Recommendation & Personalisation for Insurance

Offline accuracy and revenue are different things. We measure recommendations by experiment against the business metric, because that is the only number that pays anyone.

In insurance, claims decisions need an audit trail and a consistent basis across assessors. That single fact reshapes how recommendation & personalisation has to be built here, the guardrails, the approval points and the evidence trail are design inputs rather than things bolted on before go-live.

The workload we are most often asked to take on first is policy servicing requests, usually integrated against policy administration. We build the smallest thing that proves the case, put it in front of real users, and expand only what earns its keep.

Multi-model by default, so a provider outage is a routing decision rather than an incident. You own the code, the models where they are open-weight, and the documentation to run it without us.