Why Midmarket CEOs Should Rethink the ‘More AI + More Layoffs’ Equation

Across the corporate world, a familiar refrain echoes through boardrooms: Flatten the organization, streamline decision-making, and let artificial intelligence pick up the slack. 

The story sounds progressive — an agile pivot to the future. But for CEOs in the $250 million to $1 billion revenue range, it’s time for a reality check: The numbers don’t justify the frenzy.

Flattening in the Name of Efficiency

Large enterprises are aggressively cutting middle layers, claiming it will drive speed and eliminate bureaucracy. Target said its recent restructuring that removed nearly 8% of its corporate workforce was an attempt to “simplify operations.” Amazon announced roughly 14,000 corporate-role eliminations tied to “efficiency gains and AI integration.” 

The logic seems sound: fewer layers, faster decisions, leaner cost structure. But without a deep redesign of decision rights, workflows, and accountability, flattening can paralyze execution. Companies that skip this step often end up redistributing the same work across fewer people, creating hidden costs in burnout, rework, and strategic drift.

Midmarket CEOs must ask: Are we redesigning for the future or just following a fashion that Fortune 50 companies can afford to botch and survive?

We Have Seen This Movie Before

During the pandemic, firms went on historic hiring sprees under the belief that digital demand would permanently spike. Then came the reversal.  

Stripe’s founders admitted, “We over-hired for the world we are in,” before cutting 14% of staff. Marc Benioff, CEO of Salesforce, acknowledged the same misstep. And Meta’s “year of efficiency” was a euphemism for undoing its own overexpansion.

The lesson isn’t that layoffs are wrong, it’s that acting on groupthink is expensive. But there’s another side to the story. Companies that hesitated to adjust early burned cash and then swung the axe harder later. The real skill lies in calibrated adaptation, not speed for its own sake.

What the 2025 Data Actually Says About AI

Here’s where new evidence from SHRM’s 2025 data brief on Automation, Generative AI, and Job Displacement Risk in U.S. Employment helps separate myth from math. The study, based on 20,000 U.S. workers, found:

• 15.1% of U.S. jobs are at least half automated.

• 7.8% of workers perform half or more of their tasks using generative AI. 

• Yet, only 6% of jobs (9.2 million) are both highly automated and free from nontechnical barriers (such as client preference, regulation, or cost-effectiveness) that prevent displacement.

In plain English: The true displacement risk from AI is less than half of what headlines suggest. Most jobs are being transformed, not eliminated. Even in high-tech fields such as software engineering and finance, AI supplements human judgment far more than it replaces it.

This confirms what many CHROs have quietly observed: The hype around AI replacing people is running far ahead of operational reality. For CEOs, that’s a flashing caution light before turning reorganization into religion.

The Hidden Costs of Hollowing Out the Middle 

Middle management isn’t just a structural layer — it’s the organization’s connective tissue. Deloitte’s 2025 Global Human Capital Trends report found that fewer than 4 in 10 managers feel equipped for the people side of transformation. Yet, those same managers are the ones translating strategy, coaching new leaders, and managing the exceptions that no algorithm handles well.

SHRM’s data brief adds texture. It noted that many of the “nontechnical barriers” preventing AI displacement are relationship-based. Client expectations, trust, and tacit knowledge often determine how work gets done. Those are precisely the areas where seasoned managers excel.

Yes, some firms have excess bureaucracy. But when companies indiscriminately remove midlevel talent, they trade agility for instability. Institutional memory, culture transmission, and informal problem-solving networks vanish. The potential result: Decision latency goes up, not down.

Counterexamples: CEOs Who Refuse the Panic Button 

While tech giants chase AI efficiency, several CEOs are bucking the trend. Infosys publicly committed to avoiding mass layoffs, instead retraining over 250,000 employees on AI and automation tools while bringing in new graduates to sustain its leadership pipeline. Hologic’s management called large-scale layoffs a “failure of leadership,” preferring to utilize redeployment and temporary labor during downturns.

These leaders are not anti-efficiency. They’re simply applying sequencing discipline and investing in readiness before reorganization. In their view, cutting people before systems mature is like removing the wings to make a plane lighter.

That approach doesn’t translate perfectly to every midmarket business, but the principle holds: Stability and skill building usually beat volatility and slogans.

AI Reality Check: Productivity Frontier vs. Human Judgment 

It’s true that certain workflows — such as customer service chats, invoice processing, and code generation — show clear productivity gains from AI. McKinsey & Company’s The State of AI in 2025 survey noted that these early adopters tend to have robust data governance and standardized processes.

But those examples are the exception. Most midmarket companies still lack clean data pipelines, consistent knowledge management, and cross-functional accountability systems. SHRM’s findings reinforce this: Even highly automated roles often remain insulated from displacement due to social, ethical, and compliance factors.

So, the real CEO question isn’t “Can AI do this work?” but “Can AI do it reliably, legally, and culturally in our context?” Until the answer is yes, layoffs tied to “AI transformation” are speculative at best.

A CEO Playbook for 2026: Strategic Alternatives to a ‘Cut First’ Policy

As you enter 2026, here’s a disciplined framework for balancing efficiency with resilience.

1. Run an AI-readiness audit before making cuts. Map core use cases, required data, governance, and expected payback. If the return on investment is modeled but unproven, hold staffing steady until pilots deliver two consecutive quarters of validated results.
2. Redeploy and upskill managers into value roles. Identify managers who can transition into process redesign, data stewardship, or AI supervision. Keeping institutional knowledge close accelerates adoption and protects continuity.
3. Protect the succession spine. Pause reductions in leadership feeder roles until the new structure has proven stable. A thinned-out pipeline may save margin today but will cost years of leadership capacity later.
4. Time-box pilot projects and scale what works. Approve 90-day experiments with predefined thresholds on cost, quality, and service-level improvement. Expand only when those metrics are sustained.
5. Reset spans of control and layers with math, not rhetoric. Benchmarket optimal team sizes for your industry, regulatory exposure, and complexity. If a new span leaves insufficient room for coaching or risk management, the flattening is cosmetic.
6. Communicate like an owner, not a trend follower. Be explicit about your rationale, metrics, and review schedule. SHRM research showed that when employees perceive transparency, productivity rebounds faster after restructuring.

Some CEOs will argue that waiting until 2026 is too slow. If your AI readiness is high and results are verified, then you should move faster. But if your systems and culture lag behind your ambition, speed is not a virtue — it’s a liability.

The Deeper Lesson: Transformation, Not Trend

SHRM’s data brief ended with a reminder that transformation will touch far more jobs than it eliminates. Automation is not replacing work, it’s reshaping it. The smartest CEOs will respond not by cutting first, but by re-architecting how human and machine intelligence coexist.

AI can handle structured processes. People can handle ambiguity, trust, and change. Success in this next era depends on how well you integrate both, not on how many managers you can remove before the market opens on Monday.

The bottom line: If you are a CEO weighing layoffs in the name of “AI readiness,” take a pause. The data has shown that fewer than 1 in 16 jobs is truly automatable in both technical and practical terms. The rest depend on human judgment, relationships, and creativity — the very capabilities you risk erasing. 

The winners of 2026 will not be those who flatten fastest, but those who align technology with talent, evidence, and judgment.

Bob Goodwin is the president of Career Club, where he works with senior executives to elevate leadership performance and align business strategy with people outcomes.