Artificial intelligence is reshaping industries, but in a field like insurance, where relationships and trust are paramount, can AI actually make the experience more human?
It’s a question that weighed on Jeff Sutton (pictured), senior vice president sales and marketing at Markel Canada, as his organization explored the use of large language models and incorporating various data sets to improve decision-making, enhance our understanding of the business, and drive greater efficiency.
Like many insurance companies, Markel Canada is embracing AI to enhance, not replace, the expertise of its underwriters. According to Sutton, the company is leading with the belief that AI, particularly generative AI, can strip away the mundane and free up professionals to focus on what truly matters: judgment, relationships, and service.
“There’s a real war for talent, and as a specialty insurer, Markel Canada relies heavily on the expertise of its underwriters,” Sutton told Insurance Business. “They assess complex risks that don’t fit neatly into standard models. We want them focused on high-value tasks, making the best use of their expertise and limited bandwidth.”
The insurance industry thrives on nuance, and specialty underwriting in particular is an area where decisions require experience and professional judgment. Unlike personal lines insurance, where AI can more easily automate standardized processes, specialty insurance deals with complex risks that don’t fit neatly into algorithms.
So, where does AI fit into specialty insurance? The goal, according to Sutton, is to make underwriters more efficient, helping them reach better decisions faster without manually gathering and consolidating data. “With evolving tools, underwriters can better understand risk exposures and price accordingly for long-term sustainable growth,” he said.
This is also why Sutton underscored the importance of maintaining human relationships in an industry built on trust. He argued that AI should enhance those interactions rather than create further distance between brokers and underwriters.
“I believe Gen AI has the potential to humanize insurance by consolidating and querying information, but the critical step is still the human interaction, something that's becoming a lost art in the Canadian P&C industry,” Sutton said. “We’re so relationship-focused (but) we’ve lost the art of picking up the phone.”
While technology can provide invaluable insights, he stressed it's the personal conversations that truly strengthen relationships. Markel Canada has taken deliberate steps to reinforce this, according to Sutton, going as far as placing orange rotary phones in offices to encourage more direct conversations.
"We’re trying to break people out of the habit of emailing back and forth. Just pick up the phone and call. That’s how relationships deepen," said Sutton.
The challenge with AI in insurance is not just about implementation but perception. A new KMPG report showed that while AI adoption is accelerating in insurance, distrust for the technology remains.
The global survey revealed 46% of insurance leaders still have reservations about AI’s reliability, and only 25% fully trust AI within their companies. A majority (82%) of those surveyed acknowledge the need for robust frameworks, policies and processes to ensure regulatory compliance and responsible AI implementation.
The idea of AI making decisions on claims or underwriting policies triggers skepticism, especially when the stakes involve financial security and livelihoods. If AI can summarize data, analyze risks, and manage communications, where does that leave the human workforce?
Sutton acknowledged this quandary but sees it as an opportunity to reframe the conversation. "We have to be very careful about how we integrate AI. It’s not about replacing people. It’s about making them better," he said. "The expertise that underwriters bring to the table cannot be replicated by AI. AI can assist, it can enhance, but it will never replace the judgment and experience required to do this job well.”
One of the fundamental questions AI raises is whether it can truly make insurance feel more personal. The industry, for all its complexities, is about protecting people and businesses from uncertainty.
The danger is that AI could strip away the human element, turning insurance into a purely transactional experience. But Sutton said insurance companies can take steps to prevent this.
"What we’re seeing is that AI, when used properly, actually allows for more meaningful interactions,” he said. “If underwriters aren’t drowning in paperwork, they can spend more time talking to brokers and clients. That’s how you build trust.”
Do you agree with this perspective on AI adoption in insurance? How is your organization bridging AI use with human interaction? Please share a comment below.
