I attended a conference in Denmark recently where I met Richard Burton, professor emeritus at Duke University.
It was he who first raised the question in the headline. I thought it was an intriguing one.
Computers are increasingly doing the work that middle managers used to do – planning, decision making, quality control. Many people believe that artificial intelligence and robotics will lead to a reduction in the number of middle managers in most organizations during the next decade.
So if computers are doing the work – why not put them on the organization chart?
As Professor Burton said, if your work is not yet automated by a computer, you are probably working for one! (Like the two assistants in the diagram above.)
I guess Prof. Burton posed the question to provoke us to think about the issue.
And it’s a question that goes to the heart of how we define “organization”.
Some academic researchers define organizations as “information processing systems” – systems for collecting, transferring, and processing information.
With this view of the world, you could put the computer on the chart, because people or computers are just alternative ways of providing processing capacity to the system.
But in my opinion, this definition misses a key point.
It’s obvious that information is being transmitted and processed in organizations. It’s also clear that computers can do some task better than humans can do (or at least faster and more efficiently).
But information processing is not the purpose of organizations per se.
Instead, we need to view organizations as “accountability hierarchies” (as suggested by Elliot Jaques). Organizations have a mandate and/or mission – usually tied to external needs and stakeholder expectations.
A hospital has a legal and moral obligation to provide patient care. A school is supposed to educate students. A firm should serve its customers and other stakeholders.
To get these things done, the mandate or mission is broken down into more specific functions and tasks. People at various levels of the organization are held accountable for performing these tasks in order for the organization to fulfill its mission and mandate.
You can digitalize your organization all you want. But computers are not accountable. Only people can be accountable for performance.
A case in point: In Norway, where I live, several hospitals have had problems with their digital patient journal systems. In some cases, the system has failed to schedule appointments with patients with serious diseases who needed to take diagnostic tests.
The typical response is to blame this on a computer glitch.
But we cannot avoid responsibility by blaming a computer.
People define the functional requirements to the IT systems, people program the systems, and people follow up whether the systems do their job.
So we should not put the computer on the organization chart. The organization chart has always been – and should still be – a map of reporting lines, roles and responsibilities.
As we digitalize, the organization showed on the chart may become smaller, flatter, and more decentralized.
There will be fewer and fewer boxes.
One day there might not be any boxes left.
But by then it is no longer an organization, but a machine.
Like most discussions regarding management and organizations, this one suffers from a lack of definition of key terms. Terms like “work”, “organization”, “organization chart”, “manager” and “accountable” need to be defined before we know whether we agree with each other.
We can only guess what Burton means by any of these terms. Certainly, he is not discussing anything remotely resembling Jaques’s Managerial Accountability Hierarchy. Use Jaques’s definitions for those terms and it’s clear that Burton’s suggestion is absurd. Perhaps Burton has a definition of “manager” that allows a computer to be a manager. I cannot imagine what that would be like.
Herb, I agree regarding the importance of definitions, and in my book (where I have more space than in a blog post) I do include definitions and even a glossary at the end (I have also discussed the importance of clear definitions before, e.g., see this blog post). I also agree that Prof. Burton is not discussing Jaques’ concepts, but there is a simple explanation for that: Jaques is practically unknown in the academic world. The dominating view in academia is derived from Simon & March (first published in 1958) and more recent work by Jay Galbraith and others on “the information processing view”.