C-suite leaders are placing billion-dollar bets on AI with near-unanimous confidence. Recent reporting from the Wall Street Journal and Fast Company reveals a stark disconnect: executives see efficiency gains, while employees report anxiety, feeling overwhelmed, and spending hours each week fixing AI errors.

Here's what makes this dangerous: AI genuinely saves time—Larridin research shows 10+ hours weekly for skilled users on tasks like first drafts, data analysis, coding, and graphics. But extracting that value requires setting context, making judgments about accuracy and tone, and knowing what you're trying to learn or show—capabilities that come from experience and daily coaching by managers.

Companies are simultaneously cutting middle-management layers (the people who provide that coaching) and eliminating entry-level positions (the people who need that coaching). A new study finds that 58% of C-suite executives report no clear ownership of AI initiatives, and 75% lack proper AI governance. Yet these same leaders remain "highly confident" in AI ROI.

The gap between executive confidence and operational reality isn't merely a perception issue. It's driving irreversible workforce decisions—cutting talent pipelines, eliminating the management layer that develops judgment—before companies understand what makes AI investments pay off.

Companies are trading long-term enterprise viability for short-term earnings. They're eliminating the coaching and development infrastructure exactly when people need it most. And they won't realize the damage until it's irreversible.

For CEOs:

• 82% of CEOs are more optimistic about AI than last year, yet 58% have no clear AI ownership and 75% lack governance
• Entry-level hiring has plunged 35% since January 2023—companies are cutting the talent pipeline before knowing if AI can actually do the work
• Only 19% of C-suite executives report AI saving them 12+ hours weekly; 40% of individual contributors say AI saves them zero time
• The optimism gap isn't just perception—it's driving premature layoffs that destroy institutional knowledge companies can't rebuild

For CMOs & CROs:

• AI genuinely saves time—research shows 10+ hours weekly for skilled users on first drafts, analysis, coding, graphics
• But capturing that value requires context-setting, judgment, and coaching that comes from experienced managers
• Companies are cutting middle management (the coaching layer) while expecting juniors to magically know how to use AI effectively
• 77% of workers say AI tools are unreliable, 50%+ report confidently incorrect outputs—they're using tools without the judgment to evaluate results
• The real crisis: eliminating the people who develop judgment while scaling tools that require judgment to use well 

The Reality Check:

• Senior engineers are burning out as entry-level positions disappear, with the assumption "AI will do it"
• AI accelerates code velocity but can't design systems, test comprehensively, or work with stakeholders
• Companies that cut people before understanding what AI can actually do are creating skill gaps they may never recover from
• The fundamental question isn't "Can AI replace this role?" but "What happens when AI fails and there's no one left who knows how to think?"

Image generated with Gemini

The gap isn't subtle. And it's driving decisions with permanent consequences.

Entry-level job postings have dropped 35% since January 2023, Fast Company reports. Companies like Amazon expect to avoid hiring 160,000 people by 2027 through AI-driven automation. BT announced plans to cut 55,000 jobs and replace 10,000 with AI. Goldman Sachs and other investment banks are exploring replacing entry-level presentation work and data entry.

These are irreversible workforce decisions being made by leaders who admit they lack governance, ownership, or a detailed understanding of AI capabilities.

The disconnect isn't about who's right. It's about what happens when executive confidence drives decisions faster than operational reality can support them.

Isaac is 33, a mid-level software development engineer at a Big Tech firm. He spoke to Fast Company under a pseudonym to avoid retaliation.

At the start of 2025, entry-level engineering postings at his company dropped dramatically. The work didn't vanish—it got redistributed to senior staff with the assumption that AI would make up the difference.

"AI can straight up write better, faster, more legible code than most developers," Isaac admits. "But any seasoned engineer knows the hard part isn't writing the code, it's the design and testing."

Fewer people to delegate design and testing work to. Fewer managers to coach people on how to think about design and testing. Facebook's decision to cut middle-management layers became the blueprint: reduce overhead, let AI boost individual productivity, and capture the savings.

But middle managers weren't overhead. They taught juniors how to think strategically, provided context for decisions, caught errors before they became problems, and developed the judgment that makes AI tools valuable.

Senior developers now do both the work and the coaching—poorly, because they're drowning. "Seniors are burning out," Isaac says, "and when they leave, there's no rush to replace them, because 'the AI will do it!'"

Stanford research shows employment for 22- to 25-year-old software engineers fell nearly 20% between late 2022 and July 2025, even as hiring for older engineers grew.

