Using AI to Prevent Losses Before They Happen: Andrew Engler and Kettle's Insurance Reinvention
Hosts of The Insurance Dudes Podcast. 1,000+ episodes helping insurance agents build elite agencies.
Kettle uses machine learning trained on satellite imagery, weather, and vegetation data plus funded community-level mitigation (controlled burns, brush management, defensible space) to actually lower wildfire risk, not just price it more precisely. For agents in wildfire and catastrophe states, this is the distribution opportunity replacing the carriers walking away.
Kettle prices wildfire by lowering it, not just measuring it. Machine learning trained on satellite imagery, vegetation density, weather, and wind data generates a dynamic risk score, then Kettle funds community-level mitigation (controlled burns, brush management, defensible space) that actually reduces the probability of loss. The traditional model just reprices what carriers cannot prevent. Kettle's model creates coverage access in wildfire markets that traditional carriers have abandoned.
What is Kettle actually betting on with AI and wildfire mitigation?
Andrew Engler's background combines technology, data science, and a genuine belief that insurance as an industry has been solving the wrong problem. The traditional answer to catastrophic risk is better prediction: if we can model exactly which homes will burn, we can price them accurately and make underwriting decisions that keep the book profitable. Kettle's answer is that better prediction isn't enough, the goal should be to actually make communities safer, and then price the improved risk accordingly.
This requires a fundamentally different relationship between insurer and insured. Traditional insurance companies have no incentive to invest in reducing the likelihood of a loss, they make money by pricing risk accurately, not by eliminating it. Kettle's model inverts this: by investing in actual wildfire mitigation (controlled burns, defensible space programs, brush management), they reduce the underlying risk, which reduces losses, which makes the coverage viable in markets where traditional carriers have been retreating.
The machine learning component is what makes this possible at scale. Wildfire risk is extraordinarily complex, it depends on vegetation density, moisture levels, wind patterns, proximity to other structures, and dozens of other variables that interact in non-linear ways. Traditional actuarial models, based on historical loss data, are inherently backward-looking and consistently underestimate the risk in a climate that is changing rapidly. Machine learning models trained on satellite imagery, weather data, and real-time environmental conditions can generate risk assessments that are both more accurate and more dynamic than traditional approaches.
For agents working in wildfire-prone markets. California, Colorado, Oregon, Washington, and increasingly Texas and other states, this conversation matters immediately. The carriers retreating from their markets aren't going to come back just because prices adjust. The market structure itself has to change, and Kettle's model represents one direction that change could take.
What does the Kettle model reveal about the future of catastrophic risk?
Risk mitigation and risk transfer are not mutually exclusive. The traditional insurance model treats these as separate functions. Mitigation is the property owner's responsibility. Transfer is the insurer's product. Kettle's integration of both, using insurance as a mechanism to fund and incentivize mitigation, is a genuinely new model that could apply beyond wildfire to flood, hurricane, and other catastrophic risk categories.
Machine learning changes the accuracy of risk assessment. The agents and carriers who understand how modern risk assessment tools are changing the precision of underwriting decisions will be better positioned to serve clients, explain pricing, and identify coverage opportunities than those who are still operating on actuarial frameworks built before satellite imagery and real-time data were available. This isn't abstract, it's already changing what carriers can and can't write.
Community-level risk reduction creates community-level coverage access. One of Kettle's most interesting insights is that individual homeowners have limited ability to reduce their own wildfire risk, it's a community-level problem. When Kettle invests in community-level mitigation, it creates insurance access for entire communities that would otherwise be uninsurable. This is a model for how coverage can be restored to markets where it has been withdrawn.
The climate-adjusted risk environment requires new mental models. Agents who are still explaining insurance pricing using frameworks built for a stable climate are going to find those explanations increasingly inadequate. The frequency and severity of catastrophic events is changing in ways that historical data doesn't capture, and the pricing and coverage decisions flowing from that reality are going to increasingly require agents who can explain a more complex risk picture.
Insurtech innovation creates distribution opportunities. Companies like Kettle need distribution partners who understand their product and can reach clients in high-risk markets. Agents who invest in understanding the new approaches to catastrophic risk coverage, and build relationships with the carriers and MGAs developing them, position themselves as valuable partners in markets where traditional coverage is increasingly difficult to find.
How should agents in catastrophe-exposed markets respond right now?
If you're writing policies in any of the high-risk markets, wildfire states, coastal flood zones, hurricane corridors, you need a current understanding of what carriers are actively writing, what new approaches are being developed, and how to explain the pricing reality to clients who are experiencing dramatic increases or coverage withdrawal.
Build a list of the MGAs and specialty carriers who are still active in your high-risk markets. These relationships are your most valuable assets when traditional market access is shrinking. Reach out to your wholesale partners this week and ask specifically what's available for properties that traditional carriers have non-renewed.
Then invest in understanding how modern risk assessment works. Not at the machine learning level, at the client explanation level. When a client asks why their premium doubled, can you explain specifically what changed in the risk model? That capability builds trust and prevents the client from shopping blindly.
Finally, stay current on the legislative and regulatory environment around catastrophic coverage in your state. The regulatory response to insurance market withdrawal in high-risk areas is evolving rapidly, and agents who understand the landscape can navigate clients through it more effectively than those who are surprised by changes.
What's the takeaway for agency owners in wildfire and catastrophe markets?
Andrew Engler's vision for Kettle, using technology and mitigation to make communities more insurable rather than just more precisely priced, is a bold bet on a different future for catastrophic risk insurance. Whether or not that specific model scales, the underlying insight is correct: the old model isn't working, and the agents who understand what's coming next will be better positioned than those who are still explaining 2010 insurance to 2025 clients.
Catch the full conversation:
Level up your agency:
Listen to The Insurance Dudes Podcast
Get more strategies like this on our podcast. Available on all platforms.
Related Episodes
How AI Is Solving the Wildfire Insurance Crisis: Kevin Stein and the Delos Insurance Approach
How Niching Down Your Insurance Agency Unlocks Growth That Generalism Never Can
Tony Merwin's Measured Approach to Medicare Marketing: Why Precision Beats Volume : Part 1
Tony Merwin's Medicare Systems and the Annual Enrollment Playbook That Drives Retention : Part 2
The Insurance Agent Who Never Used a Computer : How Fundamentals Still Win in Sales
