Are We Developing AI Adoption or AI Capability?
A company introduces a new AI tool. Licences are purchased, training is provided and employees are expected to start using it.
A few months later, leadership reviews the adoption figures.
Some people use it regularly. Others have barely engaged with it. The investment has been made, and naturally there are questions about progress.
But increasing usage does not necessarily mean that people are developing the capability to use AI effectively.
And that distinction matters when deciding what to do next.
What are we measuring in AI adoption?
There are several distinctions worth making when discussing AI adoption.
Access: People have the tool available.
Initial use: They have tried it.
Repeated use: They return to it regularly.
Integrated use: It has found a meaningful place in their work.
Value-creating use: Its use improves the work, the experience or the outcome.
Someone using AI every day to summarise meetings may be a regular user. Someone using it occasionally to investigate a complex problem may be creating considerably more value.
Both uses can be worthwhile, but they tell us different things.
If the objective is familiarity, repeated use may indicate progress. If the objective is improved productivity, better decisions or customer experience, the figures need further interpretation.
And with AI, there is another consideration. People may still be discovering what valuable applications are possible in their work.
Where does learning fit into the working day?
With many business systems, adoption begins with people learning a relatively defined set of functions and applying them in their work. More experienced users may explore additional possibilities.
General-purpose AI makes that exploration more accessible across a wider range of activities.
A manager who previously relied on someone else to prepare performance reports may now be able to investigate the data independently. But that requires more than knowing how to operate the tool. The manager needs to formulate useful questions, evaluate the results and recognise what is worth investigating further.
That kind of learning takes initiative and time.
And time is not equally available across roles.
Take, for example, an advisor working in a contact centre. Handling consecutive customer contacts leaves limited freedom to experiment during the working day. A manager may have greater control over their schedule but face constant operational demands.
A business can encourage people to experiment with AI while its workload and priorities communicate that time should be spent on activities producing immediate results.
This creates a familiar tension. Managers postpone developing their people because operational problems need attention. Yet developing those people could reduce the need for management intervention in future.
The opportunity cost of learning is visible immediately. The opportunity cost of not learning often appears later.
The issue extends beyond time. Someone accustomed to closely prescribed procedures and limited discretion may need support and experience before feeling comfortable exploring independently.
There is also the question of how learning is organised.
Formal training can establish essential foundations, but introducing applications without sufficient opportunity to practise them may leave much of that learning unused. Capability develops as people apply what they have learned to real work, encounter difficulties and discover where the technology is genuinely useful.
That suggests a place for learning that continues beyond formal training: knowledgeable colleagues sharing experience, managers helping people apply AI to real work, and practical insights becoming part of wider team conversations.
This also makes use of existing expertise and management routines, rather than expecting formal training to carry the full burden of capability development.
Learning also needs a way to travel.
A leader discovers a useful way to analyse performance data with AI. The next management meeting is filled with operational updates, leaving no opportunity to demonstrate the approach.
The discovery remains individual.
Making space in existing management conversations to share applications and lessons from experimentation can help turn what one person has learned into wider capability, within appropriate operational and security boundaries.
If people are expected to discover better ways of working, what space are we creating for that learning to happen?
What happens to the capacity AI creates?
One of the most common arguments for adopting AI is that it can improve productivity and help people accomplish more in less time.
That is a legitimate business objective. But what does it mean for the people being asked to adopt it?
Consider an advisor who already handles customer contacts throughout a demanding shift. Management explains that AI will help the team process more cases.
From the organisation's perspective, this represents an opportunity to improve productivity. From the advisor's perspective, it may sound like an expectation to handle even more work.
The same message can therefore create different expectations.
An advisor might welcome AI because it reduces repetitive administration, makes difficult cases easier to resolve or creates more time to give customers the attention they need. Another might anticipate tighter targets, a more intensive working day or uncertainty about the future of their role.
