Today, our guest is Ravin Jesuthasan. He’s an author and the managing director at Willis Towers Watson and this is Work Minus Tedium. Hi, Ravin. How are you?

Good, Neil. How are you?

Doing excellent. I wanted to get into this topic of tedium. We’re going to be talking about AI, we’re going to be talking about robots, lots of different things. But first, give people a little bit of background about who you are.

Sure. So, I am the coauthor of three books on work and human capital, our most recent book being “Reinventing Jobs: A 4-Step Approach to Applying Automation to Work” that was published by the Harvard Business Review. I also am a member of the World Economic Forum steering committee for work and employment. And I’m a regular participant by virtue of that relationship at the annual meetings in Davos.

So, that’s a very prestigious role. Tell us about how you came about that. It’s obviously something we rely on a lot for good stats, good statistics coming out about work. So, tell us more about that World Economic Forum.

My employer, Willis Towers Watson, is one of the strategic partners with the forum. And through that relationship, I’ve had the great privilege of working on a number of projects with the World Economic Forum, our most recent being that we’re in the process with this year is HR 4.0, which calls for the reenvisioning of all things related to how talent is engaged with the enterprise.

Fantastic. So, that’s really much in line with what we’re wanting to talk about, how we can reinvent things, realign things. Today, we’re talking about Work Minus Tedium. So, tell us how do you define tedium?

The way I think about tedium is things that are highly repetitive, rules based, which require very little independent thought and more of a routine approach to execution. There are lots of bodies of work, probably in parts of your job and parts of my job, which fit that description. And it’s the things that if you almost go back in history to how we organize work, many would say that in the second industrial revolution, we almost dehumanized work because we brought work that was done independently into assembly lines and processes. And the quest for efficiency is what really drove a lot of the organization of work to almost increase the tedium premium, if you will.

So, originally, you said most of our work was human, and then with the industrial revolution it dehumanized. And now, how would you describe what we’re doing now? We’re trying to automate? We’re trying to pass off all those tedious tasks back to machines?

In many respects, yes. So, one of the reasons my coauthor, John Boudreau, and I wrote this last book, “Reinventing Jobs,” was because so much of the narrative around automation and work is in all candor flawed because it’s denominated in terms of jobs. And the reality, as many of the great thinkers know, is that automation affects tasks. And it affects them in one of three ways. It either substitutes for, one would say, the tedious human tasks, it augments many other human capabilities, but it also creates demand for new types of human work. And the framework that we put together really starts to deconstruct the job and ask the question of, let’s identify all the things that fall into those the three categories and the relevant type of automation, whether it’s to substitute, to augment, or to create new human work.

So, tell us about your own job. What are some tasks that you do that you would like to or are already substituting with automation?

Great question. Actually, I consider myself really blessed in the sense that I have a role that gives me a lot of freedom. I have the opportunity to innovate and create. I think probably the tasks that are the most tedious, if you will, and I’m, again, blessed to have a great team around me, colleagues who take on some of these things, colleagues who move some of these things on to different automation platforms, but it’s the things related to billing of clients, if you will, that we do a project with a client, we have to send them a bill, someone has to document all of the work that was done. And fortunately, we have great technology at our firm that allows us to automate much of that so that the role of a human, or in this case, me, is really to make sense that the client is being billed appropriately for the right work, etc.

I like the term you said the role of the human is to make sense of it. A lot of times I feel like that human element is just to look at what’s produced by the automation or the robot and say, is this reasonable? Does it seem correct? Is that what you see one role that humans can play?

I think increasingly, yes. It’s where we are still teaching the machine as it relates to where it’s substituting the work that we’re doing, it’s almost like having that human be the sense check to be the reinforcer of the learning. And there is a lot of work today which involves that. You see that across the world where, as we are applying more AI, needing humans to make sure that there’s no bias in the decision being made, needing humans to make sure that the decision being made is a logical one, and I think, increasingly, one that is empathetical. It may be the right decision made by the machine, but maybe the wrong decision from a human empathy standpoint.

