From rating workbooks to a tested rating engine
A Canadian commercial-lines MGA priced its business in Excel workbooks that had grown over many years. We rebuilt the rating logic as a tested calculation engine, built a web application on it for underwriters and brokers, and support the operations that feed it.
Before
- Pricing lived in large Excel workbooks, with logic spread across tabs and macros
- Broker quotes were slow to turn around
- Formula errors could creep into prices unnoticed
- There was no central record of who rated what, or why
What we did
- Phase 1Rebuilt the rating logic as a calculation engine, and tested it against the workbooks’ own results before anything changed for underwriters.
- Phase 2Built a web application for underwriters and brokers on that engine, with role-based limits on who can approve what.
- OngoingSupport for the operations around it: submission intake, schedule-of-values clean-up and loss-run intake.
What they have now
- Every quote comes from one tested engine instead of copies of a workbook
- Rating decisions are recorded along with who made them
- Approval limits are enforced by the application, not by memory