The State of IT Automation: New Pressures Invite New Opportunities Read Report

Episode #6: Insights from IBM: Digital Workforce and a Software-Based Labor Model

In today’s podcast we interview Gene Chao, Global Vice President and General Manager at IBM Automation

Automation, Artificial Intelligence, and Machine Learning are disrupting work environments in nearly every major industry and job function around the globe. This tectonic shift in how work gets done is almost as breath-taking as the pace at which the change is occurring. Depending on one’s perspective, this digital transformation is either exciting, unsettling, or even both. If however you run the automation practice for IBM, a Fortune 100 company, your perspective is uniquely broader, and decidedly more enlightened.

As Global Vice President and General Manager for IBM Automation, Gene Chao probably knows as much as anyone else in the world about the current state of these technologies, and the impact they’re having on the enterprise. Gene joins us to share his views on what a digital workforce with a software-based labor model will look like, why organizations clinging to rigid centralized hierarchies may not remain competitive much longer, and which metric is the worst indicator at gauging the effectiveness of automation.

Read Full Transcript

Guy Nadivi: Welcome everyone. My name is Guy Nadivi, and I’m the host of Intelligent Automation Radio. Our guest on today’s episode is Gene Chao, Global Vice President and General Manager for IBM Automation. Gene leads the entire global spectrum of IBM’s automation business with responsibilities that include building and developing IBM’s offerings across automation, artificial intelligence, and enterprise solutions. Prior to leading IBM’s automation business, Gene had a distinguished career with such global IT leaders as CSC, Hewlett-Packard, and Accenture. And most recently prior to joining IBM, Gene was the Chief Revenue Officer of IPsoft. And I can’t imagine too many other people having as broad an insight on the global automation market as Gene Chao. And so we invited him to come on our show, and we’re very pleased he’s able to take time out of his extremely busy schedule to join us today. Gene, welcome to Intelligent Automation Radio.

Gene Chao: Thank you so much Guy. Really excited to be here, and I think the topic is top-of-mind, and hopefully the audience agrees. We have a lot to talk about, so thank you again.

Guy Nadivi: All right. Let’s dive in. Gene, not too many people know this tidbit, but the company that eventually became IBM started out in the 1880s I believe building scales. So I think it’s appropriate that I start out today by asking you to “weigh in” about intelligent automation, which you talk a lot about in your own presentations, and of course is what our show’s all about. How do you define intelligent automation?

Gene Chao: Well, it’s a great question, Guy. “Automation”, amongst other terms in the industry, have become so broad, just like “AI”, just like “platforms”. We actually mean something pretty specific here at IBM. The way we define it is, think about automation has traditionally been about routines, repetitiveness. It’s about a static way of working, or a static workflow. This holds true across physical robotics, assembly line type of things, as well as what I’ll call the first generation of your knowledge worker automation. Think of things like advancing macros, or even elements like screen scraping. However, with the advancement and the ability to take advantage of new artificial intelligence elements, the advancements of machine learning, really good insights into patterns and anomalies, those micro data and business insights, a workflow rebalancing has shifted that static workflow into a dynamic workflow. To us, this is intelligent automation. Not only being able to create the static aspect of it, but the ability to learn, to adapt, to interpret with the advent of machine learning and AI, to start creating an autonomous workflow. Things that understand, things that it will be able to self-adjust and self-correct. This is the new way of working.

Guy Nadivi: So now speaking of workflow, you’ve stated that people following a process supported by technology is basically a dead-end paradigm, and the future will be one where processes are run by technology, supported by people. Can you elaborate a bit on that, and perhaps give an example where you’ve seen that next generation approach implemented?

Gene Chao: Oh, absolutely. And it’s interesting. What started out as sort of a pithy, hooky type of statement has really become the cornerstone of our methodology. And if you don’t mind, let me work backwards a little bit. You asked about a specific example. The best example I can give is a business, look at Netflix, right? It’s a perfect example where recommendations, the insights into who’s watching what and when, those “jobs” if you will are done by algorithms, by bots. They’re not being interpreted by a human or a person. People are there to support the system. They’re there to support the algorithms. And that’s what we mean by this whole shift into technology running the process, supported by people. It’s not that we don’t need people. People are a very important aspect of development and certainly the management of these things, but processes have traditionally been designed for people to follow. The new sort of paradigm as you call it is really around rebalancing that workflow. What people do, versus what the systems and algorithms can do, that’s fundamentally changed, which is why we’re trying to introduce the concept that those automation technologies, those things that actually trigger those processes, those are done by systems and technologies, rather than people igniting that spark. Does that make sense, Guy?

Guy Nadivi: Yeah. So Gene, in that future where processes are run by technology, you’ve predicted there will be a digital workforce comprised of software-based labor. What will that look like?

