When you imagine the upcoming Age of Robots, you probably see the robots replacing humans in jobs that involve manual labor. An assembly line is a good application for robots, for example.  And I assume fast food workers will soon be replaced by robots too.

But I predict that one of the first occupations that will be entirely replaced by robots will be middle management, not skilled labor. I think it will be a long time before a robot can replace a sales person or a graphic designer. But it won't be long before a computer can do project management and resource allocation better than humans. Compared to most skilled jobs, management is relatively easy. Management only becomes difficult when there are so many simple projects happening at the same time that a human can't keep them all straight. The individual tasks of management are fairly simple. Management only becomes hard when you add a lot of simple steps together until you have a complicated whole.

Computers are great for handling that sort of complexity. Put a computer in a robot body and it can walk from cubicle to cubicle handing out assignments, checking on progress, and adjusting schedules and budgets on the fly. A robot could easily juggle the complexity of dozens of projects. It could be talking to you in your cubicle while simultaneously having a phone call with another employee and texting a third without you even knowing as it happens.

I don't think it will be hard to teach computers the basics of project management. Most of the common steps for a project have a predictable order. For example, you know you need to get bids before buying hardware. You know you need to prep the space before installing the equipment. And you know you don't put the equipment into production until after it has been tested. A robot can easily learn all of the steps in a common project.

Best of all, a robot would be good at estimating the time and resources needed to complete projects. Robots that are involved in project management would share their experiences through the cloud. Eventually Big Data will help the robots determine how long the various stages of the project should take, and the resources that are needed, based on similar projects elsewhere. The robots will be free of human bias and optimism, so I would expect them to do a better job of estimating budgets and timelines than humans. A human manager will tell his boss what the boss wants to hear. A robot will be entirely objective, creating estimates based on similar projects from history. The robot won't fear being fired if he tells the boss the project won't be done before the CEO visits. For the robot, facts are facts.

One of the biggest advantages of a robot manager is that it can be a hard-ass jerk as often as that is called for. A robot might need to single out weak performers and let the rest of the team know who the problem is so peer pressure does its thing. A human couldn't get away with being so confrontational, but a robot has no feelings. It simply identifies inefficient parts of a system and highlights them.  No one would bother wasting an hour of the robot's day crying in its office or complaining about fairness.

My prediction that robots will dominate management before they dominate blue collar jobs is based on The Dilbert Principle which observes that the least skilled employees are promoted to management. You need your most skilled people doing interface design, engineering, and the hard stuff. Management is mostly about optimizing resource allocation, and that is something a robot can learn relatively easily, at least compared to most skilled jobs.

You might wonder if a robot can have enough leadership qualities to be a manager. I would point out that most humans in management have zero leadership skills, so the bar isn't set high. You can see leadership in humans when you start getting to the senior management level.  It might take a hundred years for robots to get C-level jobs. But I think robots will dominate middle management in twenty years.

Rank Up Rank Down Votes:  +98
  • Print
  • Share


Sort By:
Dec 31, 2012

Wait - is Scott saying that many middle managers are not already robots?!


I take exception to your comments about the "middle aged female workforce". Some of the most neurotic and undisciplined people I've worked with have been men!


re: "The most important thing I do is manage personalities and the inevitable conflicts that go with them".
And that is why YOUR job is safe! But I venture to suggest you may be the exception, rather than the rule.....
Dec 28, 2012
If project management were dealing with facts and logic then computers could do it. If people would just behave reasonably and do the work assigned you might not need project managers at all.

The two main functions of project managers are dealing with ambiguity and keeping things going despite it, and dealing with people and keeping things going despite them.

Dec 22, 2012
Can you build in arrogance, stupidity and self serving idiocy with a total lack of any common sense whatsoever? I.E. take all the evidence, weigh it carefully and then do completely the opposite whilst !$%*!$%* the workforce as much as possible. Oh, and drive a BMW as badly as it can.
Dec 21, 2012
Even better if you could upload celebrity personalities into the Manage-o-bots, just like you can for Sat-Navs.
If you run a small failing music store you could go down the Jack Black route. Large corporates could download Anthony Hopkins in full Lecter mode - "A cost-accountant once tried to test me.."
+5 Rank Up Rank Down
Dec 20, 2012
@ Arsonist,

I know and have known several middle managers who could be completely replaced with an email auto forward rule.

