A data lake that unlocked $9M in pipeline

The problem

Gerber Life Insurance had a data problem most enterprises have but few say out loud. Their data was federated, siloed, and lived in systems that did not know about each other. Every operational issue pulled in everyone, because nobody knew who owned what. Calls would balloon. Cost would balloon. Decisions would stall.

Their public-facing surface was Drupal. The engagement included migrations through Drupal 7 to 8 to 9. But the website work was the visible part of a much bigger story underneath. The real problem was that Gerber could not answer basic operational questions about their own business without convening a committee.

In an insurance business, that is not an inconvenience. It is a tax on every decision the company makes.

What we led

Our founder architected a cloud-based data lake that integrated Gerber's disparate systems. The job was to make data flow into a single structured layer so operational questions could be answered with queries instead of meetings.

The architectural decisions were about how to ingest federated data without disrupting the systems already running, how to structure the lake so it could feed both operational reporting and downstream applications, and how to set up the team to maintain the infrastructure after the engagement ended.

It also meant working closely with business stakeholders to convert their operational pain into the right data questions. The technical architecture was the easier half.

What this shaped at Fidget Labs

The data foundation is what makes AI possible later. In 2019, nobody was framing data lakes as "AI infrastructure." The framing was operational efficiency. But the same architectural discipline that unblocks operational reporting is what unblocks AI deployment four years later.

This is now central to how Fidget Labs thinks about AI readiness. The companies that did this work then are the ones who can deploy AI now without spending six months on a data cleanup project first. If a company is looking at AI in 2026 and their data is still federated and siloed, the work they are avoiding is the same work Gerber did in 2019. The cost of not doing it compounds.