Heritage Insurance
Igniting Innovation: Mastering Wildfire Risk with ZestyAI

Whether you’re struggling to find accurate property insights or to underwrite climate risks, finding the right partner is crucial and isn’t always obvious.
ZestyAI strives to return 10X return on customer investment.
Heritage Insurance has long been a key player in providing insurance for properties in catastrophe-prone areas. As the carrier expanded to wildfire-prone states such as California, they discovered that the traditional methods of assessing wildfire risk lacked the granularity they needed for accurate risk segmentation. Heritage needed a more accurate method to measure and quantify the wildfire risk they were writing.
Recognizing the need for a more refined approach, Heritage embarked on a strategic shift to choose a new wildfire risk partner. ZestyAI's Z-FIRE delivered granular, property-level risk assessment with surgical precision, far surpassing the capabilities of traditional methods and legacy providers.
The integration of Z-FIRE into Heritage’s operations was both smooth and impactful. The underwriting team, initially accustomed to conventional methods, found the level of detail and accuracy in the data refreshing. This not only transformed Heritage’s underwriting processes but also provided a new lens through which risk was viewed and managed. The data enabled a more nuanced understanding of risks, allowing for more informed decision-making and aligning with Heritage’s risk appetite.
The adoption of Z-FIRE enhanced underwriting accuracy and improved risk management practices. This has led to more confident decision-making, particularly in underwriting properties in fire-prone areas. The impact extended to reinsurance relationships as well, with better data driving improved terms, conditions, and pricing.
In a whitepaper commissioned by ZestyAI, Milliman performed an analysis of how ZestyAI’s Z-HAIL risk score could be used to identify property risk exposure to hail. To test the effectiveness of the model, Milliman used historical Texas loss experience from an insurance company, along with data from ZestyAI’s Z-HAIL model. Z-HAIL uses a variety of climatological data and property-specific attributes to understand and measure a property’s exposure to hail risk.
Milliman's analysis showed that ZestyAI’s Z-HAIL risk score could be used to segment properties based on exposure to hail risk. Over a period of seven years, properties with a Z-HAIL score of 10 had a reported loss ratio of 50.4% compared to only 2.4% for properties with a Z-HAIL score of 1, which corresponds to a highest to lowest loss ratio lift of 21X.
In addition, Milliman found that ZestyAI’s Z-HAIL property-level risk model also improved the insurance company’s ability to segment the risk more granularly than county or zip-code based territories, recognizing the difference in risk exposure within small geographical areas.
The insurance industry faces escalating risks from the increasing frequency and intensity of severe convective storms, which caused $60 billion in insured losses last year. Traditional risk assessment methods are insufficient to manage these growing challenges.
ZestyAI’s advanced models, Z-HAIL and Z-WIND, were applied to a carrier’s book of business over a retroactive five-year period. These models provide comprehensive coverage, strong risk segmentation, and actionable insights to optimize underwriting and rating strategies.
The models achieved a 99.7% hit rate, significantly improved risk segmentation (62X and 9.7X lift for Z-HAIL and Z-WIND, respectively), and enhanced the carrier’s combined ratio by 4 points within the first year.
A legacy book consists of the accumulation, over several years, of various sub-portfolios with different risk profiles based on different underwriting frameworks. A top-10 national carrier approached ZestyAI with a legacy book filled with properties of unknown risk and a worsening loss ratio.
Without any investment in IT resources, ZestyAI and the carrier began a three-stage process: defining the scope of the review, assessing the risk of the book, and acting on the high-risk policies.
Thanks to this simple workflow, the insurer was able to isolate the riskiest properties in their legacy book, assign the appropriate action to address this risk, and assure the continued profitability of their book. The carrier improved its action rate to approximately 60% on manually reviewed properties.
Millennial Specialty Insurance is a dynamic player in the home insurance industry. As a tech-forward company, they have made significant strides in the last five years in catastrophe-exposed locations that have been traditionally underserved by national carriers.
The journey wasn't without its hurdles. Millennial Speciality started in the challenging market of Florida and soon realized the limitations of traditional risk models when expanding across the US. The standard models, suitable for Florida, fell short in accurately assessing the complexities of wildfire risks, such as in California. This gap in risk assessment posed a significant challenge, highlighting the need for an innovative solution.
Seeking a partner that could rapidly scale and match Millennial Specialty’s pace of innovation, they turned to ZestyAI. The partnership was a natural fit, given ZestyAI's expertise in providing climate risk models crucial for underwriting and rating. ZestyAI offered more than just technological solutions; they understood the unique business needs of Millennial Specialty, making them an ideal partner in tackling the intricate Californian market.
With ZestyAI's advanced risk assessment tools, Millennial Specialty could differentiate exposure risks and structural susceptibility at a granular level. This not only allowed them to navigate the Californian market more effectively but also to insure homes in wildfire-prone areas that other companies might have avoided. This strategic partnership enabled Millennial Specialty to expand its reach and offer insurance solutions to a broader range of homeowners, standing out in terms of both technology and strategy in a competitive industry.
Historically, wildfire has been challenging to model; traditional models and Fire Hazard Severity Zone maps tend to lack the granularity needed to assess risk accurately. That not only leaves insurers with major losses but also affects their relationship with re-insurers.