The real work of AI transformation (Part 2)
Most organisations are experimenting with AI. But experimentation is not transformation.
Research shows that many employees now have access to AI tools, yet far fewer use them regularly. The challenge is rarely the technology. It is how organisations are designed.
In Part 2 of my three part series, I explore what organisations must rethink if AI is going to deliver real value.
This includes how we design work, how teams collaborate, and how leaders create cultures where people can experiment and learn.
Technology can enable transformation. But people and leadership make it real.
Part 2 of the series: Designing for the Age of AI
In the first article in this series, Designing for the Age of AI, I explored why artificial intelligence represents more than a technological shift. Like previous technological revolutions, AI will only deliver its full impact when institutions, industries, and organisations redesign themselves around new capabilities.
This second article moves from the system level to the organisational level. If AI is becoming embedded across the workplace, what does transformation actually require from leaders, teams, and structures?
The answer is rarely just about technology. The deeper challenge is organisational change.
Why technology adoption rarely creates transformation
History offers a clear lesson. Organisations often adopt new technologies long before they redesign their systems around them.
During the early decades of electrification, many factories replaced steam engines with electric motors but kept the same production layouts. Only later did managers realise that electricity made it possible to redesign factories with flexible, decentralised production lines. When factories reorganised around electric power, productivity surged dramatically (Brynjolfsson & McAfee, 2014).
Digital transformation followed a similar pattern. Early internet adoption involved putting existing services online. Over time, entirely new business models emerged, including digital platforms, online marketplaces, and subscription services.
Artificial intelligence is now entering this same phase.
According to Deloitte's State of AI in the Enterprise research, most organisations are currently using AI to improve efficiency within existing processes. Much less attention has been given to redesigning workflows, roles, and decision-making structures to fully integrate AI into the organisation (Deloitte, 2026).
Research on organisational change reinforces this concern. Studies suggest that many transformation programmes struggle because leaders focus heavily on technology and strategy whilst neglecting the human and organisational foundations needed to sustain change (McKinsey & Company, 2021). The technology advances faster than the organisations using it.
The activation gap
One of the most revealing findings in recent AI research is what many analysts describe as the activation gap.
Many employees now have access to AI tools, yet far fewer use them consistently in their work. Deloitte's global research indicates that roughly sixty per cent of employees in many organisations have access to approved AI tools, yet only around half of those employees regularly incorporate them into their daily workflows (Deloitte, 2026).
This gap highlights an important insight. Access to technology does not automatically change behaviour.
Employees need clarity about how AI should be used, how it fits into their work, and what expectations leaders hold about its role in decision-making and productivity. Without that clarity, new tools often sit on the sidelines.
This is not primarily a technology problem. It is a leadership and culture problem.
The dual heart of transformation
One of the most consistent insights emerging from both research and practice is that successful transformation balances two priorities simultaneously.
The first is customer value. Organisations must improve products, services, and experiences to remain competitive.
The second is people capability. Employees must learn new skills, adopt new behaviours, and develop confidence in new ways of working.
In my earlier article The Dual Heart of Transformation: Navigating Customer and People-Centred Change, I described this balance as the intersection between customer-centred and people-centred change. Technology often draws attention to the first objective, but sustainable transformation depends just as much on the second.
Artificial intelligence makes this tension more visible, not less.
AI tools can dramatically improve efficiency, insight, and automation. Yet the people using those tools must understand when to rely on algorithms, when to challenge them, and how to interpret their outputs responsibly. Without investing in people capability, organisations risk deploying technology that employees do not fully trust or understand.
Research on digital transformation consistently supports this conclusion. Successful initiatives depend on leadership alignment, workforce engagement, and organisational learning as much as on technical implementation (Warner & Wäger, 2019; Verhoef et al., 2021).
Technology can enable transformation. People make it real.
Rethinking how work is structured
Artificial intelligence challenges one of the most fundamental assumptions underlying organisational design: the idea that roles are defined by tasks.
For more than a century, organisations have structured jobs around task specialisation. Workers perform specific activities within well-defined roles, and managers coordinate those tasks across teams. AI disrupts this model because machines are increasingly capable of performing many routine tasks faster and more accurately than humans.
This shift suggests that organisations will increasingly define work not by tasks but by capabilities and outcomes.
Human capability will centre on areas such as judgement, creativity, ethical reasoning, and relationship building. AI capability will focus on analysis, pattern recognition, and large-scale information synthesis. When these capabilities are combined effectively, productivity and innovation can increase significantly.
