WHO CATCHES A BAD HIRE NOW THAT MIDDLE MANAGEMENT IS GONE?

For most of the last two decades, the manager sitting between the executive team and the front line did more than approve time off and run status meetings. They were an informal check on hiring decisions that nobody ever wrote down: the person who noticed a new hire wasn't ramping by week three, who caught a mismatch before it became a quarter of lost output, who had the standing and the proximity to say something before a bad call turned into an expensive one. That layer is being removed on purpose, and organizations are betting that AI can absorb what it used to do. Hiring quality is one of the first places that bet gets tested, and most companies have not built anything to replace what they cut.

The org chart got thinner on purpose

The flattening trend has a name and a timeline. In an October 2024 report, Gartner projected that by the end of 2026, one in five organizations would use AI to flatten their structure, eliminating more than half of their current middle management positions. That prediction has held up well. Average span of control, the number of direct reports per manager, climbed from 8.2 in 2013 to 10.9 in 2024 and 12.1 in 2025, according to Gallup data. Korn Ferry's Workforce 2025 survey found 41% of employees globally, and 44% in the U.S., say their organization has already eliminated a layer of management.

None of this is accidental. It's a deliberate cost and speed decision, and in a lot of cases it's the right one. The problem is what quietly goes with it.

What that middle layer was actually doing

Slayton Search Partners describes the shift as a move from a traditional pyramid to an hourglass: a well-staffed top, a well-staffed front line, and a shrinking middle in between. In a 2026 report covered by Hunt Scanlon Media, Slayton's John Doyle makes a point most flattening conversations skip entirely. When a senior hire underperforms, the instinct is to blame the candidate. The fit wasn't right, the experience didn't translate. Doyle's research points somewhere else: in organizations where the middle has eroded, new hires land in a vacuum, expected to perform without the infrastructure that used to make performance possible. That's not a talent problem. It's a structural one.

Deloitte's 2025 Global Human Capital Trends survey backs this up from a different angle. The managers who remain spend nearly 40% of their time on immediate operational and administrative problems, and only 13% developing the people they lead. The layer that used to catch a hiring mistake early is now too stretched to catch much of anything.

AI is absorbing the coordination work, not the judgment

The assumption behind most flattening decisions is that AI can pick up the slack. It can, for a specific kind of work. Scheduling, status reporting, performance dashboards, the administrative half of what a manager used to do, AI handles competently. What it hasn't shown it can do is the part that caught a bad hire: reading that a candidate's confidence in the room didn't match their answers, noticing a pattern across three references that nobody flagged individually, knowing when a strong resume and a strong interview still don't add up to the right person for this specific team at this specific moment.

There's a useful precedent here. HireVue, one of the largest AI interview platforms on the market, discontinued its facial expression analysis tool after an independent audit found the feature added no predictive value beyond what verbal and audio analysis already provided. The lesson wasn't that AI has no place in hiring. It was that some of what looked like AI judgment was never judgment at all. Automating the appearance of oversight isn't the same as building it.

What it costs when nobody catches the miss

The U.S. Department of Labor puts the minimum cost of a bad hire at 30% of that employee's first-year salary. SHRM estimates the full replacement cost at 50% to 200% of annual salary, with the number climbing toward the high end as seniority increases. For executive roles specifically, SHRM's 2025 Benchmarking Report puts the average cost-per-hire at $35,879, nearly seven times the $5,475 average for non-executive roles, and up 21% since 2022. None of those figures include the cost of the role sitting open again, the team that absorbed the gap, or the six to nine months before anyone admits the hire isn't working.

Those numbers were calculated in a world where a manager was usually somewhere in the process to catch the problem early. Remove that layer without replacing what it did, and the same bad hire now runs its full, expensive course before anyone with the authority to act even notices.

The fix is more rigor, not more layers

Rebuilding middle management isn't the answer, and most CHROs know it. The answer that holds up in the research is structure: a deliberate, repeatable process that doesn't depend on a specific person being in the room to catch what matters. Frank Schmidt and John Hunter's landmark meta-analysis of eight decades of selection research found that unstructured interviews predict job performance at a validity of just .38, barely better than chance. Structured interviews, using standardized questions and a shared scoring rubric across every candidate, reach .51 on their own and .63 when paired with a cognitive ability assessment. That gap isn't a rounding error. It's the difference between a hiring process that catches a bad fit and one that hopes somebody does.

"Nobody sets out to make a bad hiring decision," said Desiree Goldey, Founder and CEO of Do Better Consulting. "They make it inside a process that has fewer checkpoints than it used to. Structure isn't bureaucracy. It's the thing that used to be a person, built on purpose, so it doesn't depend on who happens to be in the room."

What this looks like in practice

For talent leaders operating with a flatter structure than they had two years ago, a few shifts do more than adding headcount back ever would:

• Define the decision criteria before the process starts, not while comparing finalists. If "culture fit" is on the scorecard, write down what it actually means for this role.
• Build a second reviewer into the workflow itself, not into the org chart. A calibration step that any two trained interviewers can run is more durable than one that depends on a manager who may not be there next quarter.
• Score every candidate against the same rubric, in writing, before the group compares notes. This is the single biggest driver of the validity gap between structured and unstructured hiring.
• Audit where your process still assumes a layer of oversight that no longer exists. Most companies haven't done this since before the last round of flattening.

That last one is the real diagnostic question underneath this whole shift. Not "did we flatten too much," but "what did we quietly stop checking when we did."

Common questions

What is hiring rigor?

Hiring rigor is a deliberate, repeatable evaluation process, standardized questions, shared scoring criteria, and a required second review, that doesn't depend on any one person's judgment or availability to catch a bad fit before an offer goes out.

How much does a bad hire cost?

The U.S. Department of Labor estimates a minimum of 30% of the employee's first-year salary. SHRM puts full replacement cost at 50% to 200% of annual salary, with executive roles trending toward the high end. SHRM's 2025 Benchmarking Report puts average executive cost-per-hire at $35,879.

Why are companies flattening middle management?

Primarily cost and speed. Gartner projected in 2024 that by the end of 2026, one in five organizations would use AI to flatten their structure and eliminate more than half of current middle management roles, and average span of control has climbed from 8.2 direct reports per manager in 2013 to 12.1 in 2025.

Can AI replace the hiring oversight middle managers used to provide?

AI handles the coordination and administrative work well: scheduling, reporting, dashboards. It hasn't demonstrated the judgment that catches a bad hire: reading a mismatch between a candidate's confidence and their answers, or a pattern across references that no single flag would catch. That's a structural gap, not a tooling gap.

What is a structured hiring process?

An interview format using the same standardized, job-relevant questions and a shared scoring rubric for every candidate. Meta-analytic research puts its predictive validity at .51, compared to .38 for unstructured interviews, and .63 when paired with a cognitive ability assessment.

The Precision Gap Assessment was built to answer exactly that question, where a hiring process still assumes a layer of oversight that isn't there anymore. Take a look at the Precision Gap Assessment.