United StatesFinance and insurance2016 – ongoingNeptune Insurance Holdings (Neptune Flood)
Triton automated flood underwriting engine
Neptune's Triton engine underwrites flood policies in under two seconds with no human underwriters
Problem
Flood insurance underwriting involves risk selection, pricing, aggregation and carrier assignment, which are slow when done by hand.
Approach
Neptune built Triton on its own AI and machine-learning models. It handles risk selection, pricing, aggregation and carrier assignment, and Neptune runs underwriting without human underwriters. Triton also uses KatRisk APIs to overlay flood footprint data onto exposure after a natural disaster.
Results
Takeaway
Neptune runs its entire underwriting step through one model. The loss ratio it cites covers the whole business since 2016, so it does not isolate the effect of the AI.
Categories
- Size
- Small business
- Industry
- Finance and insurance
- Business function
- Finance, risk and legalProduction and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Constellation Research reports on Neptune's description of Triton, which decides in under two seconds and runs without human underwriters. It gives Neptune's 24.7% lifetime written loss ratio and compares it with the National Flood Insurance Program.
Full details are in the original sources.