I like this discussion where we’re going. Let’s talk about augmenting. So, we talked about substituting. Next is how AI and automation will augment humans. So, give us an example of that.

One of the case studies in the book, which I absolutely love, is of an organization in the UK that is a big retail organization. They have big call center operations. And they use AI to categorize emotion in the voice of callers. So, within about 8 to 10 seconds of you calling, the algorithm, it’s a natural language processing algorithm, it can categorize emotion based on a two by two matrix of severity and nature. And so, if you happen to fall into the bright red box in the top right hand corner, your call is sped through the eight other prompts to their best customer care person. And this is someone who’s been hired because of his or her emotional intelligence, their ability to communicate with empathy, their ability to truly connect with customers at a human level, and that, not just capability, but also that desire to have those human connections. But the algorithm doesn’t just stop there. It also tells this person the words to use versus not use based on the nature and severity of emotion, as well as the idea of script. To me, it’s a great example of using AI to augment what is truly human about us.

That’s a really fascinating example. I’ve never heard of that, to be able to almost take the best of what humans can do. Because if you put a human on to listen to all those things, certain people would do better than others at being able to figure out the severity in those types of things. But to be able to teach some kind of AI to be able to do that and then augment that, that’s a really fascinating example. I love that. So, let’s go on to creating. What are some ways that AI creates new opportunities that weren’t there before?

Oh, yeah, absolutely. And that’s the area that, to me, is the most intriguing because we, as a species, and I guess every species, we see what is immediately in front of us, we see what we have today, and we are most acutely focused on where we think what we have today is being taken away, versus our ability to envision what might come next, and see and draw on some of those connections. And we’ve seen this over the last 200 odd years of industrial revolutions where going back to the Calton weavers of the 1700s who rose up, the population rose up and stoned the weavers and the machines because the power loom was transforming the textile industry back then. A couple of examples of where new work is being created as a result of automation and the advances in AI, I’ll give you one example, we wrote an HBR case about this. It’s an organization, very large oil and gas company that was bringing in automation to substitute the highly repetitive, dirty and dangerous work, dirty, dull, and dangerous are the three acronyms we often use when we think about a substitution play. But they were bringing in this automation with the goal of substituting the things that would cause people to lose fingers and toes, the things that would reduce their work life. But in the process, the presence of the automation, while it might have substituted for many of those human tasks, it also created new types of work.

And specifically, the three types of work created by virtue of these robots, one is the work of actually operating the robots. So, you needed a human to actually maneuver the robot and make it work, etc. You needed someone to calibrate that particular piece of equipment because the oil rigs would move, you would need to calibrate that equipment to ensure that it was working appropriately. And then the third body of work was actually maintaining that piece of equipment because machines break down and you need someone who is trained and skilled enough to be able to do both preventative as well as remedial maintenance. And the thing, though, I think the important thing is when these new skills that work is created, the gap between the skills that were substituted and the ones that were created can often be significant. And that’s where I think many organizations have a choice. In this case, the organization made the right choice, in my mind of saying, we’re going to reskill all the people where significant percentage of their tasks have been substituted to be able to actually perform these new tasks that have been created.

So, you said that they made the right choice, which I totally agree with. What are some examples of companies making the wrong choice?

I’ll stick with the same industry. In the mining side of natural resources, there’s so much technology that has gone in. And so, in the past where you might have had a mine employing maybe 300 to 400 people, today increasingly what you see is mines with autonomous self driving trucks, mines with remote controlled blasting equipment, and perhaps remote control in drilling equipment. And so, increasingly, you don’t need 300, 400 people. What you’ve got are four or five supervisors, who are there to step in if the equipment breaks down or if there is an issue. And so, I think that’s one example of where perhaps I think the consequence of that decision could have a huge negative impact on the societies and communities in which those organizations operate.

So, let’s come back. In your book, you say that it’s not right to jump straight to the question, what automation is possible. I think a lot of times when we talk about this topic, people say let me just look across everything I do and see what’s possible to build into automation. But why is that not the best starting point?

