Organisations seeking to deploy artificial intelligence too often focus on the technology itself rather than on how AI is embedded in daily work processes. That is the central finding of doctoral research by Inès Baer at Vrije Universiteit Amsterdam, who examined the practice of AI management within organisations.
Although AI is being applied more widely across both public and private organisations, many projects fall short of expectations or produce unexpected consequences. Baer demonstrates that these problems stem not so much from technical shortcomings, but from the way organisations manage and integrate AI systems into their operations. Her research drew on both a literature review and extensive fieldwork conducted within a public employment services organisation.
A key conclusion is that managing AI is not a one-time implementation project but an ongoing process. Organisations must continuously weigh technological possibilities against organisational needs and societal consequences. AI systems regularly create new opportunities, but also produce unforeseen effects that only become visible once the systems are in active use. Simply drafting an AI strategy in advance and then executing it is therefore insufficient.
According to Baer, success depends on continuous collaboration among a range of specialists, including managers, IT professionals, policymakers and end users. Organisations must also be willing to adapt systems when circumstances or needs change. As AI is increasingly used in decision-making, for example within government bodies, employment services and public service delivery, the need to handle the consequences of this technology with care continues to grow.
Baer emphasises that investment must go beyond software and algorithms to include training, governance and ongoing dialogue between technical experts and users. Without such attention, AI systems risk causing confusion, introducing new risks or producing undesirable effects for employees and citizens. Baer advocates a broader approach in which technology is not viewed in isolation from the people, organisations and societal context in which it operates. The findings offer practical guidance for managers, policymakers and technology professionals seeking to deploy AI more effectively and responsibly, while accounting for public values and societal interests.