Most companies have tried it, lots of companies have invested heavily in it, but few companies seem to derive serious value from it. Why is this the case?
Why? Because too many leaders are trying to avoid adapting and instead use AI to maintain the status quo. They approach AI as another change programme – another application to add to the desktop, to undertake endless proofs of concept, and they think that training everyone will be enough.
This creates the bizarre situation where an AI is used to increase bureaucracy – for example making long tedious AI reports on useless meetings – rather than using AI to stop needing the meetings.
Pointless tasks
Companies already harbour huge numbers of what my late London School of Economics colleague David Graeber called “bullshit jobs” – pointless tasks of little value. Using AI to create ever more is crazy. We need to stop.
The instinct of most leaders is to treat AI as an engineering problem – define the solution, build it, deploy it, move on. But AI joins an unstable world where competitors, customers, employees, regulators, and the technology itself are all changing simultaneously.
Over 50 years ago, systems theorist Geoffrey Vickers captured the challenge perfectly in the title of his book: Freedom in a rocking boat. Leaders face situations they cannot control because everyone else is also acting to change the situation. His subtitle remains prescient: “If we can decide upon our priorities, we can use our new machines rather than be used by them”.
Vickers was writing about automation and labour unrest, but his systems insight applies with even greater force today. The challenge is not to find the right AI “solution” but to build organisations capable of continuously adapting. That means treating AI not as a project with a start and end date but as part of the solution to building an ongoing capability for sensing and responding.
Holistic thinking
Our research suggests this requires a holistic rethinking of processes, structures, skills, and assumptions where nothing is unquestioned and AI is architected for a future nobody can predict.
This is in no way easy, as Amazon CEO Andy Jassy said in his shareholder letter this year: “You have to be willing to reimagine not only every customer experience, but also how you organise and get work done. Challenging conventions that have existed a long time (and worked) is difficult. People sometimes call these ‘change management’ obstacles. That’s true, but in times of transformational tumult, they’re almost like re-examining your faith.”
The key focus should not be on just automating or replacing human work – but thinking about how you can scale value through new innovations Will Venters
For example, AI agents work at an entirely different cadence to humans – humans work eight hours a day, five days a week – that’s less than a third of the time. So human-in-the-loop oversight of agents will be out of tempo and a drag on possible work – and can even become a liability when AI can surface patterns in real time.
Without careful thought, Rashed Iqbal, chief technology officer at RAK IDO, explains the challenge: “If I’m able to do four hours of work in one hour, if there is no structure for how the three hours that have been saved will be used, then there’s no benefit to the company.”
Instead, we need humans orchestrating AI agents that directly deliver value at their own cadence and scale – AI agents doing new previously impossible work because, as Johannes Maunz, senior vice-president for research and development at Swedish industrial tech company Hexagon explained: “With agents, there is no headcount ceiling”.
Agentic possibilities
The key focus should not be on just automating or replacing human work – but thinking about how you can scale value through new innovations that harness the new scope, speed and possibilities of agentic AI.
And what about the humans? They remain vital, but not for the reasons most leaders assume. The real value of human work has always been being adaptive – understanding customers, making sense of messy processes, and maintaining the social cohesion that makes organisations function. These are not soft cultural factors.
They are the friction-reducing work that lowers transaction costs, reduces risk, and deals with problems through creative workarounds that are a feature, not a bug, of human organisation, and the reason fragile systems keep running and adapting despite their fragility.
This is the adaptive organisation – humans focused on overcoming rigidity, spotting new value, and innovating, with AI executing rules and applying simple judgement at scale. But this division of labour demands some change in skills among the workforce. We need training in business analysis, customer insight, and systems thinking so staff can redesign how work gets done to deliver value rather than simply adding AI into existing processes.
The question is not whether AI will change your organisation. It already is. The question is whether you and your staff are adapting fast enough to direct that change.
Dr Will Venters is an associate professor of digital innovation and information systems at the London School of Economics, where he leads an MSc in management, information systems and digital innovation. He is also an Amazon Scholar within Executives in Residence at Amazon Web Services. He has written this article in a personal capacity.