The Hidden Cost of AI-Enabled Leadership: What Efficiency Can't Replace

As artificial intelligence transforms how leaders work, three overlooked challenges deserve our attention: loneliness, the difference between efficiency and support, and the diminishing role of constructive challenge.

I've been thinking a lot about artificial intelligence and executive coaching lately, but not because I'm worried about AI replacing coaches. I think we're asking the wrong question.

AI is changing how we work, access information, and make decisions. I use it every day in my own business, and I'm genuinely excited about what it makes possible. It has made me more efficient, expanded my access to information, and allowed me to accomplish things that previously required considerably more time and resources.

But as someone who spends much of her professional life working with executives, business owners, and leadership teams, I've become increasingly interested in something that isn't receiving nearly enough attention.

We have more access to information, answers, and technological support than at any point in history. But are we actually feeling more supported?

And perhaps more importantly, what happens to leadership development when the technology making us more capable also makes it easier to operate without involving another human being?

I don't believe AI is inherently making leaders lonelier. There isn't sufficient evidence to support that conclusion, and for some people, AI may provide a valuable source of reflection and support that wasn't previously available.

What concerns me is something more subtle: the possibility that we're becoming extraordinarily good at functioning independently without recognizing what we may be losing in the process.

I see three interconnected issues worth examining: leadership loneliness, the growing distinction between efficiency and meaningful support, and the potential for AI to reinforce the echo chambers that already exist around people in positions of authority.

1. AI Is Changing How We Experience Leadership Loneliness

There is an uncomfortable reality about leadership that doesn't receive nearly as much attention as it deserves. As responsibility increases, the number of people with whom a leader can be completely candid often decreases.

Executives are expected to demonstrate confidence while navigating uncertainty. They're responsible for decisions that affect employees, families, customers, and shareholders, often without the benefit of someone with whom they can openly process their doubts.

There are things leaders cannot appropriately share with their teams, conversations they may be reluctant to have with their boards, and challenges that friends or family members may not fully understand. The higher someone rises within an organization, the more complicated those relationships can become.

This isn't only an observation from my coaching practice.

Gallup's State of the Global Workplace: 2026 Report found that 31% of leaders experienced loneliness a lot of the previous day, compared with 21% of individual contributors (Gallup, 2026).

Nearly one in three leaders. Loneliness isn't simply an uncomfortable byproduct of leadership. It has implications for organizational effectiveness.

A systematic review and meta-analysis of 49 studies found that workplace loneliness was associated with elevated burnout, lower job performance, and reduced job satisfaction (Bryan et al., 2024). While these findings are correlational and don't establish causation, they demonstrate that loneliness is meaningfully connected to outcomes organizations care about.

This is where I think the conversation needs to move beyond individual well-being.

When leaders are isolated, overwhelmed, or operating without meaningful opportunities for reflection, the implications may extend beyond the individual. The quality of their relationships, their ability to support others, and the environment they create for their teams all deserve consideration.

And now we're introducing technology that makes it possible for those same leaders to accomplish even more without necessarily interacting with another person.

To be clear, I don't think every interaction with AI replaces a human conversation. In many cases, it may simply replace an internet search, an administrative task, or work someone would otherwise have done alone.

But I do wonder whether the convenience of having an always-available thinking partner will make leaders more or less intentional about maintaining the human relationships they need.

If leadership was already lonely, are we designing a future of work that addresses that isolation, or simply makes it easier to function within it?

2. More efficient doesn't necessarily mean more supported

Microsoft's 2025 Work Trend Index found that 80% of the global workforce reported lacking the time or energy to complete their work (Microsoft, 2025).

I find that statistic fascinating because we've never had more tools designed to make work easier.

We can automate workflows, manage projects, generate content, analyze information, organize communication, and now use artificial intelligence to assist with increasingly sophisticated cognitive tasks.

Yet, many professionals still feel overwhelmed by the demands of their work.

AI has enormous potential to address some of that pressure. It can eliminate repetitive tasks, improve access to expertise, and create capacity that previously didn't exist. I'm an enthusiastic user of those capabilities, and I don't think we should minimize their benefits.

But I also wonder whether we're confusing the ability to accomplish more with the experience of being better supported.

An executive can now prepare for a difficult conversation, evaluate strategic alternatives, draft a communication plan, or work through a complicated decision without ever consulting another person.

That's remarkably efficient.