Gene Chao: Oh, that’s a really good question. In what I’ll call a classic model, there’s a people-only aspect, and then there’s what I’ll call a people plus machines, or a people augmented by technology aspect. And that was good. That was a necessary step as we took a look at what our computing powers can do, what the software components can do. It’s what created the maturity of things like systems of record or ERP systems. But now you introduce a third part, which are what I’ll call the digital workforce or the virtual co-workers. And now you have a three part framework that looks like people only, people plus machines, and then a software-based labor model. That’s what we mean by the digital workforce. And there’s actually two sides of it. There is a digital workforce which is literally a software-based labor model, and then there’s what I call digitally-skilled people. So how do we move from being a pure back office accountant into an accountant that also manages technology? So those are the things that we comprise as a digital workforce.

Guy Nadivi: You’ve talked about how automated processes that touch multiple departments divided by walls will eventually cause those walls to come down, and that suggests a flattening of the organizational hierarchy away from a centralized paradigm. And as most people know, flatter organizations tend to move quicker than hierarchical ones, and generally also require employees to take on elevated levels of responsibility and greater involvement in decision making. So as automation becomes more ubiquitous, do you think organizations that cling to rigid hierarchical structures will still be competitive?

Gene Chao: That’s a very emphatic no. Those dynamics and those changes are already happening. So a couple sort of, let’s double-click into that. There’s two aspects into what I want to talk about. First is, we have the concept of trying to move everything into a stay through processing model. So we’ve taken a look at all of our work, whether it’s supporting our clients, running our business, even the way we provide consulting type of projects, and we’ve classified that work into where those routines are, what are those repetitions where there’s unique knowledge or unique work, and we’ve tried to reshape those workflows into everything straight through processing. So it becomes at a machine time, a machine speed if you will in terms of how that stuff gets done, how work gets done. That completely takes away the organizational hierarchy, because as you think about vertically integrated processes, and I’ll use the case of a banking environment. Banks have their customers, customers may have a checking and savings account, might have a 401k or investment management. They might also have a mortgage. How do you maintain those stove pipes while saying, “I’m gonna be customer-first,” and really understanding the customer relationship aspect of that? You have to break down those departmental walls. You have to get through an ability to transact, scrape through without hiccups, and you have to tear down those hierarchies.

Now, easier said than done. You touch upon one thing that is critical is the requirement of employees to take on elevated levels of responsibility. Absolutely. You have to become cross-functional which means you have to be cross-trained, and you have to be able to look across the organization and upside-down and sideways to figure out the right organizational dynamic. Those workflows will drive those new organizations. So long winded way to say “no, many businesses have already changed, they’ve already decentralized some of those paradigms”, and those are the ones leading the market today.

Guy Nadivi: It’s pretty clear automation is heralding some dramatic shifts in corporate cultures, and I’m curious, what are currently the biggest bottlenecks IBM is seeing that are preventing wider adoption of automation, and more importantly, how will those be resolved?

Gene Chao: We could probably do another show on that one. And I wouldn’t say anyone’s technically wrong, it’s just that there’s a fear factor in this. But let me land on two key things around the bottlenecks. The first one is, having a belief or trust that the technology can actually work. There’s just a general fear that’s created. So is my job gonna be lost to a bot? I wouldn’t worry about that too much, because we’re not in a place where complete swaths of people are gonna get completely eliminated, right? So there’s a fear over if I adopt this quickly and I adopt this well, we’re gonna lose our human intelligence. We’re gonna lose our human capital. That fear factor paralyzes a lot of companies. And we have this phrase which is around automation and/or AI that is responsible. Those are the types of things that we’re working on.

The second big bottleneck is really understanding the design principles of how to do this. Let me kind of shorten that and say, many of our clients are using these technologies as sort of an extra tool in their backpack as opposed to really taking a look at fundamental charter, role, and job changes in that adoption. They’re just saying, “Hey, maybe there’s an RPA, robotic process automation technology. We’ll just kind of shove it into that process, and we’ll find two hours of efficiency.” Those little points of light does not add up to an economic change. Those little points of light, as we joke around, we call those longer lunch breaks. So we’ve created some efficiencies and some productivity, but that wider adoption, that automation at scale as we say, never really happens. So we start to take a look at the strategies and the blueprinting at a wider adoption level.

Guy Nadivi: Switching gears for a moment, taking a broader view Gene, I’m curious from your vantage point, what industries have you seen benefit the most from intelligent automation?

Gene Chao: That’s a great question. There’s a bit of a complex answer there. Let me answer that in two ways, and really in a pivot. By industry or sector, we’ve noticed that those that are highly digitized, so think about banks or investment management companies, think about companies that use that next set of digital data. Those are the ones that have been the early adopters. So if I could pick the top three in terms of adoption, I would say banking and financial services. I would say the second industry is around industrials and/or manufacturing. I think that’s because they’re already used to the robotics element. And the third one that’s been up and coming very, very quickly for us, funny enough, is consumer goods and retail, because they’ve already introduced the topics on sort of a chat bot or the buying assistant. You go onto websites now and there’s a buying assistant there or a bot to help you through a purchasing process. They’ve now taken that one step further.