That’s a great question. The way automation comes into organizations is in two ways. One is where leaders lead with the technology. So, they go out and they say, I’m really intrigued by robotic process automation. I’m really intrigued by natural language processing or social robotics, as we talk about it in the book. And it’s almost like I’ve got a hammer. Let me go see how many nails I can find. And I think this is going to become a really core discipline of leaders going forward is continuously asking the question of what workflows or jobs can I transform? How can I deconstruct them and identify where these different types of automation have a critical role to play? Where are we substituting human judgment where what we’re trying to do is minimize errors and variants? Where can I augment the skills of a human being so that he or she can make that breakthrough in performance with the right form of AI? And what new human work can I create which really makes the most of that human innovation and empathy? And that typically is the more sustainable approach. It’s starting with the work as opposed to leading with the technology.

I like that a lot. You used the term robotic process automation. I think that’s a new one for most of our listeners. Can you describe what that is?

So, in the book, we looked at the three dominant critical types of automation, robotic process automation, RPA, cognitive automation, or AI, and social robotics, which is the combination of equipment, physical equipment with mobility and sensors and AI. So, it’s equipment that can work alongside humans. RPA is probably the most mature of those three. It’s been around for a long time. There is no “intelligence” within it. It’s really process automation. So, let me give you one example. Let’s say you apply for a mortgage at a bank. There is a compliance analyst at that bank who is going to take your details. He or she is going to cross tab that information against the bank’s information. They’re going to go pull your details from the IRS to validate your income and all of that, they’re going to go do a social media scan to ensure that you’re not going to be using the loan to fund some undesirable activities. They’re going to be pulling data from the credit bureaus like Equifax and TransUnion, etc., to validate your history. All of that today, that body of work, can be largely done by some form of RPA, which essentially integrates data from multiple data sets into the presentation layer. So, what it then does in that case, it frees up the individual, not to go pull data and spend a lot of time on the tedious work, which also, tedious work is also where errors happen because it’s tedious, right? We humans are not infallible. We make mistakes. We get tired. But it’s a great place for RPA to step in because it can do the processing much quicker and it can do the processing with zero errors.

So, when you think about the typical office situation, you said RPA is a technology that’s been around for a while, in terms of process automation. What are some of the new things that people can start thinking about how to apply just to any regular office situation?

I think the more interesting applications are the combination of RPA with AI, what many would call smart automation or intelligent process automation. And I think that’s really where we’re seeing this play out in across a lot of white collar work. We’re seeing it play out in work that is often knowledge based, like the work of an accountant or a lawyer, where they may be gathering lots and lots of data, and then applying a set of rules to do some analysis and to come up with some conclusions. Well, today, some form of smart automation can do all of that because you’ve got the synthesis piece that is being done by RPA and then the analysis piece that is done by some form of AI or machine learning.

So, you’re talking about white collar work. I want to know from you, from this sense of where tasks are going, where automation is going, what are some of the probably blue collar jobs that are likely to stick around for a long time and not be automated anytime soon?

It’s interesting. I think we are going to start to see the blending of what we think of being traditional white collar and being traditional blue collar because I think I’ve seen the phrase gray collar or beige collar. What it is is work that combines both the knowledge and an experience and understanding of the physical realm, so factories, warehouses, etc., with the ability to apply some cognitive capacity to that work. So, we’re seeing a lot more of this blending is in as companies are using social robotics, as I talked about a second ago. So, instead of having product move in linear fashion down an assembly line, what you have are robots that swarm around that product. And so, within that manufacturing or warehouse setting, what you have are people who can work with those robots, they can calibrate them. You have a lot more data analytics capacity because as with the robots, you have the potential for those robots to spit out a lot more data, in terms of the distance they travel, the number of pieces they move. So, you have the ability to continuously harmonize and increase the productivity of those robots because you’ve got a lot more data. And you can identify weak points in the layout of your factory, you can identify weak points in your processes, and you can look at differences between robots. So, that data analysis capacity or skill is something that we’re seeing happen a lot, be required a lot more in manufacturing and other physical environments like that.

