I assumed that the benefits of automating jobs – the promise of doing more work with fewer people – would be obvious, and that business leaders would eagerly embrace the latest and greatest automation technology as soon as it was commercially available.
In reality, though, there is an entire industry of people whose jobs it is to talk business executives into embracing automation. There are salespeople who throw big conferences and wine and dine clients, nudging them in the direction of multi-million-dollar RPA packages
. There are “digital transformation” experts who travel all over the world, telling CEOs and CTOs that they can save money by replacing 30% of their billing department with software bots. Then, once the executives are convinced, there are deployment specialists whose job it is to actually carry out the automation – parachuting into offices, analyzing processes and workflows, matching up tasks with bots capable of doing those tasks.
Consultants wield a surprising amount of power in this process. Often, they’re the people who tell CEOs which jobs can and can’t be automated, which workers should be reassigned or laid off, and how quickly it all should go. Executives rely on their advice to make decisions that govern the fates of millions of workers, yet barely anyone knows they exist. And as the economy gets more and more automated, their influence will only grow.
This week, I called Geoffrey Hodgson, a Chicago-based automation consultant with Saggezza, a global IT consulting firm. Here is an edited, abridged version of our conversation.
When you get a call from a company saying, “We want to automate part of our business,” why are they generally looking to do that?
They fall into two kind of camps. So obviously there are businesses who are solely interested in automating business processes for the purpose of removing headcount. If you’re looking at pure cost takeout, it’s “We have heavily manual processes that we know can be automated, and we don’t want to spend $50,000 for a person to do it, we want to spend $20,000 for a bot to do it instead.”
The other side is those who have bought into the grander idea of digital transformation and automation, which is enabling skilled knowledge workers to perform better, to be more efficient, to make better decisions, to remove the stuff that weighed them down in terms of being able to actually do what you want them to do.
Of those two camps, which is more common?
It’s probably 70% the FTE [full-time employees] push. But I would say that we try to help clients think about the business case of automation in a more holistic way. Because I have seen more failure with a “we just want to replace humans with bots” approach than with the the more holistic approach. When you’re when you’re looking at things from from a pure FTE/cost takeout point of view, the possibility that things don’t work out as you wanted them to is higher. And in my experience, automation programs have ended prematurely more often with with companies who are purely looking to remove human beings.
When automation projects don’t work, is it usually because the bots aren’t as good as the bosses thought, or is it because the bosses didn’t realize how many different types of tasks their employees actually did? In other words: are they overestimating the bots, or underestimating the humans?
That’s a really interesting question. There’s definitely an aspect of it where people overestimate the bots, but I wouldn’t say that it’s necessarily that they’re overestimating automation technology.
The most frequent thing that I see where automation projects fail is that companies approach the technology in too narrow of a scope. They’ve read the the articles about RPA [robotic process automation] being a game changer, and to an extent, it is.
But RPA is a gateway drug of a technology. It gets you one or two FTEs here and there, and it is effective for very rules-driven approaches. But you need the information to be structured, you need the data to be structured, you need very clear rule-based processes to do it. And in reality, those processes are few and far between.
What is the process of doing this? You get a call from a CTO of a Fortune 500 company who says, “I want to automate.” What happens from there?
The first step is an introductory call. We try and understand what the client’s objectives are, what they’re really trying to get at. Maybe we do a little bit of an education piece if they are not super familiar with the technology.
The next step, if we decided to go ahead, would be to do “process discovery.” There’s actually two ways of doing it, and the industry is undergoing a little bit of a shift about how this is done now. The old way is to sit down with business leaders and understand how the work is done today, to try to map out where those opportunities live for automation.
The newer way is to take a data-driven approach to finding opportunities for removing bottlenecks. You can apply process-mining technology to the transactional databases, to people’s desktops, to see how work is done and get millions of data points, and leverage that to build a picture of how work is done at a macro level across the enterprise.
Then, once we know what and why we want to achieve, we need to understand how. So we go through an exercise around vendor selection, and try to whittle down what would be most appropriate for the client. And then from there, it’s, do we have a business case [for automating], yes or no?
One thing I’ve been sort of puzzled by is why there’s so much jargon and euphemism that seems designed to hide what’s actually going on inside these automation projects. Stuff like “digital transformation” and “liberating workers.” Wouldn’t it be much better for workers who are at risk of being displaced by this stuff if people were actually being honest with them?
If you look at the way automation companies are marketing themselves now, a lot of it is focused on the concept of “a robot for every person.” And you could say that that is driven by wanting to perform some cynical marketing campaign to cover up what’s really going on.
But I actually really think it’s a reflection of where the automation industry actually sees its own pitfalls. You made a comment in one of your articles about the risk of “shitty automation.” Well, I think they see that, too. And so I think they’re trying to push use cases beyond just getting people out of seats, into a space where we’re really trying to leverage automation as an efficiency tool in the same way that you would leverage Microsoft Teams or Slack, or even Excel.
Obviously, there are when you start thinking automation, there are going to be jobs that are replaced by technology. It is somewhat inevitable. And I think the companies who have done best with it have been a little more open with people.
How has the pandemic changed how companies are thinking about automation?
There’s definitely an uptick in interest around automation. And I’m not yet settled on whether that is because there’s something about working remotely that is inherently advertising for an automation program, or whether it’s that people having been thinking about automation for a long time, and now believing that there is a reason to go for it because there’s this massive disruption anyway.
I’ve talked to a number of CEOs and CTOs who are deciding whether to automate their businesses. And usually, they frame it in the same terms, which is like, “Look, we don’t have a choice. If we don’t automate, our competitors will. And they will become more efficient, and we will die.” Do you find that reasoning compelling?
I mean, it’s true in the medium term. But in the long term, there has to be some level of equilibrium, right?
Let’s just operate with the assumption that the main objective is to automate jobs. There has to be an equilibrium whereby the jobs that are automated mean that there is just enough efficiency for the businesses to survive and thrive, but not so much that we have 50 million people unemployed, and therefore no money for the businesses to make.
So I think in the medium term, that argument that you were laying out for the CEOs makes sense, but in the long term, businesses have to really think about the application of ethical AI.
I mean, they never will. Nobody ever does.
But in the aggregate, are they doing the right thing for the economy? And again, that goes back to the point that automating for purely FTE takeout doesn’t just indicate a higher probability of failure. It’s really less helpful for the economy as a whole.
Is there a social stigma attached to being an automation consultant? When you tell people what you do, do they accuse you of killing jobs?
Yeah. There are definitely some really, really talented people that I’ve worked with who have quit because of that very conflict in their psyche. My former mentor at my former employer quit for that very reason. She was a fantastic person, but she just couldn’t align her personal values with the fact that even if she didn’t think she was necessarily automating away people’s jobs, her friends and colleagues and clients all thought that’s what she did for a living, and she just couldn’t live with it. So she quit and did something else.
I do get that from my friends and relatives. “This is Geoff, he automates away people’s jobs for a living.” But I feel like to survive as a consultant in the automation industry, you really have to be an advocate and a believer in the fact that it has the ability to do good.