Key Takeaways
• Traditional hierarchical oversight gives way to relational and network-based authority.
• AI and new structures (obelisk, amoeba, pencil models) reshape communication flows.
• The future manager orchestrates human-AI collaboration and maintains cultural connectivity.
• Collaborative intelligence becomes a strategic organisational capability.
Core Message
While middle layers may shrink numerically, their relational and connective value must expand to preserve coherence between strategy, AI and human execution.
How will the ‘hollowing out’ of organisation structures with a reduction in middle management change the role of senior leadership? Will they need new ways of communicating with front line employees?
Lessons from Traditional Hierarchies
My first interaction with middle managers started in the mid-80s, in Japan. I joined the Overseas Planning Department of Toyota Motor Corporation as a very junior white-collar worker helping to develop new car product information for sharing with Toyota dealerships around the world – brought in perhaps to help make car manuals from Japan more understandable to new car buyers across the world! My first real boss was an amazingly talented individual (I still think he was one of the cleverest people I have ever met) – and he was a Middle Manager. His role title was kakaricho – “Assistant Manager” or “Sub-section Chief” and he reported to the kacho (“Section Manager” or “Section Chief”). In Japanese companies at that time, such rank titles would immediately indicate to anyone inside or outside the company, your position in the hierarchy. As kakaricho, my boss was responsible for a small team or kakari consisting of about 12 people. Our unit co-existed with a myriad of other kakari all crammed into one huge, open-plan, neon-lit office – in all, about 100 or more people seated in long rows of desks with the bosses located perpendicular to each kakari row, in a classic oversight position.
The senior-most person was the bucho or “Department Head”. My lowly position meant I would hardly ever interact with the bucho: my immediate boss, the kakaricho, was the principal conduit of information from the top. Our kacho would occasionally gather us for pep talks but for the most part, I had very little to do with him. Along with my team members, I relied on the kakaricho to shape my work and assign me tasks. In those days we wrote everything – proposals, reports, faxes – painstakingly, by hand. I recall my boss coaching me to prepare Gutai An or “Concrete Proposals”: these were planning documents on which everything substantive that you wanted to say, had to fit into one single sheet of A3 (diagrams, text, tables). This succinct and lean approach predated the much-vaunted Amazon approach to proposal writing (the “Amazon 6-pager memo” – consisting of 6 sheets of A4 and no PowerPoint presentations) – by about 20 years!
In retrospect, I have come to realise just how critical my boss – and other kakarichos – were to the successful running of the business. It was in their part of the organisation and through individual mastery of ningen kankei (human relationships, human connections) that the miraculous feat of translating strategy (from the top) into “real” work (at the bottom), happened. Without their input – the input of these unsung heroes – it is hard to imagine how anything could ever have got done. And that thought has been reinforced by my interactions over the years with hard-working, dedicated middle managers the world over – from Asia Pacific through Europe to North America. If their numbers are set to reduce – from now on, into the years to come – how will strategy get translated into actual work execution? How will senior leaders communicate with front-line workers, the “doers” – in the absence of the established lines of communication of yesteryear – to actually “get things done”?
The Impact of AI on Organisational Structures
With the advance of AI, the writing is on the wall in terms of the way jobs are changing. Some are being modified, with task elements being replaced by AI: other previously unknown, new jobs are being created. Some jobs are disappearing completely. This means that concurrent with the hollowing out of the middle manager class, as AI begins to impact junior-level roles, we can also expect to see an overall reduction in the number of junior people too. Consulting firms are talking about the emergence of “obelisk”- shaped organisational structures as the pyramids of old turn to dust. It may be that in other organisations we see the creation of what I call “stubby-pencil”-shaped structures that maintain a vertical shape – with AI as the core (“the lead”) of the pencil and with slimmed-down human headcount wrapped around that core.
Other structures to be expected (because they have already been evolving for some time) are more like a collection of amoebas: they operate quasi-independently, sometimes working all on their own, sometimes merging temporarily with other work units, sometimes separating out again after a project has ended. Some, like Haier in China, maintain independent but closely connected relationships, buying and selling skills and expertise within an identifiable ecosystem. Whichever polarities in structure proliferate in the years to come, whether obelisk, pencil or amoeba, the way in which senior leaders communicate with their worker populations is going to evolve.
Collaborative Intelligence in Human–AI Organisations
One way to think about this is to consider the concept of Collaborative Intelligence – the practise of understanding how working people connect and interact. AI can help us to “see” channels of communication and even to discern “bottlenecks” where information flows slow down (for a variety of reasons and very often, for quintessentially human reasons – like simply not getting on with people you are (at least on paper) supposed to connect to and work with). Collaborative Intelligence can be broken down into three main constituent parts: knowledge (held by humans and increasingly, by AI); intellectual discourse (between humans and increasingly, between humans and AI) and collaborative technologies (examples would include email; Zoom; Microsoft Teams).
What this means is that while the absolute number of what we have known as middle managers may hollow out and decrease in real terms, their value to organisations as a channel for strategy into execution will remain. The difference is that they will learn to marshal the power of AI in the way they had to learn how to marshal the combined skills of groups of humans in the past. In fact, for the foreseeable future, their role will be to meld and merge human and AI attributes into the combined power of what will be the teams of the future.
They will also continue to enable human connections to be made between senior leaders and their co-workers elsewhere in the organisational structure, with AI providing real-time information about activities across the enterprise such that the need for some human to human conversation about “what’s going on?” will not be so necessary, because what the senior leadership needs to know will be right there in front of them, courtesy of AI.
My hope is that the middle manager of the future will be able to enjoy the kind of support afforded by AI in ways that would have been unimaginable to my kakaricho boss of 40 years ago.
My lasting memories of that kind, humble man are of smiling at him at the end of the day, saying o-saki ni shitsurei shimasu (“Please excuse me for leaving before you”), knowing that he would be staying on in the office for several hours more. He would smile, wave, light another cigarette, and dive back into his work.
