From Two-Pizza Pods to Two-Slice Pods

From Carlos Gonzalez de Villaumbrosia, “The AI Operating Model: Why 80% of Product Teams Are Stuck – and What the Top 1% Are Doing Differently” (Product School, May 27, 2026), written after ProductCon New York, his company’s conference, where he launched an AI consulting firm called Product Partners.

Villaumbrosia argues that AI tools pay off for a company only after it reorganizes around them. Drawing on the more than 1,400 product teams his training company has worked with, he says over 80% are “stuck at Level One”: people get faster drafts and meeting summaries from AI, but revenue and shipping speed stay the same. The top 1%, he says, built an “AI operating model,” covering both the AI tools a team shares and how its people are organized and trained. Smaller teams are one of four changes he lists under “People Initiatives.”

Factories hit the same lag with electricity. Steam-era plants ran every machine off shafts turned by a central engine, so machines had to sit in rows along the shafts. At first, the economist Paul David found, factories swapped that engine for an electric motor and left the shafts in place. The payoff came about forty years after the first power stations opened, once machines got their own motors and could be arranged along the flow of materials. Computers repeated it. In a study of 527 large U.S. firms, Erik Brynjolfsson and Lorin Hitt found that measured over one year, computers paid off about as well as any other investment, but measured over five to seven years, their boost to productivity was up to five times larger. The gains came, they argue, as firms rebuilt their work systems and business processes around the computers, an overhaul that may cost up to ten times as much as the computers themselves.

Dan Shipper, cofounder and CEO of the company Every, made the same pun three months earlier, in an essay about Every running each of its four software products with one person: “Two slices, to feed one person.”