Jon was recently a guest on WorkMinus Podcast, with Neil Miller. Here are the show notes:
What are we talking about?
The right way to look at automation
Why is understanding automation important for the future of work?
Leaders need to understand the current pace of automation and how we should prepare for changes.
What did Jon Gilman teach us about understanding automation?
- In many companies that implemented enterprise software, productivity actually decreased. Software on its own increases headcount and slows things down. There is too much data and it often paralyzes enterprises.
- Automation can act as an exoskeleton that allows humans to execute business processes faster.
- Automation should begin as soon as you start to notice a repetitive process. Jon says, “Anything that a machine can be doing right now, it should.” If your operational expenses are growing at the same rate as your revenue, you aren’t investing enough in automation
- Automation is moving much slower than the general public believes, but the rate of change will increase soon.
- Automation shouldn’t eliminate jobs, but eliminate the tedium from the job.
- AI was brought about by capitalism, but might lead to socialism
- Humans should double down on strategy, empathy, and interpersonal skills. “The lack of empathy will become a liability soon.” Jon also thinks we should be investing more in educating our kids on the arts and not just on STEM topics
- However, AI, might (and should) develop empathy on its own
Neil Miller: Today, our guest is Jon Gilman. He’s the CEO of Clear Software. And this episode is Work Minus the Fear of Automation. Hi, Jon. How are you doing today?
Jon Gilman: Good. Thanks for having me.
Very excited to have you on the show. You are on the front lines of a lot of the automation stuff people have been talking about. I want you to give a little bit of background of who you are, the company you run, and the type of work you do.
Sure. So, the company I run is Clear Software, as you mentioned, and we’re focused on automating business processes at large organizations. We do that through a software platform that we’ve built that allows people to get their jobs done much, much faster. And where this all originated from was my work as a consultant. So, I worked for Accenture and Deloitte for a number of years implementing enterprise software at Fortune 500 companies. And the number one thing that they found was that their productivity actually decreased after we implemented enterprise software. And they could very easily measure that by headcount increases across just about every department. So, we realized something was fundamentally wrong with business software. And what we needed to do was build an automation platform on top of it so that people wouldn’t have to jump through 27 screens to enter a customer order, or jump through 17 screens and 3 different systems to answer common questions when a customer is on the phone. So, really obvious things, but no one had really done it. So, we built this intelligent automation platform to essentially say, we don’t really care what systems you’re using. We’re going to allow you to define how you want your business processes to flow as efficiently as possible without having to worry about where is this data going to go. That’s what we take care of with our platform.
So, just to make sure everyone caught that, you said that just by using enterprise software, it reduces productivity, right?
Yeah. And we saw this across the board, every time we implemented big ERP systems. The CFO or COO or whoever was sponsoring the initiative would come back and say, “Look, man. I just spent $300 million over five years implementing this ERP system. And my call center is now three times the size it used to be. My AP department is five times the size it used to be. And my IT department is 10 times the size it used to be.” And that wasn’t just a gut feel. It’s literally look at my PNL. It’s increased significantly in all of these departments. So, when you see that hard evidence, you start to think, well, maybe we didn’t accomplish what we set out to accomplish here.
Yeah, because the promise of software is supposed to be, hey, we’ll do it better, we’ll do it faster, we’ll do it cheaper, easier, all these types of things. But that’s amazing that it really just turns it on its head that just adding software on its own actually makes things worse. Right?
Absolutely. On the consumer side, it’s had that effect. Definitely. I mean, when you look at Amazon, some of the things that you can do on online marketplaces, people are able to do their personal activities much, much faster than they used to be able to do that. But on the business side, the exact opposite has happened. We started to see fragmentation over the last 15 years where people will buy CRM software for their sales team. They’ll buy ERP software for their finance team. And all of a sudden, you end up with a spaghetti monster of IT systems all of which don’t have a great user experience. So, the business processes within them are slow, but they also don’t communicate with each other. So, you have this compounding effect of a lack of productivity, and a lack of cross system interoperability and it paralyzes enterprises.
So, when somebody is thinking about they want to add software to organize things or to digitize things, what are the reasons why productivity goes down so much? What are people doing? Why are we needing to hire so many more people to take care of the software? What’s the tedium that’s involved in that?
