When I joined Highmark, the claims platform looked operational on the surface but broke under the weight of volume and inefficiency. Hundreds of analysts were reviewing every claim, even though the vast majority followed predictable rules.

The team knew automation was a buzzword, but no one had figured out how to scale it without compromising compliance. I started with a simple question: how many of these claims truly need human eyes?

The answer was surprising. Only fifteen percent required manual intervention.

That meant 85 percent of the process could, in theory, be automated. But the tech stack wasn’t ready, and neither were the teams.

I spent months tracing each adjudication path, validating with claims experts, mapping out decision trees, and reviewing CMS rules side by side with internal business logic.

Once we had alignment, I led the creation of Highmark’s first robotic adjudication platform. It validated claims, enriched data, flagged anomalies, calculated payments, and even checked for fraud—all before a human ever touched it.

Within a year, manual intervention dropped by sixty percent. We saved six dollars per claim. Over five hundred thousand claims ran through the system in year one.

It was the first time Highmark trusted automation with money movement at scale, and it worked.

Other departments quickly adopted the framework for billing and enrollment use cases, proving the model had legs.

---