New research from Stanford Graduate School of Business has put numbers to something that many business owners are already sensing. Tech workers are leading AI adoption at 60% frequent usage, but growth has stalled. After a significant surge between 2024 and 2025, the numbers plateaued. The tools did not change. The people did. And that is the part most AI strategies are not built to handle.
Access Is Not the Same as Readiness
When organisations roll out AI tools, the instinct is to measure success by uptake. How many licences are active. How many people logged in this week. How many prompts were sent. These numbers climb fast in the early stages because novelty drives curiosity. People try something new. Some find it useful. Most figure out one or two things it can do for them and stop there.
That initial climb gets reported upward as a win. And then the numbers flatten. Not because the tools stopped working. Because the people were never given what they needed to go further.
Access and readiness are two completely different problems. Access means your team has the tool. Readiness means your team understands why it matters for their role, feels supported in building skills with it, and has been given the psychological safety to experiment without fear of looking incompetent. Most AI rollouts solve for access. Almost none solve for readiness. The Stanford research is showing us exactly where that gap lives.
The Ceiling Is Human Shaped
At 60% frequent usage among tech workers, AI adoption is not low. But it is stuck. And the reason it is stuck is not a software problem. No update, no new feature, no additional licence tier is going to move that number. The ceiling is built from something else entirely.
It is built from the team member who uses AI to draft one type of email and nothing else because nobody ever showed them what else was possible. It is built from the manager who installed the tool but never had a conversation with their team about how it fits into existing workflows. It is built from the quiet anxiety that has never been named or addressed. The fear of getting it wrong. The worry about what happens when the tool gets good enough to do more than drafting emails. The unspoken question underneath every AI rollout. What does this mean for me?
When those things go unaddressed, people find their level and stay there. That is not resistance. That is a completely rational human response to change that has not been properly explained or supported.
The Three Things Most AI Strategies Skip
The Stanford research reinforces what is visible in how AI adoption actually plays out across businesses. There are three things that consistently get skipped, and each one contributes to the plateau.
The first is the why. Not the business case why, but the personal why. Your team does not need to know that AI will improve productivity by some projected figure. They need to know what it means for their specific role. Whether it makes their day harder or easier. Whether the skills they have built over years still matter. That conversation almost never happens before the tool lands in their inbox.
The second is structured support. Handing someone a tool and pointing them to a help article is not training. Real confidence with AI comes from practice, feedback, and the ability to ask questions without feeling like you should already know the answers. Most rollouts treat training as a one hour onboarding session. That is not enough for something that is genuinely changing how work gets done.
The third is time. Businesses expect people to integrate AI into their workflows while maintaining full output at the same time. That is not realistic. Building genuine capability with any tool requires experimentation, which means some things will not work the first time. If there is no space for that, people default to doing things the way they always have and using the AI tool only when it is obvious and low risk. That is how you get a plateau.
What Responsible AI Adoption Actually Looks Like
The businesses that are moving past the plateau are not necessarily the ones with better technology. They are the ones that treated AI adoption as a people project first and a technology project second.
That starts with a genuine conversation before any tool is purchased or deployed. Not a town hall where leadership announces the direction and asks for questions. An actual conversation where people can express what they are worried about and get honest answers. What will change. What will not. What support will be available. What the organisation’s commitment is to the people whose roles are shifting.
It continues with readiness assessment. Understanding which parts of your operation are ready to adopt AI and which parts need more groundwork first. Understanding which team members are enthusiastic early adopters who can bring others along and which ones need a different kind of support. Understanding where the friction points are before they become blockers.
And it requires ongoing investment in capability, not just initial onboarding. AI tools are evolving. What your team learns to do with them today will need to be updated and expanded as the tools change. That is not a one time training exercise. It is a continuous commitment.
The Real Question to Ask About Your Team Right Now
The Stanford research is useful because it gives the AI adoption plateau a number to point at. But the more important thing it confirms is something that many business owners have already felt without being able to name it. The tool was adopted. The tool stopped being used to its potential. The business does not know exactly why.
The question worth sitting with is not whether your team is using AI. It is whether they understand why the organisation is adopting it, whether they feel genuinely supported in building their own capability with it, and whether they have been given the space and the time to actually get good at it rather than just getting by.
If the honest answer to any of those is no, that is where the work is. Not in finding a better tool. Not in sending another adoption reminder. In going back to the people and starting the conversation that should have happened first.
The technology is the easy part. The people are the hard part. The Stanford research has now confirmed what most AI rollouts reveal the hard way. The ceiling is human shaped, and no software update reaches it.
Get in touch at www.xsiv.au/#form