The Two-Front War

My vision has always been to bridge the brilliance of our Excel-based underwriting models with the power and scale of our data warehouse. To automate the data entry portion of underwriting and, ultimately, to unclog our deal flow pipeline and help us get to no’s in a fraction of the time.

But for weeks, that dream felt like a fantasy. I was stuck in a two-front war.

Backend: A slow Office Script engine running through Power Automate, constantly timing out.

Frontend: A brittle dirty-tracking system watching 500+ fields for changes. Every bug fix created two more. Progress stalled.

The breakthrough came when I stopped patching cracks.

I partnered with AI. Gemini as my cognitive partner to architect the strategy. Claude Code as my coding partner to implement it. Together, we explored every alternative, stress-tested libraries, and validated architectures. We didn’t just find a patch. We rebuilt the foundation.

A new engine. A high-performance .NET microservice using Aspose.Cells, deployed on Google Cloud Run. A model that used to time out in 2 minutes now runs in 9.85 seconds.

A new brain. A solid frontend built on Zustand and Immer. The fragile UI became predictable and stable.

The wall is gone. The bridge to the data warehouse is finally real, unlocking automation and setting the stage for AI in our underwriting process.

Building is hard. Persisting is harder. AI makes it possible.