AI Wildfire Probability

AI Wildfire Probability provides current-conditions wildfire risk assessment rather than estimates derived from historical fire patterns. Using Earth observation, the solution derives wildfire burn probabilities and fuel hazard scores from what is actually on the ground now — fuel loads, soil moisture, vegetation health, and biomass — giving insurers risk intelligence that keeps pace with changing landscapes. This helps address the limitation of traditional CAT models, which rely on historical fire perimeters and outdated fuel maps that no longer reflect current fire regimes.

The resulting intelligence can support underwriting differentiation within broadly high-risk regions, portfolio accumulation and scenario analysis, proactive reserving, and post-event claims triage. AI Wildfire Probability is distinctive in combining monthly-or-more-frequent updates, property-level granularity, and transparent contributing-factor data within a solution focused on wildfire risk assessment.

• Current-conditions assessment: Assesses risk from present ground conditions rather than historical fire perimeters and loss patterns.
• Wildfire burn probabilities: Delivers burn probability outputs updated at least monthly, and more frequently during elevated risk periods.
• Fuel hazard scoring: Produces fuel hazard scores reflecting current vegetation and biomass conditions.
• Change detection: Identifies areas where wildfire probability or fuel hazard has shifted over time.
• Transparent contributing factors: Exposes the underlying data — soil moisture anomaly, vegetation vigor, fire weather index trends, dominant wind direction, and terrain — rather than a black-box score.
• Earth observation: Uses imagery from the EarthDaily Constellation and other sources to derive wildfire risk.
• Post-event data: Provides burn perimeter and burn severity maps scored 0 to 1 to support claims triage and IBNR estimation.
• CAT model complement: Works alongside existing CAT workflows as an independent signal, without replacing existing tools.