But the ability to prepare for a difficult conversation doesn't necessarily make someone more willing to have it. Access to a decision-making framework doesn't automatically resolve the uncertainty surrounding a consequential choice. And having a sophisticated tool available at any hour doesn't necessarily address the experience of carrying significant responsibility alone.

Sometimes the problem isn't a lack of information. It's a lack of perspective, accountability, or a safe place to acknowledge uncertainty.

I've worked with enough leaders to know that many already understand what effective leadership requires. They know they need to delegate more effectively, address underperformance, establish boundaries, or stop avoiding a difficult conversation. Their challenge isn't necessarily knowing what to do. It's understanding why they aren't doing it.

That's an important distinction for organizations investing in both technology and leadership development.

If AI helps us complete work faster, do we use the capacity it creates to think more deeply, build stronger relationships, and become better leaders? Or do we simply fill that capacity with additional demands?

The Microsoft findings don't establish that AI is causing overload. They do, however, raise an important question about whether our investments in efficiency are translating into a better experience of work.

I suspect the distinction between having access to support and genuinely feeling supported will become increasingly important as more of our professional lives move into digital environments.

And that brings me to the third issue, which I find particularly concerning.

3. AI is making the leadership echo chamber more sophisticated

Leadership already has a built-in echo chamber.

As authority increases, candid feedback can become more difficult to obtain. Employees may hesitate to challenge decisions, colleagues may soften uncomfortable truths, and information can become filtered as it moves through organizational hierarchies.

The result is that leaders may receive more agreement at precisely the point in their careers when they most need constructive disagreement.

Now introduce an AI tool that is available around the clock, possesses an extraordinary breadth of information, and can help a leader develop a compelling rationale for almost any decision.

That's an incredibly useful resource, but it also introduces a problem.

AI systems can exhibit sycophancy, a tendency to affirm a user's beliefs or interpretations rather than appropriately challenge them.

In research published in Science, Cheng et al. (2026) found that AI models affirmed users' actions more frequently than human respondents in the scenarios examined. Their experiments also found that interacting with sycophantic AI could increase users' confidence in their own judgments while reducing their willingness to take responsibility and repair interpersonal conflicts.

The research wasn't conducted specifically with executives, so we shouldn't assume the findings translate directly into organizational leadership. Nevertheless, the implications are worth considering.

Imagine a leader describing a conflict with an employee. The leader presents the situation as they understand it, explains why they believe their response was justified, and asks AI for help determining what to do next.

The system can only evaluate the information it receives, and that information may be incomplete, emotionally charged, or unintentionally biased. If the response reinforces the leader's interpretation, they may leave the interaction feeling more confident without becoming any more self-aware.

That's not necessarily development. It may simply be a more sophisticated form of confirmation.

AI can absolutely be used to challenge assumptions, identify alternative perspectives, and expose weaknesses in an argument. I use it that way myself, and I think leaders should.

But there is a difference between asking a tool to challenge you and having a trusted person recognize when you need to be challenged, particularly when you haven't asked for it.

The latter requires context, judgment, trust, and a willingness to introduce discomfort into a relationship.

And that discomfort is often where meaningful leadership development begins.

Growth doesn't come from being right all the time. It comes from recognizing where our thinking, behavior, and assumptions may be limiting us.

When leaders increasingly turn to technology to process their most complicated decisions, I think we need to consider whether they're receiving the kind of challenge that produces growth or simply becoming more efficient at defending what they already believe.

Where does coaching fit into all of this?

I don't believe the future of executive coaching depends on proving that AI cannot coach. In fact, I expect AI to change the profession significantly.

If a coach's primary value is providing information, introducing generic leadership frameworks, or asking questions that can be generated in seconds, AI will make that value proposition increasingly difficult to defend.

And it should.

But the most meaningful coaching relationships I've experienced aren't built around information transfer. They're built around the ability to examine what's happening beneath the surface of a leader's decisions and behavior.

Sometimes that means recognizing a pattern across several conversations. Sometimes it means noticing the contradiction between what someone says they value and how they're actually leading. Other times it means asking a question that makes someone uncomfortable enough to reconsider an assumption they've carried for years.

A 2023 meta-analysis of executive coaching studies found positive effects across several outcomes, with particularly strong results for behavioral change (Nicolau et al., 2023). A separate meta-analysis of randomized controlled studies also identified significant positive effects of workplace coaching across leadership and personal outcomes (de Haan & Nilsson, 2023).

That distinction matters because knowledge has never been cheaper, but behavior change remains difficult.

