Jack Dorsey didn't hedge. When Block announced 4,000 layoffs — nearly 40% of the entire company — he said the quiet part out loud: "This is not driven by financial difficulty, but by the growing capability of AI tools to perform a wider range of tasks."
Amazon cut 14,000 corporate roles in the same wave of restructuring. The stated goal: eliminate bureaucracy, remove organizational layers, let algorithms handle coordination, reporting, and the decision-making that used to require a management layer.
Gartner estimates that by the end of 2026, one in five organizations will have used AI to eliminate more than half of current middle management positions.
This is not a slowdown. It's a structural shift in what organizations look like. And it's creating a hiring problem almost nobody is talking about.
What Middle Management Actually Did
Here's the thing most org charts obscure: middle managers weren't primarily in charge of people. They were in charge of information.
They translated strategy downward into task-level clarity. They aggregated status upward into executive-readable summaries. They tracked dependencies, flagged blockers, coordinated handoffs, ran status meetings, maintained the documentation trail, owned the project management tools, and made sure the right people knew the right things at the right time.
That's not leadership. That's information architecture. And AI is very, very good at information architecture.
When Amazon says it's letting AI handle coordination, reporting, and decision-making functions once reserved for managers, it's not describing a futuristic experiment. It's describing what happened when they actually deployed the tools they'd been building. The coordination function got absorbed. The management layer that was doing it became redundant.
The Org That Remains Is a Different Shape
This is the part the layoff headlines miss.
When you collapse three management layers into one — or zero — you don't just have fewer people. You have a different organizational geometry. The remaining structure is flatter, faster, and far less tolerant of ambiguity. Decisions that used to travel up and down a chain now happen at the point of contact, in real time (often inside a Slack thread, not a meeting room), with AI systems providing context and execution support.
That's great for efficiency. It's brutal for team design.
The leaders who survive in a flat, agentic org are not the managers who just got cut. The manager who was good at status meetings, at writing summaries upward, at herding stakeholders across a process — that skill set is exactly what AI absorbed. Those people, however talented, aren't well-positioned for what the org that kept them needs next.
What the flat org needs is something that didn't have a job title 18 months ago: a technical leader who can operate like an architect, execute like an IC, and make judgment calls without a decision-making layer above them. Someone who's comfortable running AI agents as part of their workflow and knows when to override them. Someone who doesn't need a meeting to get aligned — because there's nobody to have the meeting with.
The Hiring Problem Nobody Has a Job Description For
Here's where this gets concrete for CTOs: you're now trying to hire into an org shape that doesn't have established hiring infrastructure.
You know how to screen for a senior manager. You have a rubric, a panel structure, a comp band. You know what good looks like.
You don't yet have that for the person who replaces three managers in a flat, AI-augmented structure. That role has components of technical depth, strategic thinking, agent orchestration, and organizational judgment — none of which appear cleanly on a resume or score predictably on a technical interview.
What makes this urgent: you're not the only company trying to find this person. Every enterprise that went through a management layer reduction in Q1 is now staring at the same gap. The talent supply of people who have actually operated in flat, agentic orgs is tiny, because those orgs barely existed two years ago.
The people who have this experience are inside the companies that were early — the ones who built the agentic infrastructure before it was mainstream, the ones who survived leaner org structures because they were already working that way. They're not looking at job boards. They're heads-down on projects that matter to them.
What to Do About It
The job description is the first signal you send. Posting "Engineering Manager" or "Senior Director of Engineering" tells a candidate what level you're hiring at — not what the actual work is. If the job is "own the delivery of our agent infrastructure with no coordination layer above you," say that. The candidates who want that job are different from the ones who want a traditional leadership role, and you need them to self-select.
Extend the search radius beyond conventional sourcing. The person you need may have a title that doesn't match what you're hiring for. They may have been running a two-person engineering team inside a startup with no managers at all. They may be a senior IC who's been operating at the level above their title for two years. Resume scanning won't find them.
Move fast. The window on this specific talent cohort is short. When more than 52,000 tech roles disappear in a quarter and the best engineers from those companies hit the market, the strongest candidates are gone in days, not weeks. If your process has three rounds and a hiring committee approval cycle, you're not competitive for the ones you actually want.
The Question Most Companies Are Getting Wrong
The instinct after a management layer reduction is to figure out which middle managers to keep. That's the wrong question.
The right question is: what functions did that layer perform, which of those functions has AI actually absorbed, and what judgment-intensive functions remain that a different kind of leader needs to own?
If you answer the first question, you're managing the reduction. If you answer the second, you're designing the organization you actually need to build next.
The companies that get this right in the next 12 months will have a structural advantage going into 2027 that their competitors will struggle to close. The flat, agentic org that's staffed correctly is faster, cheaper to operate, and more adaptive than the layered structure it replaced.
The ones that don't get it right will cut headcount, leave the same organizational shape in place at smaller scale, and wonder why the efficiency gains aren't showing up.
VC5 Consulting works with technology leaders navigating exactly this hiring challenge — finding technical leaders built for flat, agentic organizations when the traditional search process wasn't designed to find them. If your org just changed shape, let's talk about what comes next.