Businesses have invested heavily in making AI available to their employees, but access to technology is only part of the equation. The harder challenge is understanding why some workers are transforming the way they work with AI, while others are struggling to find opportunities to use it. With research showing that 80% of employees want to use AI more than their workplaces currently allow, organisations need to rethink how they approach adoption. Rather than expecting everyone to respond to the same tools and training, businesses must recognise the different motivations, behaviours and barriers shaping how people engage with AI.
The limits of traditional technology rollouts
For years, businesses have relied on familiar ways of dividing up their workforce when introducing new technology. Employees are grouped by department, job function, seniority or experience, with training and support designed around what each group is assumed to need.
While this approach can make large-scale rollouts easier to manage, it overlooks an important reality: employees doing similar work do not necessarily have similar attitudes towards AI adoption. Two people with comparable technical skills may have very different levels of confidence in applying AI to their work.
This matters because making AI available does not guarantee that employees will use it effectively. Some may lack the confidence to experiment, others may struggle to see how it applies to their responsibilities, while more enthusiastic users can find themselves constrained by organisational processes or limited access to the tools they need.
The result is a workforce progressing at very different speeds, even where employees have received the same support. For businesses looking to move beyond isolated productivity gains and embed AI more meaningfully across their operations, understanding these differences is becoming increasingly important.
Understanding the people using the AI
What separates employees is disposition, including how they approach technology, risk and work itself. These underlying characteristics are around six times more predictive of AI behaviour than their position in the workplace hierarchy.
That means looking beyond how people rate their own AI skills and focusing instead on how they behave. This includes their confidence with new technology, willingness to experiment, comfort pursing an outcome when the path is unclear, and motivation to be the first to solve a problem.
Research involving 10,000 workers identified 64 possible combinations of these behaviours, which the analysis groups into seven distinct archetypes. Each has a differing relationship with AI that conventional workforce categories can easily obscure.
The largest groups sit somewhere in the middle. ‘Navigators’ see the value of AI but often encounter barriers such as access and permissions. ‘Stalwarts’ are willing to engage but need greater confidence in how AI applies to their work. ‘Pragmatists’ are capable but cautious, preferring to see others demonstrate its value first, while ‘Holdouts’ remain the least convinced that AI has something to offer them.
Stalwarts and Navigators alone make up nearly three quarters of the workforce. Both represent significant untapped potential, but for very different reasons. One group needs greater confidence and relevance, while the other is already willing but often prevented from doing more. Treating both as an ‘adoption problem’ risks prescribing the same solution to fundamentally different challenges.
Moving from blanket training to tailored support
The first step for businesses is therefore to diagnose the barrier before prescribing the solution. Between 72% and 82% of employees across every archetype say the AI tools available to them do not meet their needs. That means low adoption cannot automatically be solved by buying another tool or putting everyone through the same training programme. Organisations need to understand what is getting in the way, whether that is limited access, poor integration with existing workflows, unclear governance, or simply a lack of useful applications for the work they do.
The response will look different across the workforce. Stalwarts may need space to experiment and build confidence without worrying that getting something wrong will count against them, while Pragmatists may benefit from seeing AI applied successfully to recognisable tasks before changing established ways of working. At the other end, Architects and Accelerators may need greater freedom to test new approaches. Their successes can then become practical examples for colleagues, rather than leaving every employee to work out AI for themselves.
Training should follow the same principle. 81% of the most active adopters who completed AI training considered it effective, compared with 56% of the least engaged. That suggests completion rates alone are a poor measure of progress. Organisations should instead look at what changes afterwards, including whether employees are using AI more confidently, applying it to more valuable work, and overcoming the barriers that previously held them back. They can then use those insights to adapt training around where employees still need support, rather than repeatedly offering the same programme and expecting different results.
Turning AI investment into workforce progress
The next challenge for businesses is ensuring that adoption translates into tangible improvements in how work gets done. That requires a move away from measuring success through access, training participation or usage alone, towards understanding whether employees are genuinely able to do more with the technology.
There will be no single route to achieving this. Some workers will need reassurance and practical examples, while others will benefit from fewer restrictions and greater opportunities to experiment. Businesses that recognise these differences will be better placed to build an environment where AI adoption develops across the workforce rather than remaining concentrated among a relatively small group of confident users.
Ultimately, the organisations that make the most progress will be those that adapt their approach to their people, rather than expecting their people to adapt to a uniform approach to AI.
Duncan Roberts
Duncan Roberts is Associate Director at Cognizant Research. He joined Cognizant in 2019 as a digital strategy and transformation consultant in industries ranging from satellite communications to educational assessment. He has advised clients on utilising technology to meet strategic objectives and discover the art of the possible through innovation.
