illustration embodying AI

Prepare for wildfires by leveraging AI.

Predict the likelihood of a wildfire occurring at any given location over the next year more accurately, cheaply, and quickly with Terrafuse AI.

✔ Get a more granular wildfire risk score (over 250 possible scores) ranging from 1 to 10.

✔ Available in 30 square meter spatial resolution.

✔ View risk at a single location or run a portfolio analysis.

✔ Make mitigation decisions with a rank-sorted list of 27 factors contributing to risk.

✔ Integrate easy-to-use REST API with your preferred software to deliver geospatial visualizations or tabular data.

    30+ Trillion 

 data points processed

~33%

better performance than the incumbent*

4

patents

Visual_Map + Risk Graph_Risk (4).png
* performance improvement based on the Gini coefficient, which measures a model's predictive power.

INSURANCE APPLICATION

Our Wildfire AI model is validated against a portfolio with a total insured value worth over $400B and historical wildfire claims data.

magnifying-glass-icon-13.png
Make more accurate underwriting decisions by leveraging AI and climate science
coin.png

Price insurance according to actual risk

fire-extinguisher.png

Mitigate potential wildfire risks in your portfolio 

CASE STUDY

Wildfire Risk Assessment for Insurance Company in CA

Terrafuse's Wildfire AI model key results:
 

Identified ~10% gross margin improvement for portfolio
(equivalent to ~$80M for a portfolio with $1B in premiums and $135M in losses).

 

Would have prevented 71% of losses for a portfolio had it been used
(in other words, ~$96M for a portfolio with $135M in losses). This is equivalent to a ~5% (or $6.6M) improvement compared to the incumbent.

 

Achieved a lift** of 113 for a top 10 primary insurance company.

Visual_Chart_Bret feedback (1).png
* The x and y axes shown in the graphs above are of the same exact scale for the Incumbent and Terrafuse AI results.
**Lift is defined as the loss ratio of 90-100 risk percentile group divided by loss ratio of 0-10 risk percentile group. Lift measures the ability of the wildfire risk score to segment the best from the worst risks. The higher the lift, the better the model differentiates very high risk from very low risk. Typical insurance models have lifts in the single digits.

Average acreage burn has more than doubled since the 1990s due to climate change-induced wildfires, causing an estimated $90B in losses in the U.S. during the 2021 wildfire season alone. As a result, insurance companies defaulted to either blanket pricing of policies or refused to insure properties altogether, particularly in California, which drastically reduced revenue. 

The global protection gap, the difference between global economic losses and what's covered by insurers to help society adapt to natural disasters, reached an unprecedented $161 billion in 2019, or 69% of all economic loss from climate hazards.

 

Our Wildfire AI risk scores allow insurance companies to underwrite more businesses and reduce non-renewals to fewer properties than any other model - all of those which will actually burn.

Are you an underwriter interested in avoiding wildfire claims and writing more business?