I have seen this pattern up close. In conversations with companies positioned to solve AI adoption, I have repeatedly seen the focus land on the technology layer: tool selection, usage visibility, governance, prompt support, and dashboards. Those things matter. But those conversations have reinforced something important for me.
The core problem is not that employees lack access to AI. It is that organizations are treating AI adoption as a software rollout when it is really a workplace change effort.
The pattern is easy to spot. Leaders buy licenses. They compare models. They invest in dashboards, governance layers, and tool stacks. They focus on which platform to roll out, which assistant to standardize on, and how to track usage. The assumption is that if the right technology is in place, adoption will follow.
That view is understandable. AI is new, noisy, and changing fast. There are countless tools, constant feature updates, and real pressure from boards and executives to show progress. In that environment, technology feels concrete. It feels manageable. It feels like action.
But it is also the wrong centre of gravity.
The deeper issue is not whether the organization picked the right model, bought enough licenses, or created a polished dashboard. The issue is whether people are actually changing how they work. The issue is whether managers know how to lead through that change. The issue is whether old workflows, job expectations, and habits are being reworked to fit an AI-enabled workplace. That is not a technology challenge. It is a workplace change challenge.
Why the technology-first view falls short
A technology-first approach usually assumes that AI works like past digital rollouts. Install the tool. Train people on features. Encourage usage. Track adoption. Wait for productivity gains.
That logic may have worked better with earlier workplace technologies because those tools were often additive. The core job stayed largely the same. AI is different. AI changes how work gets done, how judgment is applied, how decisions move, and what managers can expect from people in their roles.
AI does not simply support the old playbook. It reshapes the playbook itself.
That is why organizations can see individual enthusiasm without seeing enterprise results. People may get small personal wins from AI. They may draft faster, summarize faster, or research faster. But enterprise adoption breaks down when teams are still working within old workflows, old measures, old role definitions, and old assumptions about value.
AI adoption is a people management issue
If AI adoption is not mainly a technology problem, what is it?
It is a people management issue.
That means the hard part is not access. It is adoption in context. It is helping people understand when AI fits into their actual work and when it does not. It is helping managers set expectations. It is redesigning workflows so AI is not bolted onto broken processes. It is helping experienced professionals let go of parts of the work that used to define their value. It is giving teams room to unlearn, test, discuss, and rebuild.
This is where many organizations get stuck.
They offer general training. They share a prompt guide. They hold a quarterly session. They track completion.
But completion is not capability. Exposure is not behavior change. Access is not adoption.
The workplace has seen this before in learning and development for years: content gets pushed out, people finish it, and the business assumes something changed when, in reality, very little did.
AI does not fix that pattern. In many cases, it accelerates it.
AI can make it easier to create more content, more job aids, more explainers, and more training materials. But faster content creation does not automatically create better performance. If anything, content overload can reduce retention and make it harder for people to know what matters in their day-to-day work.
That is why the real work of adoption sits with leaders and managers, not just IT or innovation teams.
How to support AI adoption
The idea is simple: organizations are overinvesting in tools and underinvesting in the people and processes that determine whether AI creates value. Most of AI’s value comes from people and process change, not from the technology itself. Leading organizations recognize this and invest far more heavily in training, manager support, and process redesign than many companies do today.
That matters because it changes the question.
Instead of asking, “Which AI tool should we buy?” leaders can ask better questions:
- How does work need to change?
- Which process should be rebuilt first?
- What does this mean for managers?
- What support do people need to use AI well in their real jobs?
- How will we know behavior changed, not just usage?
Those are better questions because they target business reality rather than software selection.
Five ideas are especially useful.
1. Reallocate attention and budget toward people and process change
Organizations often spend too heavily on tools and too lightly on the humans expected to use them. Yet the value comes from people and process redesign, even though many companies still concentrate spending on licenses and infrastructure.
BCG frames this as a 10-20-70 reality: about 10% of AI’s value comes from algorithms, 20% from data and technology, and 70% from people, processes, and cultural transformation.
That shows up in everyday decisions. A company may approve enterprise AI seats quickly but hesitate to invest in manager enablement, workflow redesign, coaching, or role-specific practice. That imbalance sends a message: the technology matters more than the people using it.
In practice, the reverse is closer to the truth.
2. Rebuild workflows instead of adding AI onto bad ones
Bolting AI onto legacy workflows simply makes broken processes move faster. The better approach is to rethink a workflow from the ground up and decide what an AI-native version should look like. McKinsey’s 2025 global AI survey found that, across 25 attributes tested, workflow redesign had the biggest effect on whether organizations saw EBIT impact from generative AI, yet only 21% of respondents said their organizations had fundamentally redesigned at least some workflows.
