AI Is Flattening Software Companies. The Danger Is Cutting the Wrong People.

Note: This story was written with AI assistance and edited by me.

Two software companies tied job cuts to AI in the first week of November 2026. On November 2, The Register reported that SUSE, the open source software company owned by EQT, told its roughly 1,900 employees it would reduce its global workforce by about 10 percent. CEO Dirk-Peter van Leeuwen wrote that "automation and AI are fundamentally shifting how software companies operate" and that modern tooling is "enabling teams to move faster with less operational overhead and flatter reporting lines." On November 4, NICE CEO Scott Russell explained on the company's third quarter earnings call that its cut of about 500 jobs, roughly 5 percent of about 10,000 employees, includes research and development, where "our teams are leveraging AI tools to increase productivity materially." He said NICE is also "removing layers of management and span of control." NICE expects $15 million to $20 million in restructuring charges, and its third quarter revenue rose 8 percent to $791 million.

Those are the facts as reported. Everything below is my opinion.

My view: flattening is the right call, and AI does make good engineers more productive. But the cut that pays off is the layer that only passes information along, not the senior people who decide what gets built and catch what the tools get wrong. Companies that cut judgment to hit a cost target will pay it back in quality, rework and lost clients.

Flatter is overdue

Most software organizations added management layers when hiring was easy. Every new team got a lead, every few leads got a manager, and status updates traveled up and down a chain that slowed every decision. AI tools that summarize work, track progress and draft reports remove much of the reason those layers existed. If a company can see the state of a project without three meetings, it should not pay for three meetings.

Productivity is not the same as judgment

The harder claim is the one about R&D. AI tools really do help developers write code faster. What they do not do is decide whether the code solves the right problem, fits the system it lands in or stays secure under real traffic. That work falls on experienced engineers, and when AI raises output, it raises the amount that has to be reviewed.

So if a company uses AI gains as a reason to cut the people with the most context, it is removing the brakes at the moment the car got faster.

What leaders should do instead

Cut handoffs before you cut people. Every layer that only forwards information is a candidate.

Keep the senior engineers who own outcomes. Give them the AI tools and a smaller, clearer team.

Measure results, not activity. Count shipped features, defects and client retention, not lines of code or tickets closed.

Be honest with the people who stay. A restructuring that looks like a cost cut dressed up as strategy damages trust faster than the cut itself.

A fair counterpoint

To be fair, neither company says AI is the only reason. SUSE's owner reportedly considered selling the business earlier this year, and NICE's CEO said cloud growth was "not where it needs to be." In July, Russell had told CTech, "Our plan is to grow, not shrink." Some of these cuts are ordinary cost discipline with AI in the headline. My view is only that the AI part of the story should mean fewer layers, not less expertise.

What I learned removing layers

I have run teams with too many layers and teams with almost none. The layered version always looked more organized on a chart. In practice, a simple question from a client could take two days to answer, because it had to travel up to someone who knew and back down to someone who could act.

When I flattened a team, the change was fast. Senior engineers talked to clients directly, decisions happened in one conversation and problems surfaced while they were still cheap to fix. What I never did was remove the people who knew the system best. Every time a project went badly, it was because nobody with enough experience was close enough to say no.

AI tools make that lesson sharper. A tool can draft code, tests and status reports in minutes. It cannot tell you that the client asked for the wrong feature or that a shortcut will break next quarter. That still takes someone who has seen it go wrong before.

Takeaway

AI gives leaders a real reason to remove layers that only pass information along. It does not give them a reason to remove judgment. I took a firm from $6M to $60M in under three years, and the turning point was building flat teams around senior people who owned the result.

If you lead a software team, look at your org chart this week and ask one question about each layer: does it make decisions, or does it only forward them? The answer will tell you where to cut and where to invest.