Insure Pilot
Reverse engineering a thirty year old insurance system, then rebuilding it as a modern CRM.
The Challenge
A managing general agency ran its entire operation on software built in the 1990s. Everything took three to four times longer than it should. The holding company behind the agency decided to build something new: an insurance CRM that handles claims, agencies, agents, and end customers in one place. They would own it, then license it to the other agencies still on the old platform.
Every workflow, field, and rule had to be captured by watching specialists across departments run their workflows: what was working, what wasn't, and what they wished for to cut manual effort.
The Solution
A mapping exercise that ran for months. Three sessions a week, screen shared, one area at a time. We mapped every process, along with what was required from an audit and legal perspective. Screens were captured into Figma, captures became flow diagrams, and every gap became a question for the next session.
Then we stripped it down. Hundreds of pages, duplicate workflows, and steps that only existed because of decades-old limits became 30 to 40 efficient flows, heavily automated.
From day one the build was planned around AI. A developer reading a spec fills gaps with judgment, but a build engine fills them with whatever is nearest, so the specs had to cover everything: what happens when a session dies mid report, what makes a note immutable, how a passenger marked injured in one section becomes a claimant in another. Rules like that are not visible in the interface. The only way to find them was to keep asking until someone said them out loud.
The Outcome
A thirty year old system became one modern platform with three products: a portal for end customers, a portal for agencies, and an internal system for underwriters and claims agents. All three share a single design language, so an agent, an underwriter, and a policyholder are working in what feels like the same product.
The system is in build and close to production. In testing, the improvement is significant: faster, easier to use, and requiring fewer people to manage the same book of policies. The first agency migrates onto it at launch, then it gets licensed to every agency still on the old platform.