Typically, it’s a poor user experience, combined with the fact that a job function may span many different modules or screens within a business software. So, as a good example, if I need to create a customer order in SAP, I might have to jump through 27 screens to do that. And on each individual screen, I may be keying in one or two pieces of data. And then sometimes I may have to copy and paste that data to another screen. And obviously, the problem there is that going through 27 screens to do something that’s fairly straightforward, a customer bought something. What did they buy? How much of it did they buy? Where do they want it shipped to if it’s something that actually gets shipped? And that can span many screens and sometimes many systems. Some customers may say, well, the first two steps of this happen in Salesforce and the last eight steps happen in SAP. So, you’ve got to remember to start in Salesforce, then copy and paste your data over to SAP. And people end up having to take a two week training class to figure out how to enter a customer order. And that should be a very simple thing to do.
So then, how do you define automation? That’s the core thing you guys do. What’s a layman’s term to understand what does automation actually do in all this?
It’s acting really as an exoskeleton for the human to allow them to accelerate their business processes. So, there have been a lot of terms right now that are pretty hot in the automation space. So, you’ll hear robotic process automation bandied around quite a bit and really all that’s doing is taking a lot of the old testing automation tools that we knew from the early 2000s and creating macros that run autonomously to complete tasks that a human shouldn’t be doing at all, have tests that require no human input. So, think of something like, I need to reconcile a bank statement. Reconciling a bank statement at an enterprise when you have maybe several hundred thousand transactions a day can be pretty tedious for a human but your rules don’t really change too much. There’s not a whole lot of different outcomes associated with that process. So, that’s something that can be completely automated with an RPA tool or even one of our tools which we would call intelligent process automation. But where we really thrive is where humans are involved in the process, where we’re saying, rather than you jumping through 37 screens and tediously copying and pasting data, we’re going to have you enter that data once and then we’re going to make sure that gets fed into all the various modules within a single system or multiple systems if there are more than one system involved. So that the human operator gets their job done much faster without having to go through all that tedium. So, I think of it, my physical analogy is you see a lot of folks in Costco, Costco employees and warehouses where they have those exoskeletons that allow them to lift up boxes that weigh 500 pounds, when intelligent process automation is doing is allowing humans to basically have a virtual exoskeleton that allows them to carry out these business processes much faster, almost like they have a digital virtual assistant helping them through the process.
So, if somebody is at the beginning of establishing their culture, they’re building out, they’re scaling up their business, automation probably has come across their minds, but they probably have a few reactions. One is that maybe, automation, I am against that. I’m against robots taking over jobs. So, I’m not going to pursue that. I’m only going to use humans for things. You may have other people that would say automation is something only that happens when you get to a certain scale and you have an enterprise then it’s useful, otherwise it’s not. What should be the proper approach for somebody who’s in that scaling period who’s just building a business? How should they think about automation?
Well, one, they shouldn’t be afraid of it. So, we have that issue quite a bit with our customers where we’ll initially come in and some of the folks they’re working with are thinking, “Man, this is an automation company, they’re coming in to eliminate my job.” And that’s just not true. What we’re there to do is eliminate the tedium within their job. Most job functions add value to an organization, but they also involve some day to day activities and tasks that might be pretty menial, data entry that’s duplicated over and over again, things that don’t really require thought, you’re just doing it because you were told that’s part of your job description. You just have to do it. Those are the elements that we’re trying to take out of people’s day to day lives. And unfortunately, if that’s all your job is, if you’re just doing data entry all day long, then that’s something that in the future won’t exist as a job function. But what we’re trying to do is help people who have high value jobs, customer service rep is actually a very high value occupation, because you are the face of the business when someone gets on the phone. And if we can eliminate all those annoying, “can I place you on hold while I look up your invoicing history?” all those things where the customer has to stay on the phone longer than they want to, that improves the customer experience, but also makes the life of the customer service rep easier. So, the customer service rep isn’t losing their job, they’re getting a happier customer and they’re able to have a more positive experience with them throughout the day.
So, what do you feel like is a good balance to look at? When you’re in that scaling period, you’re thinking about building a team, you need to think about both building your technology team and your human team that’s there. As you look out into the future, what do you think that smart companies right now are doing to prepare both teams and how they interface with each other?
I think sometimes it’s a little trial and error. Because obviously automation doesn’t make sense at small scale. There’s no point to automate a start up of five people. But obviously, as you start to scale, and you start to see repetitive processes, that’s an opportunity for you to optimize your business. So, I think you have to plan accordingly based on your data points. And a great way to do that is with some very easy process mining tools to show that when I’m $100 million a year company and I’m processing maybe 100,000 customer orders, or maybe 50,000 customer orders, automation is less of an incentive. But if I’m 3M or if I’m Johnson & Johnson, I’m processing 50 million customer orders a year or 100 million purchase orders a year from my vendors, there’s a huge opportunity there to increase your operational efficiency and start to plan for that. So, it’s definitely got to be a data driven mindset. But there’s also a human element to it as well. If you come in and say, like I alluded to earlier, you’re going to bring in an automation technology to help improve your business, that’s going to make people feel pretty uneasy. So, you’ve got to be mindful of the human element as well and have a little bit of empathy and understand that this isn’t all just dollars and cents. It’s also organizational change management, making sure that people still have a fulfilling work experience.
