Insurance Companies Have a New Product: Not Needing Them

When a State Farm policyholder in Texas plugs a small sensor into their electrical outlet, they probably think little of it. But that device, made by Whisker Labs under a program called Ting, is part of a change in how American insurers think about their business. State Farm claims it has now placed those sensors in more than 1 million homes across 45 states, monitoring for electrical arcing that can trigger house fires. The goal is to make sure the claim never happens in the first place. Scott Holeman of the Insurance Information Institute, a nonprofit that tracks insurance industry data and trends for consumers, businesses, and policymakers, says the shift is already underway. Building the Prevention Machine "U.S. insurers are increasingly using AI to shift from a traditional 'detect and repair' model, where claims are handled after an accident, to a more proactive 'predict and prevent' approach," Holeman said to AIM Media House . Some insurers are already producing results they can measure. The Whisker Labs program is one example. A July 2025 study by Triple-I, Whisker Labs, and Octagram Analytics found the Ting sensor program reduced annual fire damage claims by $81 per home and cut non-catastrophic fire claims by an estimated 63%. Auto insurance is seeing the same logic applied through telematics. Programs like Progressive's Snapshot and GEICO's DriveEasy use mobile apps or plug-in devices to monitor speeding and hard braking behind the wheel. "Instead of simply paying claims after crashes occur, insurers are increasingly using these tools to help policyholders avoid accidents in the first place," Holeman said. The patent filings point to significant investment. A December 2025 analysis by Evident found that State Farm, USAA, and Allstate alone account for 77% of all AI patents filed by insurers over the past decade. State Farm leads with 326 filings, USAA follows with 218, and Allstate with 136. The same prevention logic is being applied across other lines of insurance. In workers' compensation, computer vision systems scan job sites in real time, flagging missing protective equipment and fall hazards before injuries occur. In agriculture, satellite imagery combined with AI monitors crop health and soil moisture, allowing insurers to identify drought stress or disease early. Property insurers are doing the same with IoT sensors, tracking water leaks, temperature changes, and electrical anomalies in commercial buildings before they become claims. For commercial fleet operators, a 2026 Verizon Connect survey of 830 fleet professionals found telematics programs delivered an 11% reduction in insurance premiums and a 19% decrease in accident-related expenses, though that figure is industry-reported rather than independently verified. What the Research Actually Shows Independent research offers a more measured picture. A March 2025 randomized controlled trial by the AAA Foundation for Traffic Safety found that drivers enrolled in telematics feedback programs reduced speeding by 11%-13% and hard braking by 16%-21%. The improvements held during the six-week period after monitoring ended, and the results were consistent across age, sex, and race. The same study flagged a persistent problem: some drivers improved while being monitored but showed evidence of backsliding once active tracking stopped. The behavioral change is real, but its durability is uneven. A separate systematic review published in Accident Analysis and Prevention identified a deeper issue. Researchers found a strong causal link between telematics variables and accident rates, but noted that most insurers rely primarily on mileage data, leaving the predictive power of the broader dataset largely untapped. "By reducing accidents, fraud, and claims severity, the main drivers of premium increases, AI can help moderate overall losses in the system," Holeman said. "In that sense, AI is helping insurers move from being reactive claims payers to proactive risk-management partners that improve safety and help control costs." The data exists. The sensors are installed. What insurers choose to do with both is still being worked out.

Insurance Companies Have a New Product: Not Needing Them

Building the Prevention Machine

Keep reading. It's free.

Related stories