This chapter of the Art of Automation is a reduced transcript of a conversation between Jerry Cuomo and Ed Lynch, Vice President of IBM Digital Business Automation. In this discussion, Jerry and Ed take a step back and discuss the concept of Enterprise Automation as a whole; specifically, where we are today and how we got here.
Ed explains that automation is everywhere — not limited to one specific industry. He then elaborates on the future of Enterprise Automation, and how it inevitably will be centered around artificial intelligence (AI) and augmenting human work. They close by determining that automation all comes down to the human beings and that the best place to begin your automation journey is with the data.
Jerry and Ed
CUOMO:Welcome to the Art of Automation, a series that explores the application of automation in the enterprise. So, what exactly is automation, you ask, and why dedicate a podcast to the topic? Well, in this episode, my guest and I are going to take a step back and reflect on this very question and do our best to shine some North Carolina sunshine on automation, looking at it in its recent past, where it is today and where it’s heading tomorrow.
So, if your business is dependent on being digital through and through — and I have yet to meet many businesses that aren’t today — then listen in to the thoughts of one of our wisest wisemen, Ed Lynch, who has been the visionary behind IBM’s successful business automation strategy and product line since the early 2000s and is going strong today.
Few people that I’ve had the pleasure of working with balance the business and technology of automation the way Ed does. He has taught me so much about this space. So, in this episode, Ed is going to share some of that wisdom with you in a hope that you can use that insight to better prepare for the future ahead.
And while Ed will reflect on the past and present of automation, I promise his views on where automation is going is worth the listen — including how human beings and computers will continue to evolve to do what they’re best at doing. And with that, I’d like to welcome Ed to the Art of Automation. Welcome, Ed.
LYNCH: Hey, Jerry, good talking to you again. Thanks for having me.
CUOMO:So good to have this conversation with you — let’s get right into it. Ed, before we talk about where enterprises are going with automation, can you share a little bit of where they are today and why are people excited about it; and, perhaps a little bit about why now.
LYNCH: Sure. I’ve been in this space for a really long time. Let me give you like the brief glimpse of history and then tell you where exactly we are today.
I started this maybe in 2001, approximately. And back then, we were talking about system-to-system integration. We called it business integration, and that was really the synchronization of one system of record with another system of record. It was a very IT thing — it was managed with no humans in the middle, but it was a way of doing what we called contact sync way back when. And that was all about making sure that one system will work with another system.
And then we added human beings and then we added decisions and then we added planning — we added monitoring and that’s kind of where we are today. Where we are today is we’ve got a whole bunch of different mechanisms for automating work — whether the work is human work, whether the work is system-to-system work, whether the work is dashboarding work — that’s kind of where we are today.
We’ve got this trend in the marketplace which is really prevalent — and it’s become extremely prevalent with COVID — which is the hyperautomation trend. Gartner calls this “hyperautomation” — number one trend for two years in a row. This hyperautomation trend is why there are billions of dollars coming into the marketplace. Like, just yesterday, there was an announcement that Celonis had just secured a billion dollars’ worth of funding.
LYNCH: Like just completely, completely crazy. But why? Why now? Well, it’s because everybody’s focused right now on productivity. The productivity of the workers —not productivity of front-office workers with, you know, Office and Microsoft and that kind of thing — but productivity of the back-office workers, productivity of IT workers, operators, SREs, the sys progs who do work in automating IT systems. All of those different workers are under the microscope right now.
And businesses are looking for ways to get more out of the human beings and apply the human beings that are in both business and IT in really creative ways which take advantage of the skills that they have while letting the machines do the work that the machines are good at. Things that are very repetitive, right?
So, this whole space right now is going completely crazy. There are more unicorns in this space out in the market than you can believe, and it’s all driven because of productivity. As businesses look at where they’re spending human capital, they’re trying to make the humans more productive, they’re trying to make — well, make machines take over human work in some cases, applying the right labor type on to the right piece of work. So, like there’s a really simple example here.
CUOMO:Yes, sure, go ahead.
LYNCH: Procure to pay. Procure to pay is a process that most businesses have; all businesses have because they all acquire goods. Amazon procure to pay — most of us use Amazon every day. The going and looking for something in the catalogue, that’s the procure part of procure to pay. That click, buy now, that’s the automation of Amazon’s procure to pay.
There are no human beings in that loop at all. They are completely straight through processing, and that’s what everybody wants to do. Everybody wants to get their procure to pay — which, right now, employs dozens of people in every business, processing invoices, processing ERP systems, cutting checks — everybody wants to get all of those things to straight through processing.
Why? Because it doesn’t add value to the business. The procurement and payment of invoices doesn’t add any value to, I don’t know, a retailer. The retailer makes their money selling goods and they don’t want to be in the procure-to-pay business. So, really there’s a consequential, you know, thing that happens when you automate people, human tasks and you let the machines do it. So, that’s where we are as of today.
Automation as the equal opportunity employer
CUOMO:Yes. And Ed, it seems like where we are — enterprise automation — it seems to be quite the equal opportunity employer across all industries, across all job types. Is that true? Does it apply better to one than the other? Or, what’s your view on that?
LYNCH: My experience, Jerry, is that it’s everywhere. It’s everywhere because most of the things that are being automated are the horizontals, they’re not the verticals. By vertical, what I mean is the sector-specific things. There’s certainly a lot of work in sector-specific things like insurance claims and onboarding government documents, onboarding into different I don’t know, different government programs, as an example. It’s everywhere.
And the reason is because most of the things that are being automated are horizontals. Things like HR work, things like finance and accounting, things like onboarding mechanisms. Those things are being automated everywhere. Verticals are also being automated, but they’ve been automated for a long time.