This is a major shift.
Many organizations still ask where AI can fit into an existing eight-step process. The better question is whether that eight-step process should still exist at all.
Sometimes the right answer is not to automate one part of the old process, but to reduce the entire workflow to two or three steps with new roles for human judgment and oversight.
3. Stop treating adoption as simple upskilling
AI cannot be approached as light upskilling layered onto an unchanged role. In many cases, the old job itself needs to be rethought and rebuilt.
Every role now needs a meaningful digital and AI mindset, not a few prompt tips added to old routines. Harvard Business School’s Tsedal Neeley has argued that, at minimum, everyone now needs a 30% digital and AI mindset: enough fluency to use tools well, ask better questions, interpret outputs, and redesign work.
This matters because many experienced employees are not resisting AI because they dislike technology. They may be reacting to a deeper disruption. The work that once proved their expertise may be changing. Their judgment may still matter, but not in the same form. That creates uncertainty, identity threat, and real management challenges.
Good adoption work takes that seriously.
4. Replace one-off training with weekly enablement
Annual or quarterly training is too slow for AI adoption. Teams need regular, structured conversations where people discuss what is working, what is changing, and where the friction is. Ongoing managerial support and weekly forums can keep teams aligned and learning in real time.
AI adoption is not a one-time launch. It is not a campaign. It is a moving target that requires rhythm.
Teams need space to compare examples, share lessons, adjust workflows, and clarify expectations. An effective adoption plan also needs to appeal to employees’ intrinsic motivation, because people are far more likely to change how they work when they see personal meaning, value, and relevance in that change, not just corporate pressure to use a new tool. Without that rhythm, adoption becomes uneven and shallow. Gallup’s workplace research found that employees whose managers actively support AI use are 9.3 times as likely to say AI has transformed how work gets done in their organization, which helps explain why cadence, manager involvement, and relevance matter so much.
5. Measure behavior change, not just tool usage
Weak metrics can distort the whole effort.
Logins, seat counts, and general usage numbers are easy to gather, but they are poor proxies for meaningful adoption.
Better measures look at whether workflows changed, whether roles changed, whether outputs improved, and whether managers can see a real difference in how work gets done.
That is a far higher bar. It is also the bar that matters.
What this means for leaders
If AI adoption is a people management issue, leaders have to shift from tool rollout to workplace redesign.
That starts with recognizing that employees do not need more noise. They need clarity.
They need to know where AI fits into their role, where it does not, what good use looks like, what risks to avoid, and what standard they are being held to. Managers need to be able to coach this. Teams need shared examples. Processes need to be rewritten. Job expectations need to be updated.
Leaders also have to accept that adoption can feel uncomfortable.
People may worry that the parts of work they are proud of are being devalued. Managers may struggle to redefine expectations. Teams may cling to old processes because they feel safe. These are not side issues. They are central to adoption.
That is why change management cannot sit off to the side as a communication plan. It has to sit inside the day-to-day management of work.
A better way to approach AI adoption
A stronger approach to adoption looks more like this:
- Choose one workflow that matters and redesign it with AI in mind rather than adding AI to the current process.
- Equip managers to lead the shift, not just approve the licenses. Managers are the ones who help teams translate tools into real work.
- Create regular operating rhythms where teams can discuss use cases, friction points, wins, and changing expectations.
- Support unlearning. Do not assume people can calmly let go of old routines just because a new tool exists.
- Measure whether work changed. Do not confuse activity with impact.
This is slower than simply buying software, but it is much more likely to produce real results.
Where Beyond Role Plays fits
This is exactly why we focus on adoption at the people level.
The market does not need more generic enthusiasm about AI. It needs better support for the human side of implementation. It needs practical ways to help leaders, managers, and teams build the judgment and workplace behaviors required to use AI well.
That is where Beyond Role Plays comes in.
We are focused on the part of adoption that many organizations skip: helping people actually change how they work. That means supporting managers, supporting practice, supporting role clarity, and supporting the ongoing human side of adoption rather than assuming the tool will do that work on its own.
Real adoption happens when people can apply new ways of working in context, with support, feedback, and reinforcement. It happens when managers can lead the shift. It happens when organizations treat AI as a workplace change issue, not a software deployment exercise.
Technology still matters. Governance still matters. Security still matters.
But if people do not know how to work differently, none of that is enough.
The companies that get AI adoption right will not be the ones that talked the most about tools. They will be the ones that took people seriously.
Recent Comments