AI Business Strategy: Why Leadership Development Closes the ROI Gap

Companies are spending on AI like it’s oxygen. Budgets are up, transformation teams are expanding, and every vendor promises competitive advantage. The results tell a different story. A 2024 Gartner survey of 4,200+ business and technology leaders found only 48% of digital initiatives met or exceeded their business outcome targets. A 2025 BCG global survey was even more stark: 60% of respondents said their AI investments delivered little material value in revenue or cost reduction.

The pattern is consistent across industries: AI activity is visible, but defensible ROI is not.

The gap isn’t technical. It’s developmental. As leadership transformation expert Dr. Vivian Atud puts it: “AI does not create ROI—leadership does.” Technology can accelerate performance, but only when leaders know how to direct it. That’s a leadership development problem, and it requires a leadership development solution.

The Leadership Gap Behind AI’s Broken Promises

Most AI adoption follows the same trajectory. Individual teams adopt copilots and automation tools. IT procures platforms. Executives approve budgets based on vendor promises. But nobody owns the question of whether any of it actually improves the business.

This is a governance vacuum, and it exists because leadership development programs haven’t caught up to the technology cycle. Leaders are being asked to direct AI strategy with capabilities built for a pre-AI operating environment.

Nobel Prize-winning economist Daron Acemoglu argues the problem runs deeper than most leaders realize. In a recent conversation with MIT Sloan Management Review, he challenged the assumption that AI automatically improves productivity. His core argument: current incentives push AI toward automation and centralization by default. Without deliberate leadership, organizations replace human judgment with automated processes that look efficient but destroy the adaptive capacity that makes organizations resilient.

The distinction matters. There’s a difference between using AI to eliminate tasks and using AI to amplify human capability. The first saves money in the short term. The second builds competitive advantage. But making that distinction requires a kind of strategic judgment that most leaders haven’t been asked to develop.

Three Capabilities AI-Ready Leaders Need

The capability gap has three dimensions. Each one is developable through focused coaching and deliberate practice.

Strategic AI Judgment

Not every process benefits from automation. Leaders need the ability to distinguish between work that is task execution and work that requires strategic judgment. Acemoglu’s framework is useful here: AI should complement human skills where judgment, creativity, and relationship-building matter most. The leader who automates customer service without understanding which interactions build loyalty is optimizing the wrong metric.

This isn’t about becoming a technologist. It’s about developing the capacity to ask the right questions before approving the next AI initiative.

Innovation Risk Governance

Monica Caldas, global CIO of Liberty Mutual Insurance and recipient of the 2025 MIT Sloan CIO Leadership Award, models what this looks like in practice. Speaking with MIT Sloan Management Review about deploying generative AI, she described an approach that balances genuine experimentation with disciplined risk management. She came to Liberty Mutual after 17 years at GE leading transformation projects, and she brings that same problem-solving orientation to AI: push innovation forward while understanding what you’re putting at risk.

Too many leaders treat AI adoption as binary. Either they block it (missing opportunity) or greenlight everything (accumulating unmanaged risk). Developing governance capability means learning to hold both: moving forward while staying honest about what you don’t yet understand.

Cross-Functional AI Fluency

A recent MIT Sloan Management Review article made a point that many executives are missing: agentic AI tools extend far beyond software development. Features that make AI coding tools valuable for programmers are equally applicable to knowledge work that involves no programming at all. Leaders across functions need enough technical literacy to direct AI strategy, not just delegate it to the IT department and hope for the best.

This doesn’t mean every executive needs to write prompts. It means they need to understand what AI can and can’t do well enough to make informed decisions about where to invest, what to pilot, and when to pull back.

Building AI Leadership Through Coaching

These three capabilities share something in common: none of them develop through slide decks, webinars, or vendor demonstrations. They develop through reflective practice under pressure.

Coaching is built for exactly this kind of development. A coaching engagement focused on AI leadership readiness helps leaders examine their assumptions about technology, test strategic judgment through real decisions, and build the discomfort tolerance required to govern something they don’t fully control.

Consider a leader who keeps approving AI pilots but can’t articulate what success looks like for any of them. The coaching conversation isn’t about AI tools. It’s about accountability, measurement discipline, and the willingness to kill projects that aren’t delivering. Those are leadership capabilities, and they transfer far beyond the AI conversation.

The organizations closing the AI ROI gap aren’t the ones with the best technology. They’re the ones developing leaders who know how to use it.