Home / Case studies / Predictive models
$754M of underwriting & reserve accuracy
Global life & health insurer · three continents
Improvement in underwriting and reserve accuracy
Professionals in the analytics organization, across three continents
Global advanced-analytics function in the firm’s history
- 01Experience dataglobal, fragmented
- 02Predictive modelswith actuaries
- 03Committee reviewmultiple regulators
- 04Pricing & reservingassumptions changed
Reserve accuracy is a balance-sheet number someone signs — the models had to clear that bar
The situation
A global carrier with deep actuarial talent and no institutional capability for advanced analytics. Modeling happened in pockets, findings didn’t travel between regions, and underwriting and reserving assumptions were set with methods that predated the data available. The gap wasn’t talent — it was the absence of a function with the standing to change how decisions were made.
What we built
- 01A global analytics organization built from a standing start — 160+ professionals on three continents
- 02Predictive underwriting and reserving models developed with the actuarial function, not adjacent to it
- 03A governance and documentation standard designed to clear regulators in multiple jurisdictions
- 04A delivery cadence that put findings into pricing and reserving decisions, not a research archive
Reserve and underwriting accuracy isn’t a data-science metric — it’s a balance-sheet statement a chief actuary signs. The models changed assumptions, rather than merely informing a discussion about them, because the function was credible inside the profession reviewing it.