub International operates a technology portal that can “ingest” data from more than 100 different ELDs, video information from dashcams, and route optimization data to help clients assess their total cost of risk “for that mile for that driver, for that shipper,” Paul said.
The more clearly total cost of risk-per-mile is defined, and then combined with equipment and fuel costs, the better a trucking company can price its services, Paul noted. “But they also then can be more predictive in their own expected losses and premiums,” she said.
Data Use and Storage
Keeping data secure and managing it ethically are not only business priorities, experts noted.
(peshkov via Getty Images)
“It all boils down to, legally, who owns data,” said Hub’s Paul.
A trucking company owns the data in an ELD, Paul said. Beyond that, trucking companies seeking to use driver behavior data need to ask drivers to sign a telematics user agreement, as they do for motor vehicle reports and background checks, Paul advised.
Carriers can encounter reluctance or refusal from independent owner-operators when it comes to sharing data. “It can present a problem, but it can present an opportunity,” Paul said.
The independent contractor owns their driver score. A carrier can win over a driver by explaining that the score will be used to leverage lower insurance costs, which will be passed on to the contracted driver.
For a transportation company delivering data to an insurance company, security, including defending against hacks, is a consideration, Paul and others said. “I would say that’s a new and emerging issue for transportation companies,” Paul said. She also pointed to inward-facing dashcams that might capture biometric facial data. “You’ve got to deal with all the state laws associated with that,” she said.
Davis of Napa River called AI a tool to be used in decision-making. “It should not be considered a replacement for the human thought process when it comes to ethical considerations,” he said.
Data privacy should be one of the first steps in implementing “any AI program,” Davis added. “There will be those that actively attempt to breach this data and AI systems in general to gain knowledge into business practices, etc.” He recommended that a data security specialist evaluate any AI project or vendor agreement ahead of time.
AI and video telematics play an important role by giving fleets insights and evidence to better inform rates and protect against false claims, Samsara’s Wiegand said. “I see a lot of potential for driver coaching at scale and proactive training on patterns of risky behavior. Getting ahead of these behaviors before they affect insurance rates is where fleets will be able to feel a tangible impact.”
Wiegand said that choosing “the right AI-enabled dash camera is the best way to lower insurance premiums.” He said certain insurance providers look for specific dashcam features, such as audio speakers, recording capabilities or a wide view for outward-facing cameras. He noted that there are “more nuanced features” such as proximity search and on-demand video retrieval. These can help fleets pinpoint a vehicle’s location and retrieve HD video footage within minutes, he said. “For some customers, this results in hundreds of thousands of dollars saved on claims by exonerating drivers with video footage,” Wiegand said.
“As insurance premiums continue to rise, it’s become increasingly important to take a data-driven approach to risk management,” Wiegand said.
Some insurance companies are changing their approach to pricing, focusing on “how many miles and what routes and what type of driver in order to better price and achieve better profitability,” said Paul of Hub International.
Greater flexibility in rates during the span of an insurance contract is possible with AI-processed data, insurance executives said.
(peshkov via Getty Images)
If a fleet improves its safety performance score, some insurance companies now offer to decrease what they charge, Paul said. “The challenge is most trucking companies don’t have a feel for how their fleet is going to score,” and a poor score could cost them.
HDVI offers a “Safety Lookback,” analyzing fleets’ telematics data from the past 90 days when a policy is first quoted. If the data from that span shows a certain level of safety performance, a fleet can earn discounts from the beginning of the policy, the company said.
Todd Witte, vice president of insurance product for HDVI, said the percentage of customers willing to share data before buying insurance has been rising. “It’s above 50 percent,” he said.
“If you think about the old days,” Witte said, a trucking company seeking coverage might tell an insurance company about its trucks, provide driver MVRs and describe its operations area — “pretty simple stuff.”
With AI, Witte said, instead of yearly evaluation of risks, there can be monthly monitoring, enabling the insurance company to be more responsive to a carrier’s driver behavior and the fleet’s safety performance. “This data is able to paint a much more high-definition picture of the risk,” Witte said.
Another use of AI can be to counter a negative reaction from underwriters after a carrier has a major accident. Underwriters typically conclude that the carrier is a high risk, but Paul said, “Maybe the transportation carrier was just unlucky.”
AI-generated data might show that the driver involved in the loss scored high in driving performance, had no record of distracted driving, and also that the overall fleet score is above average. That information can be used to reassure an underwriter, Paul said.
“Underwriters have historically used FMCSA violation data to price insurance,” Paul said, but with driver turnover, operational changes and a different customer mix, driver and fleet safety performance changes, Paul pointed out. “Those violations that happened three years ago and two years ago may not be indicative of how that company is operating now.”
Underwriters need not use “stagnant” Federal Motor Carrier Safety Administration data “as the sole determinant in pricing insurance,” Paul said. “The AI data is fresh, it’s current.”
Want more news? Listen to today's daily briefing below or go here for more info: