Zero Correct Answers: When AI Meets Raw Claims Data

April 14
43 mins

Episode Description

Send us Fan Mail

Healthcare organizations are racing to layer AI on top of their data, but what happens when that data was never designed for the questions being asked? 

Ryan Leurck, Chief Analytics Officer at Kythera Labs, spent years in aerospace engineering before discovering that healthcare's data quality problems run far deeper than most outsiders expect. 

In this conversation, he shares research showing that large language models produced zero correct answers when applied to raw claims data, and explains why building a reliable semantic and structural foundation is the prerequisite most organizations skip. We dig into the terminology gaps between disciplines, the difference between billing data and decision-ready data, and what it actually takes to make AI useful in healthcare.

LinkedIn

Kythera Labs


See More ChalkTalkJim Episodes




See all episodes