Jon, you’re going to be one of the best people to answer this question. I’m always on the lookout for what are the core human activities that are there that’s really, at least let’s say for the next 25 to 30 years, these are only going to be able to be handled by humans. When you step into a company and you look at what can be automated, you’re probably much more aware of what the possibilities are. What do you feel like are those tasks that are just going to have to be done by humans foreseeable future?
Well, anything that involves face to face interaction or human interaction, for now, definitely still needs to be in the hands of human beings. But anything that a machine can do right now, a machine should be doing it or an automation software. So, when we think about tedious data entry or generating reports out of multiple systems, these are things that people spend hours and hours and hours a day doing and it’s not very fulfilling, you don’t feel good doing it. You don’t enjoy doing it. That’s what’s going to start to go away as a job function so that people will really be focused on things like face to face sales, face to face strategic sessions with customers and vendors. And then the robotic element of it is going to disappear.
So, what are the things that you feel like people should, even as you’re talking to your own employees, you’re thinking about your children growing up, if you have children, what are the skills you feel like are necessary to invest in? You mentioned sales, you mentioned these face to face interactions. As we think about let’s get better at being human, that’s one of the things we talk about on the show, of saying the future is human and machines working together. But we need to get better at our side of understanding what are we really good at. If you could choose anything to help your employees in the next generation to invest in those skills, what would you say they should invest in?
I think it’s definitely empathy and interpersonal skills. That’s going to be really important. Because as we start to take the hard data driven processes out of our day to day job functions, it’s going to be primarily more interpersonal. So, I want my kids, who are all very young right now, to develop empathy, but they still also need to understand data and understand technology. But we’re definitely going to be driven more towards an empathetic world as we eliminate the repetitive drudgery of day to day life. So, classic example is, if you’re a hardcore introvert and you’re technically inclined, up until today, and even today, you could basically stick that person in a dark room and have them generate reports using SQL commands, and that person would be great at it, they would never have to deal with people. But when that job function goes away, that person’s not going to really have another option. So, I think the lack of empathy is going to become a liability with folks as we automate more and more.
Tell us about the pace of automation right now as you look at the whole industry. In some ways, it’s easy to look at and say, well, maybe things aren’t changing that fast. It’s not going that quickly. But other people may be feeling it a lot in certain areas. So, from your vantage point, do you feel like things are going fast or slow? And what does that mean relatively to how they were 10 years ago, and how they’ll be 10 years from now?
It’s definitely going more slowly than I think the general public thinks. Instead of being a starry eyed futurist, I like to be a realist because I’m actually working with my customers every day. And I understand that they have clunky business processes that need to be optimized first. You can’t put automation on top of a terrible business process. So, typically, we need to re engineer or one of our partners needs to re engineer these business processes before we can put automation on top of it. And that’s slow and painful at a large organization because they may have 13 different ways of doing something and we want to get them to one standardized way of doing something. So, it moves a little slower than people think. But the positive side of it is that because of the whole RPA boom right now, organizations have started to take a look at automation and realize that there are tangible benefits to this, the dollars and cents add up. So, they’re starting to look at this and really take it seriously, which is great. But they also underestimate the huge amount of process reengineering that needs to be done. And that’s no small feat. If you’ve got a bad business process, it might take you 6 to 12 months to figure out how to get it to function smoothly and up in an optimized manner.
So, we’re talking here in 2020. And you would say, we’re still in the very early days of becoming an automated society, right?