So those are the three industry sectors that I would focus on in terms of earliest adoption. However, if I pivot to a different approach to how they use automation, there’s sort of a three-part answer. The first one is around IT service management. How do we fundamentally take back office processes and automate the heck out of it? How do we move to a virtual engineer or a virtual service agent model? The second one is really around the running-the-business processes. Think about accounting, think about HR and recruiting. Think about internal sourcing and procurement. Those are the first two areas that have been almost to the degree of table stakes that you have to go address through an intelligent automation engine. The last one is a focus on that front office. This is where that banking or financial services approach has been leading the way. Think about things like a “robo-adviser” for investment management. They’re now flowing that all the way through from advice given to their customers, making sure this compliance and regulatory aspects on doing things like tax returns and investment management. The front office has really started to adopt this approach. So that’s almost in the stack ranking. Biggest, widest adoption and IT operations and service management. Right underneath that would be the run-the-business processes, your non-core processes, and the third would be in the core business processes that they have. That make sense?

Guy Nadivi: It does. And it makes me wonder, the financial services, manufacturing, consumer goods and retail, those are three very different industries, and the processes you discussed were also very diverse. So I’m curious, from all that diversity, is there a particular metric you like above all others that best captures the effectiveness of automation to those types of enterprises?

Gene Chao: Absolutely. And I’ll actually start on the other end of the spectrum where automation has classically been viewed as a cost or productivity play. So what’s been happening is expense ratios, and I think funny enough the worse one for me is around managing headcount out of the business, so people reduction. Those are to me the worst indicators of effectiveness, because there are disconnects between operating model, use cases, and translating that into your operating expense, in this case people. We try to advise folks to go beyond that and get into taking a look at the turnaround time of a process. Measure the run time of how long a process takes, the effectiveness of the outcome. We’re introducing topics like different types of resource units. I talked about the digital workforce. Think about a human resource unit versus a software-based resource unit. Humans are managed in terms of 2,000 hours a year. Machine time, robotics, or automation is measured in 6-7,000 hours a year. So just by the effectiveness of the software technologies, that entire metric has changed. And we’re in the middle of trying to drive the new commercial terms around that. So those advancing commercial terms are really what’s gonna be able to capture the effectiveness of what we’re doing.

Guy Nadivi: You mentioned unrealistic metrics, and that leads me to wonder what are some of the most unrealistic expectations about automation that you’ve encountered?

Gene Chao: Yeah, it’s sort of coupled, and I’ll pick on the RPA community a little bit. It’s “Hey, adopt RPA and you’re gonna save 40% tomorrow.” I just haven’t seen that happen. Unrealistic expectations are really centered around the speed at which you can get these economic gains. There’s a lot of design principles, there are very many touch points and dependencies along the way before you really see a new operating model in terms of that work flow. So don’t think you’re gonna get sort of an in-quarter return on investment, especially when that bad metric I talked about in terms of people or head count take down, that doesn’t happen very quickly. And when you start looking at the cost benefit of it, you just gotta be careful. So that’s sort of that speed to realization of economics. That’s the worst one.

Guy Nadivi: Gene, for the CIOs, CTOs and other IT executives listening in, what is the one big must-have piece of advice you’d like them to take away from our discussion with regards to implementing automation?

Gene Chao: I think it’s actually two. I quickly mentioned one. The first one is understand the enterprise impacts to what you’re doing. 60-70% of the design principles of intelligent automation are common regardless of domain area. So having an automation element running your infrastructure, the baseline or common service areas are very similar if not the same to your front office processes. Now domains are obviously in question and have a uniqueness to it, but there’s a lot of what I’ll call baseline work that has to get done. So consider your enterprise strategy. The second one is be very specific on what I call “use case hunting”. That economic return, that operational return, be really specific. Can’t just be, put that in your backpack, we’re all good. It’s gotta be mindful, and it’s gotta be thought out across functional design areas.

Guy Nadivi: Use case hunting may be my new favorite term in the automation glossary. All right, looks like that’s all the time we have for on this episode of Intelligent Automation Radio. Gene, thank you very much for coming on the show today and sharing some really intriguing insights about the global automation market. It’s been great having you as our guest.

Gene Chao: My pleasure. Always fun with you, Guy, and look forward to the next one.

Guy Nadivi: Gene Chao, Global Vice President and General Manager for IBM Automation. Thank you for listening everyone, and remember, don’t hesitate, automate.

Gene Chao Photo


Global Vice President and General Manager - IBM Automation

Gene leads the entire global spectrum of IBM’s Automation business, with responsibilities for building/developing offerings, skills, and competencies across automation/autonomics, artificial intelligence, and enterprise solutions. Gene’s team encompasses this foundational business area through market/client advocacy, thought leadership, delivery assurance, and developing IBM’s ecosystem and client engagement models.

Throughout his career, Gene has held general management, sales, and executive leadership positions in the business & IT services arena. He most recently arrived from IPsoft, where he was the Chief Revenue Officer. Gene’s experience spans business and technology consulting, as well as managed/outsourcing services with leading companies such as CSC, Hewlett Packard, and Accenture. Additionally, he also has broad corporate finance experience as he began his career as a financial analyst with Shearson Lehman, and also resides on the Board of Trustees for Active Weighting Funds (ETF provider).

Gene can be found at:


Twitter: @gene_chao


More information about IBM Automation can be found at this link:

Listen To The Podcast