When we explain the benefits of AI, are we describing something employees can recognise as valuable in their own work, or mainly something the organisation hopes to gain from their increased productivity?
The capacity question applies to managers too. If AI reduces time spent preparing reports, analysing information or completing administrative tasks, does that capacity become available for coaching and improvement, help make an existing workload more manageable, or simply become absorbed by additional operational responsibilities?
Each outcome may have commercial value, but each creates a different experience of the technology and different opportunities for capability development.
Creating capacity through AI is one decision. What we choose to do with that capacity is another.
And that second decision may influence how willingly people engage with the first.
What do management responses encourage?
Suppose adoption figures remain lower than expected.
Some people may lack confidence or struggle to see where AI could help. Others may be experiencing change fatigue. The tool may be poorly suited to certain tasks, people may have had little opportunity to practise, or they may be uncertain about what greater productivity will mean for their workload or their role.
These situations can produce similar figures but require different responses.
If low usage is interpreted as reluctance, managers may be asked to reinforce expectations. That may help people overcome hesitation and discover useful applications.
It may also lead them to demonstrate that they are using the tool.
Encouraging initial use can be necessary, particularly when people have yet to discover what the technology can offer. The question is whether that encouragement results in meaningful learning and application.
The next report shows higher usage.
Has the technology become more useful? Have people developed greater capability? Or have they become more attentive to what management expects them to demonstrate?
The response to an adoption figure becomes one of the influences shaping the next figure.
If raising difficulties is repeatedly interpreted as reluctance, people may also become more selective about what they share.
Leadership can then receive increasingly reassuring figures while losing access to some of the experiences needed to understand what is happening.
Management responses also shape professional judgement.
An advisor may recognise that an AI recommendation follows the rules but overlooks an important aspect of the customer's circumstances. Adjusting the response could reflect exactly the discretion the role requires.
Yet that decision may leave little evidence in an adoption dashboard.
How the manager responds matters. Exploring the reasoning can reinforce that kind of judgement. Concentrating solely on why the recommendation was not followed may teach a different lesson.
People learn what is genuinely expected through the way their initiative, difficulties and decisions are handled.
What kind of capability are we developing?
A business may be developing compliance with AI use, capability in using AI, or some combination of the two.
The commercial difference can be substantial.
Using AI to complete an existing task more quickly may deliver efficiency gains. Developing the capability to question existing approaches and solve problems with AI may create opportunities that were previously overlooked.
In sales, AI can help someone prepare a proposal in half the usual time. It can also help them challenge an assumption about what the customer needs before writing it. Both count as AI use, but the second could change the direction of the customer conversation.
The same technology can produce very different results depending on people's ability to identify valuable applications, question its output and apply sound judgement.
Efficiency and broader problem-solving capability both have value, but they call for different expectations and forms of support.
Recognising opportunities to improve work also requires understanding the work itself, including whether a process should be automated or rethought.
Adoption figures remain a useful starting point for discussion. They can show whether people are engaging with the technology, but understanding the capability being developed requires looking at how AI is being applied, what people are learning and whether it is improving outcomes.
Examining examples of how people use AI to improve an actual task, solve a problem or make a decision can reveal aspects of capability that usage figures alone cannot show.
Understanding capability also requires looking at how management expectations and working conditions influence the way people engage with the technology.
Which brings us to a final question.
How much of the adoption we're seeing reflects people's experience of the technology, and how much reflects their experience of how we're asking them to adopt it?
Beyond AI Adoption: The Leadership Dimension
The introduction of AI can reveal more than an organisation's readiness for new technology. It can also bring into view how leadership expectations, established management practices and everyday working conditions influence people's willingness and ability to adapt.
These are among the interconnected dimensions explored through our Leadership Capability Pathways, which address how judgement, learning, contribution and performance develop across different organisational contexts.
Themes explored
Technology adoption · Capability development · Workplace learning · Employee voice · Performance measurement
Comments