AI changes organizational structure by shrinking the distance between the top and the bottom of the company. Boris Cherny, who leads Claude Code at Anthropic, describes a team where everyone codes: the product manager, the engineering manager, the designer, the finance lead and the data scientist. He estimates about 50% overlap between engineering, product and design, and expects the “software engineer” title to fade. Anthropic gives staff one shared title: member of technical staff. Anthropic is leading this change, but it will not stay in software. Every function that produces output or creative work, including design, marketing, PR, mechanical engineering and architecture, will reorganize the same way.
Executives Must Get Hands-On
Cherny’s example is an engineering manager who still writes code. I would extend that to the whole executive suite and to every output function. Think of the CMO who builds campaigns with AI, the PR chief who drafts the media plan, or the design, engineering and architecture principals who run the models themselves. The executive who uses AI is far more effective because they can create, refine, test and decide independently and sequentially. There’s no information lost, no handoffs, no interpretation needed. The executive can go from need directly to their preferred solution.
AI-Native Entry-Level Employees Carry an Experienced Knowledge Base
The other half of the merge happens at the bottom. An AI-native new hire working with a capable model has deep, experienced knowledge available on day one. The model supplies pattern recognition that once took a decade to build, so a junior designer, marketer, PR associate, engineer or architect can prototype, analyze and ship work that used to need layers of specialists.
What This Does to the Org Chart
Executives move down into the work. Entry-level employees move up in capability. The middle layers that existed to relay information and translate between functions lose their reason to exist. Cherny describes the work in five archetypes instead of job titles: Prototyper, Builder, Sweeper, Grower and Maintainer. Those archetypes fit a campaign, a building or a mechanical assembly as well as they fit a codebase.
Where Leaders F-Up
The common mistake is bolting AI onto the old chart: same layers, same titles, same distance from the work. Then the pilots stall. Three moves help:
- Use the tools yourself every week.
- Staff for archetypes, not titles.
- Pair AI-native juniors with experienced leaders who supply judgment.
Frequently Asked Questions
How does AI change organizational structure? It flattens it. Leaders become practitioners, new hires gain expert-level leverage, and the layers in between shrink.
Marc Drucker is the author of How Leaders F-Up Innovation and an advisor on innovation leadership and strategy.