Simon Glairy is a veteran in the insurtech space, specializing in how artificial intelligence redefines our understanding of catastrophe risk. With a background in AI-driven risk assessment, he offers a unique perspective on the evolving landscape of severe weather events and the technological shifts required to stay ahead of them. Today, we discuss how the industry is pivoting from traditional models to high-tech, evidence-based strategies to combat escalating global losses from increasingly volatile weather patterns.
Severe convective storms have caused over $208 billion in global insured losses recently, shifting from “secondary” to primary risk drivers. How has this surge changed the way insurers prioritize wind and hail events, and what specific metrics are now most critical when evaluating these volatile weather patterns?
The shift is massive because we have seen $208 billion in global insured losses tied to these storms over just the last three years. This staggering figure has forced insurers to stop viewing hail, wind, and tornadoes as minor “secondary” perils and instead treat them as primary threats to their annual balance sheets. We are moving away from broad regional assessments to hyper-local metrics, such as the specific durability of a building’s envelope against high-velocity wind. It is no longer enough to look at historical averages when the frequency and severity of these events are breaking records every single season.
Traditional risk assessment often relies on historical norms, but AI is now being used to analyze unstructured data like satellite imagery and infrastructure age. How does moving toward an evidence-based model change the advice given to clients, and what steps are involved in validating these granular analyses?
Moving toward an evidence-based model allows us to look at the “ground truth” of a property rather than just its zip code’s history. By utilizing satellite imagery and AI, we can pinpoint the exact state of a roof or the specific age of infrastructure that might fail during a storm. This changes our advice from vague warnings to actionable steps, like recommending a specific reinforcement for a structure that the data shows is vulnerable. Validating these granular analyses involves cross-referencing AI outputs with real-world loss data to ensure our predictive models are actually reflecting the current, volatile climate reality.
Factors like population growth in hazard-prone areas and aging infrastructure are driving up loss costs alongside climate volatility. In what ways can AI identify specific vulnerabilities in construction aging or vegetative exposure, and how do you translate that data into a concrete resilience strategy for a property owner?
We are currently seeing a perfect storm where population growth into hazard-prone areas meets infrastructure that simply was not built for today’s volatility. AI excels here by scanning for “vegetative exposure,” identifying exactly how close trees are to power lines or buildings, which is a major driver of loss during wind events. For a property owner, this translates into a concrete resilience strategy, such as proactive maintenance on a 20-year-old roof identified as high-risk by our aging patterns analysis. By addressing these specific vulnerabilities—such as the degradation of natural ecosystems or construction wear—we help clients move from being reactive to being truly prepared.
While many executives believe AI-adopters will outperform their peers, a significant portion of the industry has not yet integrated these tools. What are the primary hurdles for companies that haven’t adopted AI, and can you share a scenario where AI-driven data successfully mitigated a loss traditional modeling missed?
The primary hurdle is often a “legacy mindset” where firms are hesitant to replace decades-old modeling processes with something as complex as AI. Despite 73% of executives believing AI-adopters will outperform their peers, only about one in four insurers are actually using these tools for severe convective storm risk today. I have seen scenarios where traditional models completely missed a looming loss because they did not account for recent ecosystem degradation near a new housing development. AI-driven data flagged that specific area as a high-risk zone for flash flooding and wind damage, allowing the carrier to adjust terms and provide the client with a mitigation plan months before the storm hit.
What is your forecast for the future of AI-driven risk management in the insurance industry?
My forecast for the future of AI-driven risk management is a complete integration where real-time data becomes the heartbeat of every policy. We will see a shift where organizations that do not adopt these data-driven approaches will struggle to remain solvent in an increasingly volatile climate environment. Within the next decade, I expect AI to move beyond just assessment into automated, real-time loss prevention, where sensors and satellite data trigger immediate protective actions for properties. It will no longer be an elective tool but an essential survival mechanism for the entire insurance industry to thrive.