Research from a controlled study of knowledge workers found that AI tools increased both the speed and quality of outputs in areas such as writing and analysis. Crucially, the strongest results occurred when workers redesigned their workflows around AI collaboration, rather than simply using it as an occasional tool (Noy & Zhang, 2023).
Human and AI collaboration
As AI systems become more capable, organisations will increasingly operate through hybrid human and machine collaboration.
In these environments, humans and AI systems perform complementary roles. AI can rapidly process large amounts of data, identify patterns, and generate options. Humans provide context, judgement, and ethical oversight.
Managers therefore shift from supervising human labour to orchestrating a system that includes both human and machine contributors. This may appear a subtle change, but it represents a significant shift in leadership practice.
Leaders must learn to evaluate AI-generated insights, understand the limitations of algorithms, and guide teams in deciding when to rely on automated recommendations and when human judgement should prevail.
New organisational structures
As AI capabilities expand, new organisational structures are beginning to emerge.
Some organisations are establishing small internal teams responsible for orchestrating AI across the enterprise, focusing on integrating data, tools, and workflows across departments. Others are experimenting with cross-functional teams where domain experts, analysts, and technologists collaborate closely, using AI tools to support rapid problem solving.
These structures reflect a broader shift away from rigid hierarchies toward more adaptive organisational networks.
Research on digital transformation has shown that organisations navigating technological change must develop new dynamic capabilities, allowing firms to sense emerging technologies, experiment with new approaches, and reconfigure resources quickly in response to change (Warner & Wäger, 2019). Artificial intelligence accelerates the need for exactly these capabilities.
Culture and psychological safety
Whilst organisational structures matter, culture often determines whether transformation succeeds.
Technological experimentation inevitably involves uncertainty. People learning to work with AI tools will sometimes generate incorrect outputs, misinterpret results, or choose ineffective approaches. Organisations that punish mistakes find that employees avoid experimentation altogether.
Research on organisational learning emphasises the importance of psychological safety in enabling innovation. Teams are more likely to experiment and share knowledge when they feel safe raising questions, admitting uncertainty, and testing new ideas (Edmondson, 2018).
When leaders model curiosity and thoughtful risk-taking, teams are more likely to explore new possibilities. When leaders emphasise rigid control or perfection, innovation slows. Artificial intelligence makes this cultural dimension more important, not less.
Transformation as an organisational capability
The most important shift many organisations must make is moving from viewing transformation as a temporary project to treating it as a permanent organisational capability.
Digital technologies, including AI, ensure that change will continue to occur. Organisations that thrive in this environment do not simply manage occasional transformation programmes. They design systems that allow them to continuously adapt. This includes building leadership capability, strengthening governance structures, and developing cultures that support experimentation and learning.
Artificial intelligence may be the technology that accelerates this shift. But the underlying challenge is not technological. It is organisational.
Looking ahead
Artificial intelligence is often discussed in terms of algorithms, automation, and technological capability. Yet the real work of transformation happens within organisations. It involves leadership, culture, governance, and the design of systems that allow people and technology to work together effectively.
In the final article in this series we will widen the lens again. If organisations must redesign themselves to harness AI, governments and industries must also adapt. The next article explores how public policy, industry ecosystems, and responsible governance will shape the future of the AI economy.
References
- Brynjolfsson, E., & McAfee, A. (2014). The second machine age: Work, progress, and prosperity in a time of brilliant technologies. W. W. Norton & Company.
- Deloitte. (2026). State of AI in the enterprise: The untapped edge. Deloitte Insights.
- Edmondson, A. C. (2018). The fearless organization: Creating psychological safety in the workplace for learning, innovation, and growth. Wiley.
- Kotter, J. P. (1995). Leading change: Why transformation efforts fail. Harvard Business Review.
- McKinsey & Company. (2021). Why transformations fail and how they succeed.
- Noy, S., & Zhang, W. (2023). Experimental evidence on the productivity effects of generative artificial intelligence. Science, 381(6654), 187–192.
- Verhoef, P. C., Broekhuizen, T., Bart, Y., Bhattacharya, A., Dong, J., Fabian, N., & Haenlein, M. (2021). Digital transformation: A multidisciplinary reflection. Journal of Business Research, 122, 889–901.
- Warner, K. S. R., & Wäger, M. (2019). Building dynamic capabilities for digital transformation. Long Range Planning, 52(3), 326–349.