Will AI flatten job levels?
Some teams will need fewer managers. Most will not. A layer cut on a forecast is hard to put back.
Short answer: AI will remove management layers in some functions, not everywhere. Teams whose managers mostly route work and compile reports can run flatter. Teams where managers coach new people, judge quality or sign off on risk cannot. Decide job family by job family, from what managers do. Keep a career step for the people a cut affects.
This is general information, not legal advice.
What the data and the rules say
Flattening is already under way. Two in five employees in a global workforce study from April 2025 reported fewer management layers at their employer. Over a third felt less sure of their direction as a result [1].
It is not spread evenly. Average US team size per manager rose by more than one person between 2024 and 2025, yet the typical (median) manager still leads about six people. A small group of very large teams pulls the average up: Roughly one manager in eight now has 25 or more reports, while two in three have fewer than ten [2].
In the US, flattening can also change how a role is classified. To be exempt from overtime as an executive, a manager must, among other tests, customarily and regularly direct the work of two or more full-time employees or their equivalent. The test counts employees, not software [3]. A manager left with one direct report and a set of AI tools may no longer meet that test.
What it means for a company of 50 to 500 people
Flattening fits coordination work. Where a team lead mainly assigns tickets, tracks status and writes the weekly report, AI can take much of that and one manager can lead more people. Support and operations teams are the likeliest candidates.
It fails where managers build people and judgment. A manager who trains new hires, reviews work before it ships or signs off on regulated decisions does work that does not shrink with better tools. Widen that span and quality or development is what slips.
Cutting a layer removes a career step. The team lead role is often the first promotion people aim for. Remove it and strong performers need another way up, which usually means a specialist track. Our guide on whether a dual career ladder is worth it covers what that track has to promise.
The remaining manager job changes level. A manager leading 20 people with AI support does different work from one leading six. Level that role on its scope and decisions, not on headcount alone, as our guide on re-leveling AI-changed jobs explains.
Our view
We would not cut a layer across the company because a forecast says AI will. We would look at each job family and ask what its managers spend their week on. Where most of it is routing and reporting that AI now handles, merge the layer and give the people in it a specialist path. Where most of it is coaching, review and sign-off, keep the layer and leave the span alone. In the US, check the classification of any manager whose team shrinks before you change their pay or hours.
Where we help
Our job architecture and leveling work sets the levels for each job family, including manager and specialist tracks, so a change in layers does not break pay or careers. To discuss your own position first, Book a free call.
Sources
- Korn Ferry, Workforce 2025: Power shifts (survey of 15,000+ professionals in ten markets, published 7 April 2025), checked 2026-10-31
- Gallup, span of control and team size (published 13 January 2026), checked 2026-10-31
- US Department of Labor, Fact Sheet #17B: Exemption for executive employees under the FLSA (revised August 2024), checked 2026-10-31