AI in Insurance
AI Could Break the Gartner Hype Cycle | Insurance Innovation Reporter
AI is moving so quickly that it may upend the traditional hype cycle, with carriers under pressure to balance innovation, regulation, and customer trust.
In today’s technological boom, it is increasingly difficult to separate hype from reality. With so many breakthroughs in medicine, space exploration and even autonomous driving it is not advisable to bet against even the boldest of efforts. So, perhaps humans will indeed live on Mars one day. In recent times the naysaying against hype is generally about the when it will happen lesser about the if it will happen.
The property/casualty insurance space has seen its share of hyped concepts like blockchain and virtual reality but none as much as the recent exuberance for AI. And for good reason. Insurance practices revolve around information gathering and validation—whether in underwriting, claims, or risk management.
Actuarial mathematic sciences are applied for trending and pricing. And a host of other internal functions and external processes are just a few ways to describe insurance at-a-glance. All of which may benefit from AI tools and agents near-term and into the future. Insurance talent shortages, high costs of insurance for consumers and businesses, changing risks, demand for loss prediction and prevention are just some of the bigger challenges in which AI may come to the rescue.
by Stephen Applebaum and Alan Demers, as featured in Insurance Innovation Reporter
AI Model Insurance: The Emerging Market for Algorithmic Risk Transfer - FourWeekMBA
AI Model Insurance emerges as insurers develop new products to cover hallucination liability, bias lawsuits, and performance failures in AI systems.
AI Model Insurance represents a rapidly emerging market where traditional risk transfer mechanisms adapt to cover the unique uncertainties of artificial intelligence systems—from hallucination liability and bias lawsuits to performance guarantees and regulatory compliance, creating entirely new actuarial challenges and insurance products.
As AI systems move from experimental tools to critical business infrastructure, the risks they create demand sophisticated insurance solutions. Unlike traditional software that fails predictably, AI systems fail in novel ways—hallucinating facts, exhibiting unexpected biases, or degrading performance mysteriously. This unpredictability transforms how we think about technology insurance and creates opportunities for innovative risk transfer mechanisms.
Insurance Leads AI Adoption. It’s Time to Scale | BCG
Despite the insurance industry’s advantages, most companies haven’t fully realized AI’s potential. Those who show commitment, upskill their workforces, and boldly innovate will lead.
Key Takeaways
- Despite the insurance industry’s advantages, most companies haven’t fully realized AI’s potential. Those who show commitment, upskill their workforces, and boldly innovate will lead.
- This industry is particularly well suited to AI because of its deep data reserves, experience with analytic decision making, staff proficiency, and recent gains in productivity.
- Companies begin enthusiastically with pilot projects, but only 7% successfully bring their efforts to scale.
- Overcoming the challenges depends on fostering a company culture that accepts change and accountability and that allows people to work across internal silos.
The insurance industry has outpaced other sectors in its early embrace of artificial intelligence.
Insurance Leads AI Revolution But Struggles to Scale Beyond Pilots - Risk & Insurance
Only 7% of carriers have successfully implemented AI across their organizations despite industry advantages, BCG research finds.
While the insurance industry has emerged as a leader in artificial intelligence adoption, matching technology and telecommunications sectors, only 7% of insurance carriers have successfully scaled AI initiatives beyond pilot programs, according to BCG’s Build for the Future 2024 Global Study.
The insurance sector possesses unique advantages that position it for AI success, BCG said. Deep data reserves spanning decades of customer behavior, extensive experience with analytical decision-making, and a workforce already proficient in data analysis create ideal conditions for AI implementation. Recent deployments demonstrate this potential, with leading firms achieving productivity gains exceeding 30% when equipping service and operations employees with AI-powered knowledge assistants, according to the report.
One large insurance carrier processing 50,000 daily claims communications now uses customized GPT models to draft customer messages, with human agents reviewing outputs for accuracy, the report noted. By deploying this technology across all customer interactions, the company has enhanced both brand consistency and operational efficiency.
The industry’s AI adoption journey reveals three distinct categories of companies:
- Locked into Pilots: Focused on isolated use cases with annual investments under $5 million.
- Broad Experimentation at Limited Scale: Investing $5 million to $25 million annually across fragmented initiatives.
