We are firmly in the digital age, awash in data generated on every surface and in every layer of every business. Yet, despite decades of investment in technology, time, and effort, many organizations are still not seeing meaningful returns. A global survey of over 4,200 business and technology leaders conducted by research firm Gartner in late 2024 found that only 48% of digital initiatives had met or exceeded targeted business outcomes. Other surveys have had similar results, and the view is even more discouraging when studies focus on AI investments. For instance, 60% of respondents to a 2025 global survey from BCG said their investments in AI have delivered little material value, either in increased revenue or lowered costs.
We’ve been building toward this digital age for well over 30 years — so, what’s the problem? In 2020, Linda A. Hill, Sunand Menon, Ann Le Cam, Karina Grazina, and Lydia Begag of the Leadership Initiative at Harvard Business School began studying what it takes to lead in the digital era. Six years on, their research data reveals a consistent pattern: Leaders who are more successful with digital transformation tend to place early and sustained emphasis on building what Hill’s team calls a “digitally dexterous” workforce — one in which employees are both willing and able to use new technologies to advance the organization’s goals.
The research offers strong evidence that leadership attention to culture, learning, and skills plays a central role in whether digital investments pay off. As Hill and her coauthors explain in “Why Digital Dexterity Is Key to Transformation,” this is a long-term leadership commitment that begins with reframing the challenge itself.
This issue isn’t limited to advanced technologies like AI. Even in areas that have long relied on data, such as customer experience, success depends less on how much information is available than on leaders’ ability to decide what deserves attention. Having hundreds of different metrics to evaluate customer experience might seem like an advantage, but managing and applying insights from all of them can be difficult, according to authors Charles H. Patti, Maria M. van Dessel, and Steven W. Hartley. If your company is collecting more CX data points than you can make good use of, you’ll want to read “A Smarter Approach to Measuring Customer Experience.”
Nowhere is the need for human judgment clearer than in organizations’ rush to adopt AI, particularly large language models and generative tools. These technologies can be powerful, but only when employees are able to test and contextualize their outputs. Research conducted by Steven Randazzo, Akshita Joshi, Kate Kellogg, Hila Lifshitz, and Karim R. Lakhani reveals a potentially disturbing trend: When research participants tried to push back on and validate an LLM’s conclusions, the AI pushed back even harder, using a variety of techniques to actively persuade users to accept its results. In “Validating LLM Output? Prepare to Be ‘Persuasion Bombed,’ ” the authors show how easily the technology can overpower human judgment — not because it’s “smart,” but because users lack the knowledge, confidence, or skills to challenge it effectively.
Organizations fail at digital transformation not because their technology is weak but because they haven’t prepared their people to use it well.