After two stalled portal projects, a specialty insurer let us re-sequence the program: canonical data first, product operating model second, broker experience third. The third portal attempt shipped in half the time of either failure — and kept shipping.
Two previous "digital front door" programs had died the same death: beautiful broker portals built on top of four policy systems that disagreed about what a policy was. Every screen needed bespoke integration, every release broke another, and the program burned out at the experience layer without ever touching the cause.
We spent the unglamorous first six months building the canonical submission-to-policy data model and the service layer to expose it — with underwriting, not IT, as the arbiter of every definition. Product squads were stood up around journeys rather than systems, funded persistently instead of by project. Only then did the broker experience work begin, and it moved fast because every screen drew from one source of truth.
“The first six months produced nothing a broker could see, and it was the most valuable six months of the program.”
Facing something similar? Talk to the partner who ran this →