The things that are being automated now — why, robotic process automation (RPA) — is going completely crazy is because we’re automating things that are horizontals — like horizontal workflows that are the same no matter which sector you’re in.
CUOMO: So, Ed, thank you so much. Makes so much sense. At the top of show I promised the audience that there are a few people that I know who are more qualified to talk about where automation is going. So, let’s talk about that now; let’s talk about the future of automation. Where is it going and perhaps touch a bit on the role of AI. What role does AI play in that future?
LYNCH: I think those two thoughts are intimately tied together. I think the future of automation — we have automated, we found mechanisms to automate lots of different types of work. But the thing that we’re getting better and better and better at is augmenting human beings, making them more productive. And that’s where AI and machine learning and deep learning and GANs and all the rest of these different techniques are coming into play.
They are mechanisms to transform data like raw data into patterns and then apply analytics to the patterns to give you predictions — and generate probability distributions on what’s going to happen and allow individuals to make decisions based on probabilities rather than just based on one zero facts. This kind of Bayesian decision-making is very, very different, but it’s also extremely helpful in augmenting human beings.
When a human being used to try and evaluate churn, as an example — the probability that some particular customer is going to drop your product — it used to be that you had a lot of instinct, you had a lot of observation, you had a lot of investigation. What’s the past history?
Now we have churn models. A churn model is nothing more than just absorbing all the data, applying some analytical techniques and generating a probability distribution which says the likelihood of, you know, Fred or Jerry leaving your bank is .97
And what that does is that automatically, because of the nature of it it, allows you to automatically start doing things, the preemptive stop churn. It also applies in risk and in lending and onboarding and sentiment analysis and on and on and on. Like all of the various different things that have dozens and dozens like thousands, millions of data points, the machine can absorb them because the machines are capable of absorbing them. So, the future of automation — I think the future of automation is AI.
CUOMO:Wonderful. I couldn’t agree with you more there, Ed. Can you paint a picture of the future state of enterprise automation — maybe kind of continue on what you were doing? And how is it going to or how is it already impacting business? And can you share some key metrics to measure success by?
LYNCH: Sure. The future of automation — when you dig underneath the covers — the classification that I put on things is, I say that some work is done by human labor, some work is done by digital labor. And if you think about it that way, then combining human labor and digital labor into a hybrid workforce — all of sudden a business manager can make a very active decision about where to put the work. Do I put this type of work, this task, do I give it to machine, do I give it to a human?
Now what that comes down to, that decision point comes down to, what the heck are people really good at? Well, they’re good at things like judgment, they’re good at things like strategy, they’re good at thinking out of the box. Machines are terrible at that stuff. They’re terrible at judgment, they’re terrible at reading emotions, they’re terrible at interacting with human beings.
So, if you can make a decision as a business manager, you know, this task is much better done by a human being than done by a machine — fantastic. Then let’s figure out a way to augment the human being. If this particular task is much better done by a machine than by a human being for whatever the good reasons are, then fine, let’s give it to a machine. That active hybrid workforce management is where we’re headed.
And I see it emerging right now. I see it in robotics, I see it in the application of AI and machine learning algorithms to augment human beings. So, the future state is you get to a very, very efficient back office.
And the single KPI that I keep in my mind when I think about this — I put myself in the shoes of a chief operating officer and I say, what is that person really concerned with? They’re concerned with dollars per employee — how much revenue does the company make, divided by how many employees you have, and they’re trying to continuously optimize that KPI.
LYNCH: They can optimize the KPI by driving the numerator, driving loyalty, driving customer journey, driving sentiment, interaction, driving NPS, as an example. That’s how you drive the numerator. You drive the denominator by making effective use of the employees that you have. You can drive the denominator to zero and get a really good dollars per employee, but you don’t have any people and that’s not going to be an effective business. So, there’s a balance point.
And that future state, if you think about dollars per employee that gives you the hint about where we’re headed. They’re trying to manage dollars in the numerator, sometimes dollars in the numerator require human beings because you need people to hold hands and do customer support and that kind of thing.
CUOMO:Yes, makes sense. Ed, could you just quickly comment regarding future state? The market seems to be split and quite galvanized on business automation and IT automation. Will it continue to be that way?
LYNCH: I don’t think so. I think the right north star to be thinking about is whether you’re in IT or you’re in business, you have people — you have human beings — and the people are doing something. If you think about the optimal way of deploying the people, whether it’s in IT or in business, that’s where automation comes in.
And so, whether you are optimizing the work of an operator or you’re optimizing the work of a sys prog or of an SRE or of an invoice processor or somebody who’s doing claims, it all comes down to the human beings.
And that human being thing — obviously human beings are extremely valuable. We don’t want human beings to be completely displaced, because then you have no business. But the proper application of human beings, whether it’s in IT or in business, I think it’s kind of the same. And the same technologies, the same automation technologies come into play. And they’re slightly different, obviously. You’ve got, in ticket management it’s slightly different than processing claims. But it’s still kind of the same. And so, that’s where I see things going.
It all starts with data
CUOMO:Makes sense, yes. And Ed, last question for you is, where do you recommend a user start to best position themselves for the future that you’ve just shared with us?
LYNCH: Well, I think from a user point of view, like when I talk to my clients and I say, how do you get going with this stuff? You start with the data. You have to start with the data. And the data — how rich the data is, you know, how plentiful it is — that’s going to tell you what pot of gold you’re starting with.
For worker bees who are trying to get into this space, I would recommend data science. I would recommend focusing on skills that human beings are good at and making sure that those suckers are on your resume. Like you’ve got stripes on your arms saying, yes, I’m good at this stuff.
CUOMO:Okay, Ed, that’s certainly impressive and inspiring. I want to thank you for being on the Art of Automation today.