As climate change intensifies the severity and frequency of weather-related events, insurers are investing heavily in technology to manage both risk and response. According to Carolyn Murnaghan (pictured), sustainability and climate risk leader at Deloitte Canada, the most significant breakthroughs are happening in two areas: predictive modeling and claims workflow.
On the front end, climate models are increasingly being integrated into underwriting. Insurers are using advanced analytics to gauge where concentrations of risk could strain their portfolios, and adjusting appetite accordingly.
“A lot of work is being done on the modeling side to understand where those potential future impacts as a result of climate may be, and bringing that information into underwriting so that there is a management of those risks,” Murnaghan told Insurance Business.
But the most visible impact for policyholders is unfolding during claims. When catastrophic events strike, speed and efficiency can make a material difference in recovery. “As a claim comes in, an individual is engaged with and can process their claim in particular using any type of speeding up technology, like AI to support the claims manager in that process,” she said. These tools are increasingly being deployed to streamline workflows, cut delays, and improve communication during peak claims periods.
For insurers, deploying artificial intelligence in claims isn’t simply a matter of plugging in new software. Murnaghan stressed that success depends on preparing organizations across three dimensions: people, processes, and technology.
“When we're speaking about AI-ready operations, there are lots of different things that need to be in place for AI to be used effectively,” she said. “The individual who's using it needs to understand its benefits and its limitations so that they can engage with it and use it effectively. It needs to be integrated into the flow of work so that you aren't having to go out into separate solutions in order to leverage that benefit from AI.”
She added that strong governance is just as important as technical capability. AI must be managed in a controlled and ethical way, with safeguards to ensure outputs remain within organizational expectations. That means embedding AI into day-to-day workflows, not bolting it on as an afterthought.
According to Murnaghan, her team applies this framework when assessing which insurance functions can be automated and to what extent. “It’s about making sure that all of those components – people, process, and technology – are at a sufficient maturity that AI can be used within a controlled and effective way,” she said.
Beyond claims handling, Murnaghan pointed to climate modeling as another area where AI is beginning to transform insurance. Traditional climate models are typically framed around long horizons, with projections extending 20 years or more into the future. While these outlooks are critical for understanding systemic risks, insurers also need tools that can guide underwriting and portfolio management decisions on a much shorter timeline.
“Climate models are generally considered predictive at a 20-year time horizon,” Murnaghan said. “But what we really need to understand is what does that look like in a shorter time horizon, ideally one year, which is very difficult, but five years is one where there is some predictive information.”
AI and machine learning, she explained, are enabling insurers to “downscale” long-term climate data into more actionable insights. By layering future climate projections on top of historical claims data, carriers can begin to anticipate where risks are most likely to increase within the next five years. This evolution is already shaping how Canadian insurers think about concentration risk, renewal decisions, and portfolio exposure across different regions.
Murnaghan noted that Canada’s exposure to multiple perils makes this shift even more urgent. Unlike many countries that grapple primarily with one or two dominant risks, Canada faces a broad spectrum – from hail and floods to wildfires, windstorms, and both extreme heat and cold. This variety complicates underwriting and amplifies the value of AI-driven modeling. By connecting long-term projections to short-term operational decisions, insurers are better positioned to adapt to this complex risk environment.
That shift means climate data is moving beyond the realm of high-level strategy and into day-to-day underwriting. “We’re getting into the point where that is starting to become of interest in the underwriting process,” Murnaghan said, noting that Canadian insurers are actively exploring how to integrate these AI-powered insights into routine decision-making.