 Entry-level jobs aren't just about execution. They're the training ground where people develop pattern recognition, learn to see what matters, build relationships that become institutional memory. Critically, they're where people learn from managers who coach them daily—showing them what good looks like, how to spot problems, when AI output is brilliant versus confidently wrong.

Eliminate entry-level roles, cut middle management, and you destroy your talent pipeline and the coaching system that makes AI productive.

Matt Beane, author of "The Skill Code," points this out: the expert-novice approach to skill-building has existed for 160,000 years. AI is disrupting it in less than three years, and companies are treating that disruption as an efficiency gain rather than a risk.

What happens in five years when senior engineers burn out and leave, and there's no one behind them who knows how to think? AI will write code. No one will know if it's solving the right problem. There will be no managers left to train them.

Here's what the confidence gap overlooks: AI delivers real productivity gains. Larridin research shows experienced users saving 10+ hours weekly. AI excels at first drafts, data analysis, coding, graphics, slide creation—the execution work that used to consume days.

But achieving a 10-hour productivity gain requires skills that most companies are eliminating.

To extract value from AI, you need to:

• Provide rich context about the problem you're solving
• Know what you're looking to learn or show before you start prompting
• Exercise judgment about accuracy, tone, and appropriateness
• Recognize when AI output is brilliant versus confidently wrong

Studies show the cost when people lack this judgment. Carnegie Mellon found AI coding agents fail basic tasks 70% of the time on complex, existing codebases. AI customer service agents produce incorrect information more than 50% of the time, according to Asana research. The METR study found that AI coding tools slow down experienced developers because time spent correcting errors exceeds the time saved.

These capabilities—context-setting, judgment, pattern recognition—don't come from using AI tools. They come from experience and coaching. Not professional coaching, but the daily coaching that happens when managers develop their people.

And companies are systematically eliminating that coaching layer.

Facebook cut middle management. Others followed. The logic seemed sound: AI makes individual contributors more productive, so you need fewer managers. Cut the overhead, boost short-term earnings.

But middle managers weren't overhead. They were the people who taught juniors how to think, provided context for decisions, caught errors before they became problems, and developed the judgment that makes AI tools valuable instead of dangerous.

Without that coaching layer, Asana research shows employees spend 4.5 hours weekly fixing AI errors. Mark Hoffman, Asana's Work Innovation Lead: "AI can make work look faster on the surface, but it can also create a lot of cleanup work—double-checking outputs, correcting errors, and redoing steps that were based on faulty information."

The cycle accelerates: Cut managers to save money. Junior people lack coaching to develop judgment. They can't effectively use AI because no one teaches them what good looks like. Productivity suffers. Senior people burn out trying to fill the gap. More cuts follow.

One Reddit post about a company halting entry-level engineering hiring received hundreds of responses from people experiencing the same pattern. One commenter: "Not sure what the plan will be after the knowledge transfer is over."

The question C-suite confidence isn't asking: AI saves 10 hours weekly for people with experience and judgment. What happens when you eliminate the people who develop that judgment, stop hiring people who could learn it, and expect AI to bridge the gap?

People need to learn how to work effectively with AI. And AI still has a lot to learn. There's no shortcut through this transition—it requires the coaching and development infrastructure companies are cutting to boost quarterly earnings.

The most telling finding isn't about AI capabilities. It's about leadership readiness.

The 2026 study finds that 75% of companies lack AI governance frameworks and 58% have no clear ownership of AI initiatives. Meanwhile, 81% of C-suite executives believe their company has a "clear, actionable policy for AI guidance." Only 28% of individual contributors agree.

Research from The Conference Board reveals that CEOs identify AI simultaneously as a top investment priority, a leading external risk, and a governance concern. Yet they're accelerating investment while hoping governance catches up.

30% of global CEOs (38% in the US) identify AI as the leading factor that could negatively affect their business in 2026, ranking it above political polarization. They see the risk. They're investing anyway. Without governance. Without ownership.

This is how Salesforce happened. Marc Benioff bet big on AI capability, fired 4,000 people, then admitted six months later they "massively overestimated AI's capabilities." Customer satisfaction plummeted, institutional knowledge was lost.

But Salesforce is learning publicly. Most companies will make the same mistakes privately—discovering too late that confidence isn't a substitute for readiness, and that institutional knowledge can't be rebuilt with better prompts.

The optimism gap reveals a deeper failure: most C-suite leaders are asking the wrong question.

They're asking: "Can AI do this job?"

The right question is: "What human capabilities become exponentially more valuable when AI handles commodity work?"

