Headlines are for Heroes

In 1934, J. Edgar Hoover got word that his agents had located John Dillinger. Before confirming the capture, before validating the intelligence, he went to the press. Headlines declared victory.

Dillinger slipped away. The headlines remained.

Last week I wrote about Hoover doing the casework before building the FBI - about doing the investigation yourself. But there’s a second part to this story: knowing the difference between finding a suspect and making an arrest.

Between building something and shipping something that matters.

Right now, every company is claiming AI transformation. “We’ve scaled AI across the enterprise.” “Our workforce is AI-powered.” “We’ve deployed hundreds of use cases.”

The reality? Most are still in pilot mode. They built a POC that worked once, presented it to leadership, and put it in the newspaper.

They found Dillinger. They didn’t make the arrest.

Here’s where we are:

• We’ve shipped AI into production in 30+ places, scaling them across teams - several with measurable ROI
• We’ve partnered with one function to fundamentally transform how work is done, from reauthoring goals to full process reimagination
• We’ve piloted another dozen things. Some will scale. Some won’t. Some looked brilliant in the demo and fell apart when they hit real users
• And we have a backlog of high-confidence ideas that haven’t proven anything yet

I know the difference. And I’m not putting pilots in the newspaper.

This is where most AI “transformation” falls apart - in the gap between proof-of-concept and proof-of-value.

A POC that works once in controlled conditions isn’t scale. A pilot with ten enthusiastic early adopters isn’t adoption. A dashboard showing usage isn’t impact.

The arrest isn’t made until:

• It runs in production without you babysitting it
• People use it without being asked
• You can measure the value it creates

Until then, you’re just creating more AI theater. More headlines that age poorly. More FOMO that drives bad decisions.

Do the investigation. Build something real. Test it honestly. Then put it in the newspaper.

Don’t be Hoover claiming Dillinger when the work isn’t done.