Problem is, none of them were smart enough to figure out how to set it up.
Dec 20, 2012
To be more efficient in the workplace I'm writing a macro to replicate myself. All attempts to date have failed.......but I'm still working on it.
+2 Rank Up Rank Down
Dec 20, 2012
I wonder what that scenario would do to the arrogance factor.

One of the marketing project managers I work with is clearly on the ropes. She is nice, efficient, well organized and with adequate technical skills. She is competent in her role, except that one side of the development team has decided they don't like her. They think she's an idiot because she doesn't understand the technology very well. I do believe marketing needs to get out of the sushi bar and spend *a lot* more time with both sales and development - but at the same time, this person does at least employ other marketing people who do understand the technology. Her job is to make the trains run on time. She is in charge of executing the marketing plan - making sure the assets are created, categorized, launched and tracked appropriately. She does have to approve the assets themselves, but she relies on input from others to make her decisions.

The development team is sabotaging her. They don't review assets or contribute to the process in anyway. They make a public show of being helpful and supportive in meetings and then do exactly zero when the cameras are turned off.

I work with multiple teams - and, from my perspective, the "smarter" marketing managers don't produce better work than she did before she lost support of the development team. (Now they do.)

You can find ways to sabotage even a robot manager, but I wonder if they would choose to do so. I think their enmity stems more from arrogance and disrespect than it does from any actual job performance issues. A robot that was merely efficient, organized, competent, etc. would probably do just fine.
+12 Rank Up Rank Down
Dec 20, 2012
The entire premise is flawed by assuming that management decisions are made based on logic and facts. Middle managers that present factual based plans to the executive level are almost always drummed out as being not team players, or not utilizing their resources efficiently.

Logic/Fact: Pulling a resource to work on an execute pet project for a week will remove that resource from working on his or her already scheduled project thus extending the delivery date.

Real Life: There is no relief given to the original project schedule and the resources are asked to put in the extra time to get them both done.

The middle manager that gets both done succeeds in real life in the short term. In the long run the resources will quit or transfer and overall costs will increase due to turnover costs.

I see this play out every day!

Dec 20, 2012
1 on Stuis comment defending middle managers? 4 on delius1967s comment defending middle managers? Surprising. I would have thought more Dilbert fans had it in for middle management.
Dec 20, 2012
Marshall Brain's story Manna runs with this ball. It's interesting and a good, quick, easy read, like one of Scott's thought experiments.


Dec 20, 2012
[The robot won't fear being fired if he tells the boss the project won't be done before the CEO visits.]

It may not fear it, but it will still be fired/melted down. That information will be in the cloud of Big Data. Whilst a rohbert may have no sense of self preservation, it will have a sense of efficiency. And it will be much more efficient if it is not pounding on the door of a furnace, from the inside.
Dec 20, 2012

Your sentance should have read like this: "As a manager, I take exception to your generalisations...now please sit back and anjoy as I make a generalisation of my own."

I do not think that robots will replace mid-level managers. Their job isn't hard enough to warrent the hardware. Most of the mid-level managers I know could be replaced with an auto-forward e-mail rule.
+20 Rank Up Rank Down
Dec 20, 2012
In the 1700s nearly everyone worked on the land. Now fewer than 1% of people make all the food for the rest of us.

In the 1800s everyone flocked to manufacturing. It sucked up lots of labour. Now manufacturing is a few people tending the machines. Not just outsourced, manufacturing jobs are falling fast.

The web taught us how many things can be automated. We buy holidays, cars and houses using a web form. We get our money from a hole in the wall. We get better customer service from automated systems at Amazon than we do from bricks and mortar retailers.

This means that companies make money without needing to employ people. Wealth generation is no longer linked to employment. Money used to be distributed in wages, and those wages then used to fund government and public services. That model is broken.

Governments will probably ignore the problem - in which case unemployment will soar and taxes will continue to rise as fewer earners pay for more unemployed, until there are too few earners to pay for the services. The system will break down under its own weight.

We need to think of a new model. What are your thoughts?
Dec 20, 2012
This future is already here - and it is called Business Process Management.