Absolutely. I mean, anytime anybody predicts where we’re going to be in 10 years, they’re always catastrophically wrong. I think five years ago, I was watching CNBC. Some nutcase on there was talking about how nobody was going to be driving in the year 2020. Well, now it’s 2020. And I’m still driving and everybody’s still driving. So, I think people tend to overestimate the acceleration of technology. So, it’s going to be that way with automation as well. So, I would say by 2035 or 2040, we’re going to see most of the data entry jobs eliminated, most of reconciliation related jobs eliminated, but there’s still going to be quite a few job functions out there. And maybe when we look out at a grander timescale, 50 to 100 years from now, we’re going to be talking about basically having no jobs at all. Over time AI is going to develop empathy. So, that’s actually Alan Trefler, who’s the founder of Pegasystems, he had a great article on Forbes a couple weeks ago about how it’s necessary for AI to be built with empathy, and my counterpoint to him was, well, we’re human beings. So, for 10s of thousands of years, we’ve been exposed to various stimuli. We’ve developed biases, but we’ve also developed empathy. So, it stands to reason that a computing engine who follows the same amount of stimuli as we do in their environment, but as a much faster evolution, can develop empathy much faster. So, I think we’ll start to see artificial intelligence develop empathy, and then we can start to apply it to specific scenarios like maybe handling a customer call. So, as we start to accelerate that growth, we’re going to see just about every job function disappear, which means labor disappears. So, how do you distinguish yourself from your peers when there’s no labor? Right now you can say I’m just going to work harder than my peer and I’ll get promoted and the rest is history. But the problem is when there is no labor, there’s no way for you to distinguish yourself from your peer. And I think eventually what’s funny about automation is it was brought on by capitalism but it may end up leading us to, essentially, a basic social income state where everyone receives a stipend and they live their lives, and they have a lot of free time. And maybe they work an hour or two a week. But it’s interesting that capitalism may eventually lead to socialism.
These are fun discussions to have. Another question I like to ask people is what would you do personally, if you were getting a basic stipend for all your needs, everything was met, would you still be making software? What would you be doing?
Probably not. I think I would be writing and making music.
Yeah. Which I think if we look to this future society where these things are possible. And we’re at this cusp where we can actually see it as a possibility like that. That’s a great chance to think about these things. And like you said, we’re not there yet. We’re not going to be there next year or in five years. So, it’s not something we need to necessarily assume is going to happen quickly. But I think we are at a place now where we can start to plan for that. We can start to put things in order to get there. So, I guess what would you say? What’s one building block that we can do now to build a future where people can enjoy life to that extent where they can do whatever they want?
Well, I think schools need to stop taking away art programs and things that force students to think and use different parts of their brain. I love STEM and I think it’s one of the most important fields on the planet, but we need to understand history, we need to understand art, because all of these things shape the way we interact with people. And when I see schools starting to drop their art programs, I’m thinking this is really terrible because we’re going to need that in the future when STEM is less of a priority. People are spending most of their time exploring their interests, they might not have any interests, if they haven’t been exposed to anything. So, I think that’s really important. And we’ve seen it happen in the 20th century, because the 20th century probably did more for productivity than all of humanity up to that point. People were working 80 hours a week in 1900. And then it was down to under 40 by the year 2000. And that gave way to a lot of free time for people to explore interests. And because of that, in the 20th century, we saw this explosion in every art form. We even saw new art forms get created, saw film, rock and roll music, all these different great art forms, and there’s only going to be more and more time to do it. So, it’s pretty exciting from that perspective. But it’s all because we’re spending less time working.
We recently had Andrew Barnes on the show who was talking about a four-day workweek. He’s saying we’ve gone from 80 hours to 40 hours. It’s about time to cut it down a little bit more. Are you on board with that? You want to go to 32 or 30 or less?
Yeah, I think 32 is achievable right now. And then, over the coming decades, we’re going to see that scaled back quite a bit. But, ideally, we should preach that utopian state where you’re working one or two hours a week. That might be the year 2100. But some day maybe my kids will live to see it, but I won’t.
Well, cool. I love these discussions. It’s fun to dream out in the future. Let me ask just one thing to bring us back to what people can do now when they’re thinking about it. As we’re talking to leaders of these scaling companies, when they’re thinking about automation, what’s one sign that they should be embracing or one area of their business where they should really think about? If they’re starting to adopt CRM systems, ERP systems on a big level, what’s a way they can start in the right way so they don’t get to mammoth size and have to go back and re engineer everything? What advice would you give to somebody at that stage?
At that stage, you want to pay attention to your data. So, as your revenue is growing, if your opex in certain departments is growing at the same pace, that’s a problem. That’s where you need to start looking at some forms of automation. So, if you grow from 50 million to 100 million in revenue, yet your customer service department went from 5 million a year in opex to 10 million a year in opex, there’s a problem there. So, when you start to see things moving linearly with your growth, there’s a huge opportunity there for process improvement and automation.
Yeah, I like that. That’s a really good opportunity. Cool. Jon, thanks so much for being on the show. We appreciate you coming on, sharing your insights. You’re obviously on the forefront of these things. So, we look forward to staying in touch with you and hearing from you again soon.
Awesome. Thanks, Neil.