Companies are pouring staggering amounts of money into artificial intelligence. IDC projects global spending will surpass half a trillion dollars within the next two years. Boardrooms are talking about it, tech vendors are promising it, and learning and development teams are feeling the pressure to show they’re ‘doing something with AI.’
And yet, the results are almost invisible. MIT recently reported that 95 percent of AI projects fail to deliver measurable outcomes. Despite the unprecedented investment, productivity gains are elusive, employee adoption is shaky, and the business case often collapses under scrutiny.
How can we surround ourselves with the most powerful technology in human history, spend billions deploying it, and still struggle to prove it makes us better? The answer isn’t hiding in the models or the code. The real story and the real risk are sitting right in front of us.
Advance both, or fail at both
Executives often frame AI transformation as a choice: is this a technology problem or a people problem? In reality, it’s both, and treating it as anything less is a recipe for failure.
Many companies treat AI like a software upgrade. They focus on integration, licenses, and technical training only to find that adoption stalls and employees quietly revert to their old ways of working. Others swing toward culture and change management, ignoring the fundamental skills required to use AI effectively. Unfortunately, they underplay the reality that AI isn’t just another tool. It’s the most powerful technology we’ve ever placed in human hands.
That’s why leaders can’t afford either/or thinking. If you only push the tech, you’ll be part of the 95 percent of failures. If you only focus on people, you’ll underestimate what’s at stake. Success requires advancing both at the same time, with equal intentionality.
I saw this tension firsthand when a company preparing to deploy Microsoft Copilot asked me to help drive adoption.
Case study: Beyond “how-to”
The client was a global healthcare company with tens of thousands of employees worldwide. They were rolling out Copilot as part of their Office 365 roadmap and had engaged me to create a change management strategy. Their expectation was clear: focus on the technical “how-tos,” pair it with a structured communication timeline, and ensure people felt informed along the way. In short, ensure the rollout felt orderly and smooth.
On paper, that plan made sense. However, in the very first strategy meeting, I posed a question that shifted the conversation: “Do you know how your people are already using AI today and how effectively they’re using it?” The room went quiet. Everyone knew employees had access to AI tools, but no one had any data on how those tools were being used or misused.
Identifying that gap was a turning point. There was an immediate recognition that what we were trying to accomplish wasn’t limited to adoption or “how to.” It was about equipping people to make wise decisions with AI. Instead of focusing narrowly on technical training, we began with an assessment designed to show employees where they stood today, highlighting their gaps, and pointing them toward development opportunities. The aggregate results offered insight into where to focus support for the company. For employees, it was a personalized roadmap to grow with the technology, not just learning to click the right buttons or type the right prompt.
There was resistance at first. The word “assessment” triggered fears: Would this be used to evaluate performance? Was it a backdoor to replacement? However, when employees saw how the process was framed as developmental, not punitive, trust grew. People realized the company wasn’t trying to extract more from them but signaling an investment in their future.
When Copilot launched, the difference was striking. Employees were prepared to be effective. Surveys showed people felt confident using the tool and believed the company genuinely cared about their ability to thrive in an AI-enabled workplace. Adoption exceeded expectations, satisfaction scores held steady, and innovation began bubbling up as teams discovered new ways to apply Copilot beyond the initial roadmap.
In the end, success wasn’t about luck or timing. It came from reframing adoption from a technical rollout to a capability-building strategy. It required balancing the “how-to” with the skills and judgment people would need to use AI effectively.
Why 95% fail, and why this worked
Now, it’s worth acknowledging this healthcare company’s story is the exception, not the rule. MIT’s research and most people’s anecdotal experience make that abundantly clear. However, those failures aren’t caused by flawed software. They’re caused by flawed framing.
Most leaders are still thinking about AI as a technical shift. They’re trying to flip the switch, train employees on features, and check the box. However, there’s a big difference between creating awareness and changing behavior, a responsibility that falls squarely on L&D leaders. Unfortunately, adoption metrics often mask the deeper reality that people may know how to use the tool but don’t know when or why to use it effectively. Without judgment, integration, and discernment, the technology sits idle. Or worse, it amplifies mistakes.
And here’s the part that surprises many executives: building capability doesn’t require adding months to the timeline. In the healthcare company’s case, the adjustments to the rollout plan were statistically insignificant. The launch went forward largely as scheduled. The difference was that when it went live, people weren’t just compliant; they were confident. The impact wasn’t in the timeline; it was in the outcomes.
This is the decision point every organization must face. You can move fast and hope for the best, or you can invest slightly more up front to avoid joining the ninety-five percent of failures. Speed without capability is a race to the bottom. The companies that succeed will be the ones that treat AI adoption not as an IT project but as a leadership challenge.
AI as amplifier
A message you won’t typically hear in the headlines is that AI isn’t going to magically solve our leadership challenges. It was only ever going to expose them. And, AI in particular doesn’t erase human strengths or weaknesses; it magnifies them. When people are equipped, AI accelerates progress. When they’re not, it multiplies the chaos.
That’s why the future of AI isn’t a race to roll out the next tool or upgrade to the latest model. It’s a race to prepare people to use whatever they have well. The companies that win will be the ones that recognize success is not about features. It’s about capability, and that’s where learning and development has the most critical role to play. You have to go beyond completion and communication plans and focus on sustainable effectiveness.
AI is here for good, and it’s extremely powerful. The stakes are higher than anything organizations have faced before. However, don’t ever forget the story of AI won’t be written by the algorithms we build but by the humans we choose to become.
Try Big Think+ for your business
Engaging content on the skills that matter, taught by world-class experts.
