By Isabelle Bousquette
As corporate America looks to redesign the workplace for the AI age, there's a new kind of team gaining traction: the "pod."
Smaller than a traditional engineering group, pods are designed to move faster to build and iterate on products. They're also more cross-functional, including not just engineers but also designers and applied scientists. And critically, all that expertise is concentrated in just a handful of human workers (anywhere from one to eight), as well as AI agents.
For years engineering teams have been slowly favoring smaller and smaller teams in the name of speed and agility, but the growing capabilities of AI coding assistants and other agents that can potentially reduce the time-to-ship are allowing for even smaller pod-size structures. With AI agents doing more of the actual software development, including coding and testing, it takes fewer human workers to build products at scale.
Earlier this month, when Coinbase announced it was laying off 14% of its workforce, in part to reorganize itself for the AI age, Chief Executive Brian Armstrong wrote that he was doubling down on AI-native pods.
"We'll be concentrating around AI-native talent who can manage fleets of agents to drive outsized impact. We'll also be experimenting with reduced pod sizes, including 'one person teams' with engineers, designers, and product managers all in one role," he said.
The benefit of pods is speed, agility and the ability to do more, faster, with fewer resources, said Rob Witoff, head of platform at Coinbase. The company now has a team of three people working on an AI adviser project, he said, adding that historically an undertaking like that would have required 10 to 15 people.
"There are these exponential gains when we have fewer people because you're spending a lot less time in meetings and reviews and getting people on the same page," said Witoff. "This is I think, fundamentally, why startups are fast, because they've got a couple hyper-aligned people with very little unnecessary overhead."
Pods are the next step in an ongoing engineering org evolution. In recent decades, so-called scrum teams -- cross-functional groups focused on deploying and iterating quickly -- have replaced a slower, step-by-step engineering methodology known as "waterfall."
"The pod structure is just the next evolution of the scrum team, just with some very powerful AI tools that allow you to be going much faster than the past," said Dan Diasio, Ernst & Young's global consulting AI leader.
The better the technical tooling, in this case AI agents, the smaller the human teams need to be, said Deepak Singh, a vice president at Amazon Web Services.
In the early days of Amazon, he said, Jeff Bezos famously advocated for the idea of the "two-pizza team" -- that is, any team should be small enough that it could be fed with two pizzas.
Today, Amazon still has its "two-pizza teams," but, hypothetically, a 16-person, two-pizza team might be broken up into two smaller cross-functional pods of eight, each focused on a specific project, according to Singh. Everyone in your pod is also on your two-pizza team, but not everyone in your two-pizza team is in your pod, he said.
"If you have a large team, you spend half your time just talking to each other and trying to figure out what needs to be done," said Singh. "The goal has always been velocity and one of the best ways of increasing velocity is making decision-making faster."
Write to Isabelle Bousquette at isabelle.bousquette@wsj.com
(END) Dow Jones Newswires
May 18, 2026 08:00 ET (12:00 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments