The US Underwriting Stack — legacy, modern, AI-native
What US carriers run to price risk, bind policies, and manage the book. The stack has a firm modern incumbent in Guidewire and Duck Creek for policy administration, a modern-to-AI-native transition in progress at the rating layer, and an AI-native entrant redefining how submissions get triaged.
Legacy
Pre-cloud workflow — people, paper, on-prem software
Modern
Cloud SaaS, classical ML, template-based automation
AI-native
Built around deep learning or LLMs from day one
| Action | Legacy | Modern | AI-native |
|---|---|---|---|
| Commercial submission intake Receiving a new-business submission from a broker or agent, triaging it against the carrier's appetite, deciding whether to quote. | Email inbox, shared drive, Excel rating sheet | ||
| Submission document extraction Turning submission documents — loss runs, schedules of values, broker slips — into structured data an underwriter can price. | Underwriter reads each attachment by hand | ||
| Rating and pricing Building and deploying the actuarial models that turn risk factors into premium. | Actuarial notebooks in SAS or R, Excel pivots, mainframe rate tables | ||
| Policy lifecycle administration The system of record: quote, bind, issue, endorse, renew, cancel. | Mainframe PAS from the 1970s-1990s | — gap in our coverage | |
| Underwriting portfolio management Looking at the book of business the carrier has written and reshaping UW decisions accordingly. | Quarterly Excel reports for portfolio committees | ||
| Distribution and product recommendations Matching customer profile to product at the point of sale — agent's desktop, bancassurance teller, embedded-insurance checkout. | Agent memory of the tariff binder |