What about trade jobs? We had a guest on, Byron Reese, who talked about he felt like plumbers would be one of the least likely jobs to be automated just because every situation was new. It was very difficult to program something to be able to handle those types of things. These skills like electricians, plumbers, welders, and types of things, do you feel like that’s also, at least for the short-term, not something likely to be automated? Or do you think that, too, will also be subsumed?

I think I’ve learned over my 30-plus year working career to, particularly of late, to never say never. But I do think that those things are much, much further out than we think. And in large part, if you go back to our framework, Neil, it’s because, yes, you might think they’re repetitive, but they’re actually incredibly variable because the context and the environment is so unique from one home to another, from one circumstance to another. And it’s the same reason why I think it’s going to be fascinating to see this narrative around truck drivers play out, longhold trucking is probably ideal because the number of variables is much less. But when it comes to shorthold, whether it’s within a city, particularly within metropolitan areas, I think it becomes incredibly complex because there are so many more moving parts, there are so many more nuances. And where machine learning works well is where you have modularity in data, and you can train the data and you can get a predictable outcome each time. So, I think the application of automation in the physical realm is going to be much more nuanced than many of us think.

As someone who works with a lot of humans and with a lot of robots, what are some of the roles, some of the jobs that you will be very sad to see get passed over to robots, that you really want to see stay with humans in terms of a human touch or some kind of connection you make? What is on your list of things you really hope that humans hold onto for a long time?

Oh, boy, that’s a great question. I think each of us can probably identify transactional repetitive activity where we like the human connection. I was watching a thing at McDonald’s where they have automated, I think, the frying of chicken nuggets, something like that. So, you drive up and you say what you want and the machine immediately if you say chicken nuggets, it starts cooking them, etc. So, they’ve almost eliminated humans from that process. I would categorize my response in two ways. One is I can fully appreciate from a micro perspective, from an individual organizational perspective, and maybe even the consumer perspective, why that makes sense because it drives more efficiency, it reduces errors, and it’s that step two of our framework where what we’re trying to solve for is both minimizing errors and minimizing variance with a standardized product. And so, I totally get that. I think what I’m more concerned about is the rapid commoditization of work, which then leads to the rapid automation of that work, and the social implications of that. And if you think of the role of some of those jobs, they’re often on the ladder to people acquiring skills of what it’s like to be in a work environment, it’s for young people to have the opportunities for acquiring knowledge, of learning to work with others, learning to collaborate, learning to work with equipment, learning to take orders from the manager or supervisor. I really worry about the social consequences of some of those decisions because what it does is that it creates what some have called the hollowing out. It doesn’t create the pathways and experiences that people need in order to, over time, become productive in society.

I totally resonate with what you’re saying because, on the one hand, I don’t care who makes my chicken nuggets. I want that person to have a job, I want them to be okay, but if it’s a machine or not, it doesn’t matter to me. But I do appreciate that attitude you get when you learn how to work in a line like that, how you learn about efficiency. Even in my own experience, I really value the times I’ve had working in factory settings and understanding those types of things. So, that’s a really good point I haven’t heard before. So, Ravin, tell us more about how we can stay in touch with you and the work that you’re doing.

I am on Twitter and I’m on LinkedIn. And I do, on average, a speaking engagement once a week on the future of work. And I travel all over the world for some of this stuff. I actually did four such presentations last week, including three in Copenhagen. and I post a lot of my materials on both LinkedIn and Twitter. So, I’d love to have your listeners follow me on both platforms.

Yeah, absolutely. And there’s a lot of stuff to engage with, a lot of things. Do you have any new books coming out?

We’re thinking about one down the road, but nothing for the foreseeable future. But lots of articles. We’ve got about 10 on the Harvard Business Review website that, if your listeners are interested, they’re obviously more than welcome to use those. It’s hbr.org.

Fantastic. Ravin, thanks so much for being on the show. We really appreciate you sharing your insights and we look forward to learning more from you.

Thank you, Neil. It was my pleasure.