Why AI Adoption Fails: The Individual Choice Problem Worth Solving
Most AI adoption programs fail for the same reason most technology rollouts have failed for the last twenty years, and it has nothing to do with the technology. People adopt change one person at a time, by choice, or they don't adopt it at all. You can mandate a tool. You cannot mandate a decision that happens inside someone's head.
This came up in a recent conversation on my podcast, AI Literacy for Entrepreneurs, with Grace Gravestock, a change management leader who spent over 20 years running large-scale technology projects, including work on the $67 billion Dell EMC integration after 17 earlier attempts had failed. Her read on why AI pilots stall matched something I'd been circling for a while: leadership treats AI adoption as a technology rollout. It's actually a psychology problem wearing a technology costume.
The Root Cause Isn't Training
When AI adoption stalls, the instinct is to add more training. More workshops. Another vendor demo. A refresher session.
Grace's read, from two decades of watching this exact pattern across ERP systems, CRM rollouts, and now AI: "It comes even before we get to training. The root of it is really having the buy-in on an individual level for people for the change." Training assumes people have already decided to change. Most haven't. You can train someone extensively on a tool they've privately decided not to fully use, and they'll attend every session and change nothing about how they actually work.
The bottleneck isn't skill. It's a choice that hasn't been made yet, and no amount of training addresses a choice.
The Three Steps People Actually Go Through
Grace's framework for change, which she's building into a book called Three Secrets to Making Change Fun and Easy, breaks the individual adoption process into three steps: choice, alignment, and openness to new ways of working. (Get on her book waitlist and a see short video on the three secrets here: funandeasychange.com)
Choice comes first, and it's not optional to skip. People need to feel they've decided to adopt AI, not been told to. Grace pointed to a consistent pattern in her large-scale projects: the biggest resistance almost always comes from the people who've been at the organization longest, and specifically from the subject matter experts whose buy-in everyone else is watching for. If they hold back, the people around them read that hesitation as permission to hold back too.
Alignment comes second. Choosing to adopt AI only sticks if it connects to where someone sees their own future going. Grace described this from her own experience: she bought a Mac Mini specifically to commit to using AI more seriously, and it sat unused for months. Buying the tool wasn't the choice. Deciding her future included working with AI, rather than around it, was the actual step, and that took time to settle.
Openness to new ways of working comes last, once the first two are in place. This is the part that looks like adoption from the outside; actually using the tool, building new workflows. But it only holds if the first two steps happened first. Skip to step three without steps one and two, and you get exactly what most companies see: people going through the motions in training, then quietly reverting to their old workflow the moment nobody's watching.
The Fear Underneath the Resistance
There's a specific fear driving a lot of this, and it's more direct than most leadership teams want to admit. During the conversation, I raised something that came up repeatedly in earlier research we'd done talking to leaders about AI adoption: people are afraid they're training their own replacement.
Leadership usually addresses this by simply denying it: "That's not going to happen." Which isn't totally true. It's a dismissal, and smart employees, the ones you most need adopting AI well, notice when a real fear gets waved away instead of addressed.
Grace's says: AI will likely replace a lot of "the worker bee stuff". The people who keep their jobs and thrive are the ones who learn to orchestrate rather than execute. That's not a comfortable message to deliver for most leaders, but it's a more honest one than blanket reassurance, and honest messaging is what actually earns trust during a change process.
What Leadership Communication Gets Wrong
Grace's second core point from two decades of change projects: none of this works if leadership isn't visibly bought in first. "The leaders need to be bought into the change. Nothing else can happen if the leadership is not aligned with the goal. People feel that, they see it, they know it."
But leadership buy-in alone isn't enough. The messaging also has to answer one specific question for every group affected, what Grace calls the WIIFM radio station: what's in it for me. Not the company. Not the department. The individual person being asked to change how they work.
Grace shared a concrete example from a past project: the turning point wasn't better training or a bigger budget. It was tying the project's outcome directly to senior executives' bonuses. Suddenly the leaders driving the initiative had something personal at stake, and that personal stake showed up in how visibly and genuinely they championed it.
(Further reading: How to Prove AI ROI to Your Leadership Team (Before They Cut the Budget))
What This Means for Your AI Rollout
If your AI adoption has stalled despite reasonable training and a real budget, the fix probably isn't a better vendor or a longer workshop. It's addressing three things Grace's framework and two decades of project data point to directly:
Name the fear instead of dismissing it. If your team is worried AI threatens their role, "that's not going to happen" isn't reassurance, it's a signal that leadership isn't taking the question seriously. Address it honestly: some tasks will change, and the people who learn to work alongside AI rather than against it are the ones who come out ahead.
(Further reading: How to Stop AI From Replacing your Team (And Use It to Grow Instead))
Get your longest-tenured experts on board first, individually. They're the ones everyone else is watching. If they hold back publicly or privately, that hesitation spreads through the team faster than any training program can counteract it.
Answer WIIFM for every group, not just the company as a whole. What does this change actually mean for a specific person's day-to-day work, their time, their advancement? A rollout that only talks about company-wide efficiency gains will lose the room.
(Further reading: How CEOs and CMOs Should Lead AI Change Management)
The Short Version
AI adoption fails when organizations treat it as a technology deployment instead of a series of individual decisions. People need to choose to adopt AI, see how it aligns with their own future, and only then become open to genuinely new ways of working. Skipping straight to training without addressing choice and alignment produces attendance, not change.
If leadership doesn't visibly buy in first, and doesn't answer what's in it for each specific group being asked to change, the rollout will stall regardless of budget or tool quality.
You can hear the full conversation with Grace Gravestock on episode 275 of AI Literacy for Entrepreneurs, where we go deeper into her three-step framework and what it looks like in practice on large-scale change projects.
If you want to see where your organization's AI adoption is actually stuck, the buy-in stage, the training stage, or somewhere else, the NorthLight AI Readiness Audit gives you a structured picture in about 15 minutes. Run the audit now.