Most experienced leaders don't need another explanation of why delegation matters or how to approach a difficult conversation. They need to understand why they continue behaving in ways that undermine the outcomes they're trying to achieve.

Technology can contribute meaningfully to that process. AI coaching tools are improving, and I expect them to become increasingly effective at facilitating reflection, offering structured feedback, and supporting professional development.

But I also believe there is something distinctly valuable about a relationship with someone who knows your context, remembers what you said three months ago, recognizes when you're avoiding an issue, and cares enough about your development to disagree with you.

That relationship takes time to build. It requires trust, consistency, and a willingness to engage in conversations that may not produce immediate or comfortable answers.

In a business environment increasingly optimized for speed and efficiency, those qualities may become more valuable, not less.

Coaching isn't the only place they exist. Mentors, peers, colleagues, and healthy organizational cultures can provide essential opportunities for connection and challenge. Professional coaching serves a particular purpose within that broader ecosystem, especially for leaders who need confidential space to examine complex decisions and persistent behavioral patterns.

The opportunity isn't to replace AI with human interaction or to suggest that technology cannot provide meaningful support.

It's to recognize that different forms of support serve different developmental needs, and that organizations should be intentional about how they invest in both.

The question organizations should be asking

As businesses accelerate their adoption of artificial intelligence, I hope we begin evaluating its implications for leadership with the same seriousness we bring to productivity, efficiency, and return on investment.

We're investing heavily in technologies designed to make leaders more capable. Are we investing with the same intention in the relationships, developmental experiences, and organizational environments that make them better?

Are leaders gaining perspective, or becoming increasingly comfortable with their existing assumptions? Are they experiencing meaningful support, or have we simply given them more sophisticated tools to manage everything independently? And as we continue removing friction from how work gets done, are we being careful not to eliminate the constructive friction that helps people grow?

I don't think the answer is to resist technological progress. The benefits of AI are too significant, and the opportunity to make work more effective is too compelling.

But I do think we need to be deliberate about what we preserve as our professional lives become increasingly digital.

For much of modern business history, information and access to expertise were expensive. Technology has fundamentally changed that equation. Answers are increasingly accessible, and the cost of obtaining them continues to decline.

Attention, trust, meaningful relationships, and the willingness to challenge another person don't scale the same way.

Perhaps that is where the real opportunity lies for leadership development and coaching in an AI-enabled workplace: not in competing with technology for answers, but in helping leaders develop the judgment, self-awareness, relationships, and behavioral capabilities necessary to use those answers well.

AI may make us more productive, more informed, and more capable of working independently. Those are extraordinary advantages, but they don't automatically make us better leaders or leave us feeling more supported.

And as organizations continue investing in the technologies that will shape the future of work, I think we should be equally intentional about investing in the people expected to lead through it.

AI can make answers cheaper, but it doesn't necessarily make leadership less lonely, and it certainly doesn't eliminate our need for other people.

As AI becomes one of your most trusted thought partners, who in your professional life still knows you well enough (and has your permission) to tell you when you're wrong?

References

Bryan, B. T., Andrews, G., Thompson, K. N., Qualter, P., Matthews, T., & Arseneault, L. (2024). Loneliness in the workplace: A mixed-method systematic review and meta-analysis. Occupational Medicine, 73(9), 557–567. https://doi.org/10.1093/occmed/kqad138

Cheng, M., et al. (2026). Sycophantic AI decreases prosocial intentions and promotes dependence. Science, 391(6792), Article eaec8352. https://doi.org/10.1126/science.aec8352

de Haan, E., & Nilsson, V. O. (2023). What can we know about the effectiveness of coaching? A meta-analysis based only on randomized controlled trials. Academy of Management Learning & Education.https://doi.org/10.5465/amle.2022.0107

Gallup. (2026). State of the global workplace: 2026 report. https://www.gallup.com/file/workplace/707798/state-of-the-global-workplace-2026-download.pdf

Microsoft. (2025, April 23). The 2025 annual Work Trend Index: The Frontier Firm is born.https://blogs.microsoft.com/blog/2025/04/23/the-2025-annual-work-trend-index-the-frontier-firm-is-born/

Nicolau, A., Candel, O. S., Constantin, T., & Kleingeld, A. (2023). The effects of executive coaching on behaviors, attitudes, and personal characteristics: A meta-analysis of randomized control trial studies. Frontiers in Psychology, 14, Article 1089797. https://doi.org/10.3389/fpsyg.2023.1089797