- Strategic Leaders: Investing $25 million or more—often reaching $50 million to $100 million—and use AI to redesign entire business processes.
New Economist Impact report finds AI is reshaping insurance – gradually | SAS
Global industry leaders foresee a workplace where humans, AI agents collaborate
Insurance leaders are beginning to see artificial intelligence (AI) move from hype to real-world impact, according to a new report from Economist Impact and sponsored by SAS. The research finds that emerging technologies such as generative AI haven’t yet transformed the industry, but they are demonstrating gains in productivity.
And insurance leaders see a major role in particular for agentic AI – autonomous systems capable of performing even complex tasks almost independently – suggesting that the insurance firm of the near future might feature a hybrid workforce in which human agents and AI agents work hand-in-hand on tasks such as underwriting, product development, claims processing and more.
“Insurers’ workforces will become hybrids of human employees and agents collaborating closely, with some agents working largely independently under human oversight,” said Jodie Wallis, Global Chief Analytics Officer at Manulife, one of the executives whose expertise shaped the report’s findings.
Gen AI Sprouts Ears and a Mouth | Insurance Thought Leadership
A new generation of Apple AirPods enables AI applications that will improve insurers' customer service.
Perhaps the most convoluted conversation I've ever had occurred when my wife and I visited a tiny winery in Tuscany in the late 1990s. The winery was tucked into the stone city wall of Montepulciano, on a street barely wide enough for pedestrians, let alone our rental car. There was no sign above the tiny door, just the street number a sommelier had written down for us. An elderly man answered the door... but spoke no English, while we spoke no Italian.
We experimented and found he understood Spanish, which both my wife and I spoke because we had recently lived in Mexico. In addition, he spoke French, which I still mostly understood from three years in Brussels. So Kim and I would ask a question in Spanish, which he translated in his head into Italian. He responded in French, which, after some fumbling, I'd translate into English. And away we went — English to Spanish to Italian to French and back to English. We learned he was a sixth-generation wine maker, heard about and tasted his wines and purchased two dozen wonderful bottles.
An announcement last week from Apple will let us use AI to cut right to the chase: translating from any language to any other language in real time and via voice, not just text. The new AirPods won't guarantee you the wonderful morning that Kim and I spent with the charming Italian winemaker but will help the insurance industry with customer service.
The headline of the New York Times review of the Apple announcement pretty much says it all: "The New AirPods Can Translate Languages in Your Ears. This Is Profound."
Paul Carroll, editor-in-chief, Insurance Thought Leadership
Financial Results
US P&C insurers boost profitability in H1'25 despite cat losses: Moody's - Reinsurance News
Moody’s Ratings, the credit rating agency and research firm, reported that US property and casualty insurers delivered stronger profitability in the first half of 2025, even as California wildfires in the first quarter drove substantial catastrophe losses.
According to Moody’s, a representative sample of 20 rated insurers posted net income of $27.0 billion for the period, a 13% increase compared with the same timeframe in 2024. The improvement was fuelled by rising personal auto results and robust investment income, which more than offset heavier catastrophe claims.
Moody’s noted that catastrophe losses reached $16.1 billion for these insurers in the first half of 2025, up 35% from a year earlier and representing the largest first-half loss burden the group has experienced in many years.
While wildfire activity was the main driver in the first quarter, the second quarter saw relatively lighter catastrophe costs. Moody’s also pointed out that reserve releases, particularly in personal auto and workers’ compensation, provided further support to overall earnings.
Commentary/Opinion
Vehicle Literacy: Let’s Talk About Cars | Insurance Thought Leadership
Despite having VIN data, auto insurance companies remain surprisingly illiterate about the cars they insure. There's no excuse.
Sometimes a picture says it all. The stock photo above shows 14 individuals working on the same vehicle and focusing on separate things. None of them knows everything the others have seen. None knows everything about the car. None necessarily documents anything they just did. They just complete a task — none talking to each other.
Sounds like every auto insurance company I have visited in the last 25 years — no exceptions.
Marketing, advertising, agency, distribution, acquisition, rate/quote/bind, service, claims, etc.: We all know those are about a car, just not exactly which car and what's inside it.