When AI writes code, human skills in system design, stakeholder management, and strategic problem diagnosis become more important, not less. When AI automates customer support, human skills in reading account risk, navigating escalation politics, and building trust become irreplaceable.

Companies winning with AI understand this. They're not replacing expertise with automation. They're using AI to scale expertise, make judgment accessible, and accelerate learning instead of replacing it.

One strategic accounts leader—we'll call him Marcus—managed $110M in recurring revenue by creating an AI knowledge base trained on his frameworks, deal recordings, and decision-making patterns. His team could ask: "What's Marcus's playbook for handling a renewal with a skeptical CFO and technical debt issues?"

The AI didn't make decisions. It made Marcus's expertise accessible when he couldn't be in the room. The judgment calls—reading customer politics, knowing when to push and when to wait, adapting to communication styles—those were still human.

Marcus's region outperformed every other territory. Not because AI replaced thinking, but because it leveraged thinking that already worked.

Compare that to companies eliminating entry-level roles. They're not scaling expertise. They're destroying the pipeline that creates it.

1. You see the productivity gains but not the coaching gap. The confidence disconnect between C-suite and individual contributors signals that people doing the work see problems leadership doesn't—specifically, they lack the judgment and context to use AI effectively without daily coaching.
2. You're underinvesting in the capabilities AI requires. Strategic communication, diagnostic thinking, influence without authority, cultural intelligence—these determine whether AI investments create value or chaos. And they require the coaching infrastructure companies are eliminating.
3. You're making irreversible decisions based on incomplete understanding. Cutting entry-level positions, eliminating middle management, and restructuring departments can't be undone. Once institutional knowledge and coaching infrastructure are gone, you're starting from zero.

Companies that succeed with AI in 2026 won't be the most optimistic. They'll be the most disciplined:

• Clear ownership before expanding investment
• Governance frameworks before scaling deployment
• Systematic coaching and development during transition
• Understanding that both people and AI have a lot to learn

The alternative is becoming another Salesforce—discovering too late that you eliminated the people who knew how to do the work and the managers who could teach others.

Most CEOs are treating AI as a technology investment. The winners are treating it as a leadership and development investment.

AI commoditizes execution. That makes human skills in judgment, influence, and strategic thinking more valuable—but only if you invest in developing them. And that requires the coaching infrastructure companies are cutting.

The executives winning with AI aren't replacing human expertise. They're amplifying it. Using AI to scale judgment, make institutional knowledge accessible, and accelerate capability development.

But that requires investing in capabilities most companies have neglected:

• Strategic diagnosis – Seeing patterns AI can't recognize, knowing which details matter
• Influence without authority – Getting buy-in, building trust, enrolling people in change
• Cultural intelligence – Reading organizational dynamics, adapting communication, creating psychological safety
• Organizational design – Restructuring for effectiveness when AI changes what's possible

And critically, it requires the daily coaching that develops these capabilities. Not professional coaching—manager coaching. The kind Facebook eliminated to boost quarterly earnings.

Companies making these investments aren't cutting leadership development budgets while adding AI budgets. They're doubling down on both—because they understand that AI without people who know how to use it is just expensive chaos.

The question isn't whether AI will transform your organization. It will.

The question is whether you're investing in the coaching and development infrastructure that determines if that transformation creates value or destroys it.

If you're a CEO, CMO, or CRO making AI investment decisions right now, you're facing pressure: boards want ROI, investors want efficiency, competitors are moving fast.

That pressure creates a dangerous temptation—to cut costs and boost short-term earnings by eliminating "overhead" like middle management and entry-level positions.

But the companies that will exist in five years aren't optimizing for next quarter. They're building the coaching and development infrastructure that makes AI valuable.

They understand there's no shortcut through this transition. People need to learn how to work effectively with AI. And AI has a lot to learn. That requires the patient, daily coaching from managers who develop judgment, provide context, and teach people what good looks like.

You can't prompt-engineer institutional knowledge. You can't automate judgment. You can't replace decades of pattern recognition with dashboards.

But you can invest in developing the capabilities AI makes more valuable. You can build governance before scaling deployment. You can preserve the coaching infrastructure that helps people navigate this transition.

Most companies are betting on AI. The smartest ones are investing in the people and coaching infrastructure that determine whether those bets pay off.

More GTMFlow Blogs on AI and Leadership

• Why AI is making great leadership more valuable

• Want career security? Start investing in being an AI‑Integrated Leader

• AI and Leadership: Using AI to Control vs. Empower People