Many recent deployments of AI are proving their worth in reducing costs and claims. Nationwide and Whisker Labs' Ting devices proactively prevent electrical fires; telematics data is helping truck fleets cut collision costs; and voice AI is providing faster responses to policyholders.
Read more of Digital Insurance's recent coverage of AI, and click the links below each item for the full story.
AI-powered sensor helps prevent electrical fires for Nationwide policyholders
Nationwide is working with Whisker Labs, which claims its Ting sensor program has prevented more than 27,000 potential electrical fires across the U.S. Insurers pursuing similar predict-and-prevent partnerships should structure agreements around activation and outcomes — not hardware distribution — to ensure sensors are actively identifying hazards. Carriers can leverage engagement data, activation rates and confirmed "saves" to refine underwriting eligibility, pricing and customer incentives. Ting, which plugs into any outlet and uses AI to detect electrical arcing, recently expanded to frozen-pipe alerts, addressing another major loss category. Nationwide distributes the sensor free to eligible policyholders, offsetting costs through projected loss reduction.
Read more:
AI telematics cuts collision costs 19%, survey finds
Insurers underwriting commercial auto and fleet policies should factor AI-driven telematics data into pricing and loss models: SambaSafety's 2026 Driver Risk Report found AI-enabled systems produce a 19% average decrease in collision costs across fleets. The report, drawn from 28 million driving events logged October 2025 to March 2026, shows 84% of fleets rate telematics as very or extremely important to safety.
Read more:
Transparency key as 68% of consumers open to voice AI
Insurers deploying AI voice assistants must prioritize transparency and accuracy to meet rising consumer expectations, according to Sonant's Insurance AI Report. Of 1,000-plus consumers surveyed, 68% are willing to use voice AI for faster responses, and nearly 80% are open to AI-powered voice assistants. However, 64% cite incorrect information as their top concern, followed by data privacy at 63%. Consumers are most receptive when AI identifies itself as non-human and is positioned as after-hours support.
Read more:
How insurers can design AI to keep humans in the loop
McKinsey research shows AI could automate 57% of U.S. work hours, but fewer than one-third of companies have scaled it durably across the enterprise — a gap insurers should heed before overhauling underwriting workflows. The practical model: use AI to handle intake, document classification and guideline checks, while routing uncertain, conflicting or reputationally sensitive decisions to humans by design. Carriers and MGAs should also move underwriting guidelines from static PDFs to testable, version-controlled frameworks that allow outcome modeling before deployment. The strategic priority is not maximum automation — it is identifying which underwriting decisions require contextual judgment or ethical accountability, then building systems that protect and amplify those functions.
Read more:
40% of adults fear walking alone; insurers eye safety benefits
With 40% of U.S. adults afraid to walk alone at night near their homes, according to Gallup, employers are expanding voluntary benefits to include 24/7 personal security platforms — a shift with direct implications for insurers. As workforce fear rises, insurers face potential pressure on medical claims, productivity-linked losses and employee turnover costs.
Read more:
How insurers can break AI out of pilot mode
Fewer than 15% of insurers have fully integrated AI into their financial and operational cores, creating compounding margin erosion as settlement cycles stretch beyond two months and budgets drain on manual error correction. To close the gap, carriers must take three concrete steps: build a unified data fabric that connects claims, underwriting and policy administration systems; adopt agentic AI workflows that move beyond rigid robotic-process automation to context-aware automation; and deploy low-code platforms that let business-side staff iterate on risk models in weeks rather than months. Carriers that delay rebuilding their data infrastructure now risk falling permanently behind competitors already advancing toward agentic workflows.
Read more:
Travel insurance gaps leave insurers, policyholders exposed
Travel insurance sales reached $5.56 billion in 2024, up 46% from 2019, but widespread claim denials tied to war, airspace closures and government disruptions are fueling consumer disputes and reputational risk for carriers. Insurers should audit exclusion language — particularly "acts of war" clauses — for clarity and enforceability as geopolitical volatility increases exposure to litigation and regulatory scrutiny. Compliance officers should review how "blanket exclusions" are disclosed at point of sale, given the documented gap between consumer expectations and actual coverage.
Read more:
Legacy modernization risk falls as AI cuts cost, complexity
Insurers clinging to legacy core systems — many dating to the 1970s and '80s — face compounding operational and regulatory exposure as institutional knowledge walks out the door with retiring staff. Rather than pursuing costly full-platform replacements, which often duplicate systems and balloon transformation budgets, carriers should prioritize cloud-based, in-place modernization to preserve embedded business logic while improving scalability. AI has materially reduced modernization timelines and investment levels.
Read more:
47% of customers buy insurance online
Nearly half of home and auto insurance customers now purchase policies digitally, yet satisfaction scores are falling, according to JD Power's 2026 U.S. Insurance Digital Experience Study. Service segment satisfaction dropped four points to 695 out of 1,000; shopping satisfaction fell 12 points to 523. Insurers should prioritize comparison tools; customers were nearly twice as likely to consider a purchase when such tools were available (39% vs. 12%). Chatbots and virtual assistants show clear ROI: users scored satisfaction 132 points higher than non-users, yet only 11% of shoppers encountered them. Closing that deployment gap is the most immediate opportunity for conversion improvement.
Read more:
AI claims imagery hits only 80% accuracy, experts warn
Insurers relying on AI to assess aerial and satellite damage imagery should pair those tools with skilled human analysts — AI alone achieves only 80% accuracy, meaning 1 in 5 determinations may be wrong, according to McKenzie Intelligence Services' David Heathcote. Common AI misreads include flagging baseball diamonds as destroyed structures, misclassifying cars as buildings and missing wildfire ash or floodwater in swimming pools. Synthetic aperture radar can resolve cloud and smoke obstruction, but still requires expert review. Heathcote recommends a hybrid model: AI for speed and data volume, human analysts for complex or ambiguous cases.
Read more:
Insurers lack AI talent as adaptation concerns surge
A new EY survey of 106 insurance company executives found a 19-percentage-point jump in respondents concerned about staff ability to adapt to a changing risk environment. More than 25% of firms are still evaluating AI's impact — a figure EY's Stu Doyle called surprising. Insurers cite improving productivity without adding headcount, hiring AI specialists and measuring three-year impact as top concerns. Routine risk management and data-science roles face the greatest AI-driven disruption.
Read more:
This roundup was created with AI assistance. A Digital Insurance editor reviewed each item before publication.
