Managing General Agent (MGA)
Real-Time Underwriting Data
A real-time operational data store that gave an MGA a live view of its book at the point of underwriting decisions.
An MGA was underwriting against yesterday’s data. Position reporting ran overnight, so by the time an underwriter could see aggregate exposure, capacity consumption or binding trends, the picture was already stale, and any question that crossed systems meant a request to the data team and a wait.
What we built
A real-time operational data store sitting alongside the transactional systems, fed continuously rather than batched.
- Streaming ingestion from policy administration, binding and claims sources
- A model designed for underwriting questions: exposure, capacity, aggregation, performance by programme and producer
- Serving layer supporting both interactive queries and the screens underwriters already used
- Governed and lineage-tracked, so figures could be defended to carriers and capacity providers
Why it mattered
Underwriting is a sequence of decisions made under time pressure with incomplete information. Narrowing the gap between what has happened and what the underwriter can see directly improves decision quality, and removes the data-team bottleneck from routine questions.
The same governed store now underpins the MGA’s reporting to its capacity providers, so operational and external numbers reconcile by construction.