How Gennaro Brooks-Church Turns Underwriter Corrections Into Smarter Systems

Most software stays exactly the same after it launches. You install it, you use it, and unless a company pushes an update, it works the same way on day one hundred as it did on day one. Gennaro Brooks-Church, Founder and CEO of Cazimir and Brooks Energy, has built his AI systems to work differently. Every time an underwriter fixes something the system got wrong, that correction becomes part of how the system learns going forward.

The Problem With Systems That Never Change

A lot of automation tools in insurance are built on fixed rules. They follow the same logic every single time, no matter how many times a team points out that something isn't quite right. Gennaro Brooks-Church has said this approach misses the point of what AI can actually offer an underwriting team. If a system never improves after it's deployed, it stays stuck at whatever level of accuracy it had on launch day, even as it processes thousands of submissions.

He built his technology to do the opposite. Instead of treating the system as finished once it goes live, he treats every correction an underwriter makes as new information the system should learn from.

What Happens When an Underwriter Fixes a Mistake

Here's the basic idea. When Cazimir processes a submission, it pulls data out of documents like ACORD forms, Statements of Values, and loss runs. Sometimes it gets something wrong, maybe a number is misread, or a field is misclassified. When an underwriter corrects that mistake, the system doesn't just fix that one instance and move on. It uses that correction to adjust how it handles similar situations in the future.

Over time, this means the system starts to reflect the specific standards and habits of the team using it. A deployment that's been running for six months should understand a client's formatting quirks and internal preferences far better than it did on the very first day.

Why This Builds Trust, Not Just Accuracy

Gennaro Brooks-Church has been clear that this isn't only about getting better numbers. It's also about building trust with the people using the system. Underwriters are more likely to rely on a tool that visibly gets better the more they use it, rather than one that repeats the same mistakes over and over with no sign of improvement.

This is part of a bigger philosophy he's talked about often. He doesn't want AI to replace the judgment of underwriters. He wants it to support that judgment, and treating corrections as valuable input, rather than something to quietly ignore, is one of the clearest ways that shows up in practice.

Keeping the Learning Process Transparent

One challenge with systems that keep learning is that they can start to feel less predictable over time. Gennaro Brooks-Church has worked to avoid that by keeping the learning process tied to traceable data. Even as the system adjusts based on new corrections, every output still links back to its original source document. Underwriters can always see where a piece of information came from, even as the system behind it continues to improve.

This matters because a system that learns but can't explain itself creates a different kind of problem. Gennaro Brooks-Church has said the goal is for the technology to get smarter without becoming harder to understand or trust.

Why This Approach Fits Insurance Specifically

Insurance underwriting isn't the same everywhere. Different brokers format documents differently. Different carriers have different internal standards. A system that can't adapt to these differences ends up creating extra work instead of saving time. By building AI that learns directly from the corrections made by real teams, Gennaro Brooks-Church has tried to make sure his technology fits the specific way each organization actually works, rather than forcing every client into the same rigid process.

A Simple Idea With a Big Impact

At its core, the idea behind this approach is fairly simple. Mistakes are useful. Instead of treating them as failures to move past quickly, Gennaro Brooks-Church treats them as the exact information his systems need to keep getting better. It's a small shift in thinking, but it's a big part of why the technology behind Cazimir and Brooks Energy continues to improve the longer it's used, rather than staying stuck at whatever it could do on the day it was first switched on.