Artificial intelligence is increasingly discussed through the lens of disruption, particularly its potential impact on employment. Yet focusing primarily on the number of jobs that may disappear risks overlooking the more significant economic development taking place. AI is beginning to change how companies organise work, how individuals apply their expertise, where businesses direct capital and, ultimately, how productivity is created. The transition is therefore likely to be less about a simple replacement of labour and more about a gradual reorganisation of the relationship between technology, people and capital.
This distinction becomes clearer when considering how AI is being adopted inside organisations. Most jobs are not made up of a single activity; they consist of a combination of routine processes, judgement, communication, problem-solving and institutional knowledge. AI is particularly effective at compressing the time required for some of the more repetitive elements of this work, from analysing large volumes of information to drafting documents, retrieving knowledge and identifying patterns. As these capabilities improve, the value of AI may increasingly come from allowing people to redirect their time towards activities where experience, context and judgement matter more.
This view is increasingly reflected in labour-market research. The International Labour Organization’s 2025 assessment estimates that one in four workers globally is in an occupation with some degree of exposure to generative AI. Importantly, however, its research concludes that transformation is more likely than widespread replacement because most occupations continue to require meaningful human involvement. That distinction matters. Exposure to AI does not necessarily mean the disappearance of a profession; in many cases, it means that the way the profession is practised will evolve.
Early evidence also suggests that this evolution can produce meaningful gains in productivity. Research by Erik Brynjolfsson, Danielle Li and Lindsey Raymond, based on more than 5,000 customer-support employees, found that workers using a generative-AI assistant increased productivity by approximately 14% on average. The benefits were particularly pronounced among less experienced employees, suggesting that AI can help distribute knowledge and established best practices more efficiently across an organisation. While one study cannot represent every industry, it illustrates a broader possibility: AI can function as an amplifier of human capability rather than simply as a substitute for it.
That possibility becomes particularly significant when considered at an organisational level. Productivity improvements have historically been central to economic growth because they allow businesses to generate greater output from the same underlying resources. AI introduces the prospect of extending that principle across a wide range of knowledge-intensive activities. A financial professional may analyse information more quickly, a software engineer may accelerate parts of development, a physician may receive assistance in processing complex information, and a customer-service employee may draw on a much larger body of institutional knowledge in real time. The value created in each case is not necessarily the elimination of the individual performing the work, but an increase in what that individual can accomplish.
At the same time, greater access to AI does not remove the need for expertise; in many cases, it may make expert oversight more important. AI-generated responses can be convincing without always being complete, accurate or appropriate to the circumstances in which they are being used. A useful parallel can be seen in healthcare. The widespread availability of medical information online has given patients far greater access to knowledge about symptoms and diseases, but it has not removed the need for doctors. Information without clinical context can be misinterpreted, important distinctions can be missed, and individuals may draw conclusions that a trained professional would assess very differently after considering medical history, examination and other evidence. AI presents a similar challenge in many professional settings: generating an answer is not the same as determining whether that answer is correct, relevant or appropriate.
For this reason, blind reliance on AI can itself become a source of risk. As these systems become more capable, the ability to evaluate their outputs may become as important as the ability to generate them. Experienced professionals will still be required to challenge assumptions, verify information, recognise omissions and understand when a conclusion requires deeper investigation. In areas involving financial, legal, medical, operational or reputational consequences, the cost of a plausible but incorrect answer can be significant. AI can accelerate analysis and broaden access to information, but responsibility for interpretation and decision-making ultimately remains with the person or organisation using it.
This has important implications for the skills organisations will value. As routine information processing becomes easier to automate, domain expertise, commercial judgement, creativity, leadership, relationship management and the ability to frame the right questions may become more important. The most effective professionals may not simply be those who know how to use an AI system, but those who understand their field well enough to recognise when the system is right, when it is wrong and what may be missing. In that sense, AI may reduce the value of certain repetitive activities while increasing the value of expertise surrounding them.
The result could be a labour market characterised by considerable movement but also considerable creation. The World Economic Forum’s Future of Jobs Report 2025 estimates that a combination of technological change, demographic trends, economic developments and the energy transition could create approximately 170 million jobs globally by 2030 while displacing around 92 million, producing a net increase of 78 million roles. These figures should not be interpreted as an AI-specific employment forecast, but they are useful in illustrating how significant technological change can generate new occupations at the same time as others decline.
Many of the new areas of economic activity associated with AI are already becoming visible. The development and deployment of increasingly capable systems require semiconductors, computing infrastructure, data centres, electricity, cooling, connectivity and cybersecurity. At the application layer, businesses are developing specialised AI systems for industries ranging from healthcare and financial services to manufacturing, logistics and professional services. Beyond the technology itself, demand is also emerging around implementation, governance, data management, monitoring and the redesign of business processes. What begins as a software innovation therefore has implications across both the digital and physical economy.
This broadening of the AI ecosystem is important because transformative technologies rarely create value in isolation. Their economic impact tends to expand as complementary infrastructure develops and organisations learn how to incorporate the technology into existing processes. The internet, cloud computing and mobile technology followed similar patterns: the underlying innovation was important, but much of the eventual economic value was created by businesses that used the infrastructure to redesign products, distribution and customer experiences. AI may develop in much the same way. The largest opportunities may ultimately emerge not simply from possessing the technology, but from applying it to solve economically significant problems.
There is also an important difference between adopting AI and benefiting from it. Access to increasingly sophisticated models is becoming more widespread, which means technology alone may not provide a lasting competitive advantage. Businesses still need proprietary knowledge, high-quality data, trusted customer relationships, effective distribution and the organisational ability to integrate AI into existing workflows. They will also need appropriate controls and human review to ensure that speed and automation do not come at the expense of accuracy, context or accountability. The companies that create enduring value may therefore be those that combine technological capability with strong institutional knowledge and disciplined human judgement.
The implications extend beyond individual companies. As AI reduces the cost of certain forms of analysis and production, it could enable smaller organisations to access capabilities that were previously available mainly to businesses with substantial financial and human resources. A smaller company may be able to conduct deeper research, automate administrative processes, personalise customer interactions or build software with a much leaner organisation. This has the potential to support entrepreneurship and create new business models, while also increasing competitive pressure on larger incumbents. In this sense, AI may influence not only productivity within companies but also the structure of competition between them.
For investors, AI is therefore better understood as a long-term economic transition rather than a single technology theme. It is likely to reshape routine work while increasing the importance of expertise, judgement and human oversight, alongside creating new sources of productivity and economic activity. The investment opportunity will lie in identifying businesses that can combine technological capability with specialist knowledge and convert that combination into durable improvements in efficiency, decision-making and competitive advantage.
Aceana Group, Insights