Back to the picture above: No one even counted #15, the person in the driver's seat.
That's what I'm talking about: We ignore our customer.
Worse, we blame them for not knowing their car, and we ask them car questions all the time.
It's not like we don't ignore everyone else in the insurance value chain, but not knowing the customer's car when we talk about their car with the customer seems like an epic ball drop.
We ask them all sorts of things about their car that we could just look up ourselves most of the time. Not because we are mean but because we lack basic information to communicate in a literate way about specific car features and values across our many insurance transactions.
Telematics, Driving & Insurance
Catena: $5 Million Raised For Its API For Telematics And Logistics Data
Catena announced it has raised $5 million in a seed funding round. This round was led by Floating Point and saw participation from several noteworthy investors, including Shaper Capital, Teamworthy, Plug and Play, SpringTime Ventures, Liquid 2, Blue Moon, and Blue Impact Supply Chain Ventures.
In addition to these investors, more than 20 prominent logistics executives from companies such as Augment, Fillogic, FreightWaves, OTR Transportation, Reliance Partners, TalentSolvers, and TruckSmarter have also joined in the investment. With this funding, Catena has raised a total of $8 million since its inception in January 2024, positioning itself as a frontrunner in freight data infrastructure.
The company was founded by three industry experts who bring a wealth of knowledge and experience from various sectors, including finance and technology. Jeremy Baksht, who has held senior positions at Walmart, Bloomberg, and Citigroup, leads the team alongside Travis May, a co-founder and former CEO of Datavant and LiveRamp, who currently serves as CEO of Shaper Capital. The third co-founder, Mike Goynes, previously served as the data officer at Interos and Optoro. Together, their extensive backgrounds make them exceptionally well-suited to develop a robust data layer tailored for the freight industry.
Arity launches Geosight for ZIP-level driving data
Arity has introduced Geosight, a new solution offering insurers ZIP code-level driving behavior data to refine territorial pricing. The dataset, derived from the driving activity of 15% of U.S. drivers and covering 96% of populated ZIP codes, is refreshed monthly and delivered in CSV format for immediate use.
Geosight captures metrics such as speeding, phone use, and braking patterns, moving beyond static geographic models to provide a behavior-based view of risk. By enabling earlier trend detection and more credible pricing assumptions in low-volume areas, Arity positions Geosight as a plug-and-play tool for insurers to sharpen segmentation, support filings, and respond more quickly to market shifts.
Claims
Aon Launches Specialist Claims Solutions Aimed at Reducing Claims Leakage
Aon plc launched Specialist Claims Solutions, an end-to-end service across the claims process, including leakage diagnostics, claims administration/run-off services, debt recovery and commutations and inwards claims handling support.
Aon has also formed a collaboration with ARMStrong Insurance Services, a provider of claims audit and recovery services to the property/casualty industry.
Veronica Judice and David Griffiths will lead the newly formed team as global co-heads of specialty claims solutions. They will both report into Siccardi.
Protection Gaps
Insurers excluding coverage in shifting environment
Coverage gaps exposed amid persistent challenges in how domestic terrorism is defined and insured
Recent acts of violence in the United States have underscored persistent challenges in how domestic terrorism is defined and insured. As businesses face mounting risks and legal uncertainties, industry leaders are examining the evolving landscape of coverage and liability.
Chris Kirby, president of Starwind Specialty’s political violence & terrorism (PVT) program, is closely watching the evolving risks facing US businesses as incidents of violence with ideological motivations continue to make headlines.
“While these events meet the criteria for domestic terrorism, the US federal government currently lacks a criminal statute to formally charge individuals with domestic terrorism,” Kirby said.
Kirby pointed to the 1995 Oklahoma City bombing as a foundational example of the legal ambiguity surrounding domestic terrorism in the United States. “Despite the scale of the attack, McVeigh was not charged with terrorism but with criminal offenses, including the use of a weapon of mass destruction and multiple counts of murder,” Kirby said.
He noted that the threat has not diminished since then. “Although that event occurred nearly 30 years ago, the threat has only intensified. Between 2001 and 2021, more than 540 domestic terrorism incidents were reported in the US, with federal investigations into such activity increasing by 357% over the past decade,” Kirby said.