You're About to Automate Away Your Safety Net | with Tom Redman

Everyone’s racing to hand their work off to AI agents, but most corporate AI projects are failing, and bad data is a big reason why. Tom Redman, known as “The Data Doc,” built the first data quality lab at Bell Labs and has spent four decades in the least glamorous, most important part of the AI story: the data underneath it. In this conversation, Tom explains why cleaning up bad data is a losing battle, and why the companies that win focus on the data they’ll create tomorrow instead. Plus, learn why line managers and AI experimentation rarely mix, how a “lab, factory, and bridge” structure fixes that, and why the customer-supplier model is the habit that separates real AI value from slop. You’ll Learn: Why data cleanup never works and how to fix bad data at the source Why every AI project should start with the customer and what “good” looks like to them, not with data How to run the Friday Afternoon Measurement to gauge your team’s data quality in a single afternoon Why the only real AI transformation happens “between people’s ears” The “lab, factory, and bridge” model for separating AI experimentation from day-to-day operations Resources mentioned:  Tom’s book People and Data Data Quality Solutions The Friday Afternoon Measurement The “bridge” concept from statistician Diego Kuonen (article written with Roger W. Hoerl and Thomas C. Redman) Producer: David Grabowski Theme Music: Aleksey Chistilin