Skip to content

AI Is Changing What’s Possible in Commercial Auto Insurance

Arissa Dimond

AI in commercial insurance refers to the use of machine learning, predictive analytics, and continuous data monitoring to improve underwriting accuracy, automate risk assessment, and intervene before claims occur; shifting the industry from reactive loss management to proactive risk prevention.

By the numbers:

Commercial auto insurers are under real pressure. Claims severity has surged 36% since 2020. Nuclear verdicts reached a median of $23.8 million. And despite 54 consecutive quarters of premium increases, underwriters still faced $4.9 billion in losses in 2024.

The industry has been asking the same question for years: how do we get ahead of losses rather than chase them?

Artificial intelligence is the most credible answer the market has produced. Not because it's new technology; insurers have experimented with AI for years, but because the data infrastructure to support it is finally in place. Telematics. Continuous monitoring. Normalized behavioral data from millions of commercial drivers. AI now has something meaningful to work with.

The debate has shifted from whether AI will revolutionize commercial insurance to identifying which insurers will harness its power to gain a competitive edge, and which will be left struggling to catch up for the next decade.

The Shift AI in Auto Insurance Is Already Happening

The numbers alone signal how seriously the industry is moving. The global AI in insurance market is projected to grow from $10.8 billion in 2025 to $176.6 billion by 2035—a compound annual growth rate of more than 32%. Within auto insurance specifically, growth forecasts range from 8% to 15.5% CAGR through 2035, driven by behavioral risk scoring, telematics adoption, and automated underwriting platforms.

Adoption reflects that momentum. According to the NAIC’s Big Data and AI Working Group, 88% of auto insurers report they use, plan to use, or are actively exploring AI models in their operations. In a recent survey of insurance leaders, AI was cited as the top technology priority for 2025 by 36% of respondents—ahead of big data, analytics, and cloud infrastructure.

Commercial vehicles are among the fastest-growing segments driving this investment. Fleet telematics systems generate a continuous stream of behavioral data, such as speeding events, harsh braking, and distraction patterns, that AI is uniquely equipped to process and act on. The infrastructure exists. What’s changing is how effectively insurers are connecting to it.

What AI in Commercial Insurance Risk Assessment Does

It’s worth being specific, because “AI in insurance” gets applied to everything from chatbots to underwriting engines. In commercial auto, the applications producing real outcomes fall into a few distinct areas.

  • Risk assessment and underwriting. AI models analyze continuous streams of MVR data, telematics behavior, and CSA safety scores to produce dynamic driver risk profiles, not point-in-time snapshots taken at renewal. This means underwriters can see how a fleet’s risk is evolving throughout the policy term, not just what it looked like when the policy was written. The result: more accurate pricing, better risk segmentation, and less adverse selection.

  • Loss control at scale. Manual loss control is constrained by headcount and geography. AI changes the math. Continuous monitoring across an entire book of business lets risk control teams identify which accounts are trending in the wrong direction and prioritize interventions accordingly, without waiting for a claim to prompt action. SambaSafety’s 2025 Telematics Report found that fleets combining continuous monitoring with targeted training reduce violations by an average of 77% within 12 months.

  • Targeted driver training. Generic safety training has a limited shelf life. AI-powered training workflows match specific behavior events to relevant course content, so a driver flagged for distraction gets distraction-specific training, not a broad defensive driving refresher. The intervention is timely, relevant, and measurable.

  • Fraud detection. AI is reducing insurers’ risk of fraudulent claims and cyberattacks by up to 70%, according to Technavio. The same technology that enables better risk assessment also surfaces anomalies, including AI-generated fraudulent accident images and voice cloning attempts, that would be nearly impossible to detect manually at scale.

The Execution Gap Is the Real Problem

Insurers widely recognize the value of AI in insurance risk assessment. The adoption challenge isn’t awareness—it’s execution.

“Fleet operators are flooded with data and can’t make sense of it,” said Matt Scheuing, CEO of SambaSafety. “More data doesn’t always equal the answer.”

The same is true for insurers. Telematics data arrives from dozens of providers in inconsistent formats. MVR data has state-by-state variation. Claims data lives in separate systems. Without a way to normalize and connect these inputs, AI models have nothing reliable to work with—and the potential value of all that behavioral data goes unrealized.

This is the execution gap: the distance between having data and being able to act on it. Only 25% of commercial insurers describe themselves as fully capable of handling large amounts of telematics data, while more than a third acknowledge their infrastructure needs meaningful enhancement. The insurers closing that gap aren’t necessarily the ones with the largest internal technology teams. They’re the ones building the right data partnerships.

How Insurers Can Close Commercial Driver Risk with AI

The commercial auto market isn’t short on data. What it’s short on is data that’s clean, connected, and actionable when risk control teams actually need it.

That’s the problem the SambaSafety platform is designed to solve. The platform pulls from more than 3,000 sources — MVR, CSA, and telematics from 100+ providers — and normalizes them into a single driver risk profile. Not because aggregation is technically impressive, but because fragmented data is the specific reason AI-driven risk programs stall before they scale. You can’t build a reliable risk model on inputs that vary by state, provider, and format.

The AI layer sits on top of that foundation. When a behavior event occurs, it matches the event type and vehicle category to relevant training content. It closes the loop: alert, intervention, completion, outcome. Portfolio Insights gives risk control teams a book-wide view of which accounts are improving and which need attention, so limited resources go where they’ll move the needle.

“Our job is to separate the signal from the noise,” said Matt Scheuing, CEO of SambaSafety. “We distill all of that data into one simple driver risk profile that’s clear, actionable, and predictive.”

Fleets using this approach reduce violations by an average of 77% within 12 months. A multi-year study found consistent platform use led to a 22% reduction in claims frequency and a 50% reduction in claims involving bodily injury.

Getting the data foundation right is only part of the equation. As AI takes on a more consequential role in underwriting and risk control decisions, regulators are paying close attention to how those decisions are made and by whom.

AI Governance: The Part Insurers Can’t Skip 

AI capabilities and AI governance must evolve in tandem. The NAIC’s Big Data and AI Working Group is proactively developing robust model laws to address the use of AI in the insurance sector, focusing on transparency in underwriting decisions, enforcement of fairness standards, and accountability for third-party AI systems.

Insurers that take charge of building their governance structures now will be well-prepared for the regulatory landscape ahead. Additionally, the escalating threat of fraud demands urgent action. With the rise of generative AI, malicious actors can effortlessly create false claims and impersonate policyholders. Insurers that fail to invest in AI detection capabilities are not just passively waiting; they are significantly increasing their risk exposure and endangering their bottom line.

The Window for Early Advantage Is Narrowing

Insurers already deploying AI-powered underwriting, risk scoring, and risk control workflows are building data advantages that compound over time. Models trained on years of behavioral data become more accurate. Risk control programs with established feedback loops produce better outcomes. The gap between early movers and late adopters will widen, not close.

That doesn’t require building everything from scratch. The most effective approach pairs a clear internal strategy with the right external data partnerships—accessing proven infrastructure and AI capabilities without the time and cost of building in-house.

“We’re at a pivotal moment where we can move from reactive risk management to proactive risk prevention,” said Scheuing. “Risk is not random. If insurers and their policyholders consistently use data and risk tools, they're likelier to see better outcomes.”

The commercial auto market is grappling with a significant profitability challenge that mere premium hikes won’t rectify. To truly address this issue from the ground up, the industry must leverage AI-driven risk intelligence. By integrating this technology with the right data and workflows, we can begin to turn the tide and improve profitability.

Survey

You may also like...

View all