Every business is a collection of processes. In a typical business 70% of these is waste.

By defining these processes and automating them, major savings can be made, a better customer service is provided, compliance with regulations is built in and there is an audit trail of every step.

The system starts with a form such as a travel booking, a mortgage application or a purchase. As this progresses through the organisation, checks are made on the applicant, services are scheduled, product inventory is checked etc. all aiming to provide whatever separates the customer from their money.

In the past this was done by email, phone call or meeting. Now the form tells you when it is ready to be approved/rejected, with all the data you need to make the decision. If you don't handle it in a given time scale it gives it to someone else, or escalates it to your boss.

BPM not only eliminates the mundane tasks - shuffling paper, checking stuff and looking things up, but it makes sure that people have what they need when they need it - no more time spent looking for that communication you received last week.

This is a $10bn industry already used by all major banks, most utility companies and, here in the UK, by the Welfare Department. It will eliminate a whole tier of useless non-jobs - the Wallys of this world.
Dec 20, 2012
Reading this, I almost get the feeling that Scott wrote it as a trick -- that at some point he's going to say, "Wait, sorry, did I say robots? I meant to say Eskimos. But now, see how your reaction is different? That says something about you!"

Being middle management is anything but easy. (If it is easy, why are so many people so bad at it?) I've spent about 15 years now being a first- or second-level manager, and what I've realized is that the work that Scott describes, handing out assignments, project planning -- that part IS easy, but it isn't the most important part of the job. The most important thing I do is manage personalities and the inevitable conflicts that go with them.

The difference between an effective employee and a slacker is usually not their levels of ability, it is their motivation. You usually can't get a job without ability, because ability is what companies tend to focus on in interviews. If you want to know if someone can write a piece of code, ask them to write a piece of code. Motivation, on the other hand, is ephemeral and hard to get at. There's no simple way to get at it in an hour-long talk, so almost nobody tries. Instead they base their opinion of someone's motivation on how excited they seem, or what they say about their last job, or any number of B.S. things that have no correlation to actual motivation.

Back to managers: since motivation is the difference maker, what makes a good manager is not their ability to shuffle papers or balance budgets, but how good they are at motivating people. If your employees believe in you, believe in your ability to lead them to success, they will work so much harder for you that it is like having 50% more people on your staff. Great managers, far from focusing on the day-to-day trivia that robots could, indeed, do better, instead spend their time winning the confidence of their employees and peers. Virtually everyone, even people who are self-motivated to a high degree, knows the feeling of working for a manager that you have no confidence in, and how hard that makes it to really put in your best effort. The opposite is also true, though unfortunately many fewer people have ever had the chance to experience it.

The other key attribute of a great manager is an ability to tell who CAN be motivated. Some people cannot. They are there to pick up a paycheck and simply don't care about anything else. These people need to be eliminated from the org, regardless of their ability.
Dec 20, 2012
In most cases upper management rises from middle managment, how do you replace or promote upper managers if middle mangement simply becomes a computer program?
0 Rank Up Rank Down
Dec 19, 2012
Scott, I hope you've read this. It mentions a fast food restaurant's manager being replaced with an application called "Manna".


It's a quick read (30-60 min) that talks about a robotic future both in dystopian and utopian terms.
Dec 19, 2012
I agree that computers can out-schedule humans. On the other hand in many areas the most important role of the middle manager is to hire outstanding first line people, not just schedule them. Hire the right individuals and good things happen. Computerizing the hiring function will take some time...
+1 Rank Up Rank Down
Dec 19, 2012
As a manager, I take exception to your generalisations.

It's obvious that you have had some bad experiences with dud managers in a past life, but no need to tar us all with the same brush. I'd love to see a robot do my job, trying to keep the middle aged female workforce from constantly panicking, over reacting and otherwise imploding on a daily basis. Anyone without experience in being a nagged husband (and that includes a robot) would be reaching for the Koolaide or self destruct button after a week.
+18 Rank Up Rank Down
Dec 19, 2012
The main purpose of middle management is to give the common employee the illusion of the chance of a promotion not leadership. The second is to take or pass along blame for making decisions. The upper management will never take the blame for anything.
Get the new Dilbert app!
Old Dilbert Blog