Chat with Lee Coulter, The "Godfather of Cognitive Automation"

‍

Pascal Bornet, Chief Data Officer, Aera Technology, recently sat down with Lee Coulter, Operating Partner, Acresis. Coulter can easily be described as the "godfather of cognitive and intelligent automation. In this wide-ranging interview, the pair discuss Industry 5.0, digital transformation, and the future of intelligent technology.

-----------------------------------------------------

What follows is an edited transcript of the video.

Pascal Bornet: Hello, everyone. Welcome to our Cognitive Automation Community Fire Chat. Today, it's my pleasure to welcome a distinguished guest. We have with us today Lee Coulter. Lee, how are you today?

We are super happy to have you with us today. Lee, over the coming seconds, I'll try to introduce you, okay? It's a very complex exercise given the richness of your professional life and career over the last years. So, first of all, in my view, for me, you're the godfather of cognitive and intelligent automation. You're the chair of the IEEE standards on intelligent automation, IEEE being the organization in charge of setting the global business and technology standards. In this context, has set the first definitions and frameworks of cognitive automation, and I think it was in 2016, '17. Can you remind me? When was it?

Lee Coulter: Yeah, we started in 2015 and our first standard was published in 2017. Then we released one in 2019, and our third standard implementation and management methodologies was actually published on Tuesday, IEEE 2755.2-2020. We'll find that standard. We're pretty excited about it. We'll be doing some PR on that here over the next month.

Pascal Bornet: To give a bit of background yourself, you are 30-plus year executive. Your career is just amazing. You've held leadership positions in prestigious companies, such as General Electric and Kraft Foods, and more recently Ascension, transformAI. I mean, really a challenge to summarize such a rich career in a few seconds, but here is a tentative.

You started at GE for 15 years, where you held multiple leadership positions in the healthcare, capital business divisions, and in the IT business outsourcing. After that, you moved to Kraft Foods as a chief administrative officer, where you led the global implementation of more than a hundred shared service centers in more than the nine functions across the world. After that you joined Aon Corporation, where you led the global IT transformation. You then moved and founded and led Ascensions, which is a globally recognized captive business process outsourcing. A few weeks ago, you just completed the sale of transformAI, basically a hypergrowth automation business that you built and led for the last two years. Congratulations for that, again. On top of this, you have published more than a hundred papers. I mean, we've seen you in many podcasts. A recognized thought leader on the topic.

Lee Coulter: Thank you. I'm really excited to be here with you, Pascal.

Pascal Bornet: Thank you. So, without waiting further, let's start with the first question. I've heard you speaking about Industry 5.0. What is the difference between Industry 5.0 and 4.0? Can you tell us more about that?

Lee Coulter: Sure. It's a really important topic, and some of the best thinking around this actually comes out of Singularity University, and I'm a member of Abundance 360, which is related to that. So this notion of Industry 4.0 and 5.0 are based on the key differentiator is convergence. What does that mean? The power of telecommunications of 5G by itself. So we'll take 5G as one thing. It has an innovation curve and its advancement, in terms of its speed and latency and bandwidth, is A.

Then separately, so we'll put that aside, we'll just say that's 5G. Then separately over here we have IOT. IOT is a whole field of sensors and monitors and data. This likewise is on an exponential curve, the world of IOT devices, and then maybe I'll put cloud and big data as another area of innovation, which is moving on its own innovation trajectory.

Each of these, as you look at them, we could say that's an Industry 4.0 thing. We're seeing rapid disruption. We're seeing incredibly exponential growth in the innovation in the area. Now, when you contemplate the combination of these things, it's like one plus one plus one is 75, and that's really Industry 5.0, so what does that mean? So, if I just take 5G and IOT just by themselves, and I apply that to agriculture, we now have fully automated farms that nobody ever visits. The equipment maintains itself, every plant gets watered and fertilized specific to its needs, the weeds are burned out with lasers by robots that go over by, drones monitor the health of every single plant on the farm.

This is not science fiction. This is the real power of convergence, when these different technological areas now come together, and they themselves, each one of them on an exponential path, has exponential to a power when you put them together and you look at the power of these things together. There's a host of examples of different places that individually we're seeing exponential innovation, we would just call in isolation 4.0, but what's really happening so fast around us right now is the convergence. I'll use a quick example for that.

So, there was an X prize about, oh, it wasn't long ago, maybe less than a year ago, and it wasn't a big one. I think it was $500,000, and it was to the first team that built a working tricorder from Star Trek. It had to perform I want to say 12 or 13 vital functions, and the X prize was one in like six weeks. This team, I want to say it was somewhere in Asia, I want to say it was in Japan, they basically, pulling all off-the-shelf technologies, assembled a tricorder built around a smartphone, and they won this thing in a month and a half. They expected it was going to run its full three-year course. No, because in sensors and in telecommunications and data and everything that they needed in order to create a tricorder all of those technologies were moving very, very fast in their own respect, and when you put them together, that tricorder is really an Industry 5.0 sort of dynamic. So did that make any sense, Pascal? I know it was a lot of words.

Pascal Bornet: So it's about the convergence of technologies that are currently existing, and these conversions create synergies. The first question that comes to my mind is what is preventing us to get there right now?

Lee Coulter: Well, interestingly standards are one of the big challenges, and we'll just talk about IOT, telecom, and big data. We don't have globally accepted standards on how IOT data shall be gathered, captured, and stored. So, if we had, and you know I'm a fan of standards, because standards allow the industry to innovate much, much more rapidly. I ask this question all the time: would you personally go purchase a router, an internet router, a wifi router, that didn't say 80211 dot something, right? No, of course not, because that's the standard. In every USB stick you put in your computer, it fits and it follows USB 3.0, 4.0, 5.0. This really allows for dramatic impact to... The standards speed up innovation and our ability to collaborate and co-innovate together, which I think is really important.

So, if we look for evidence of Industry 5.0, we see its nascent, its beginnings. I follow with great interest full self-driving, or we'll just call it level one through level five autonomy in vehicles. I'll be more vague about that, because most people don't realize that most of the cargo ships today, except for the time spent in port, it's fully autonomous. Aircraft, except for the time leaving and arriving, it's fully autonomous.

And arriving it's fully autonomous people don't really realize the level of autonomy in other vehicular situations we have. But each of those is an example of this convergence of many different kinds of technologies that make this sort of capability possible. And as we look out at the things that are emerging today, or that have emerged in the last 2 to 10 years, a lot of them wouldn't be possible if some of these individual technologies had not all been moving through just exponential innovation and change. I even look, Pascal, at intelligent automation. And we'll talk about that in some significant aspect, but here's a great example of the convergence that has occurred in just the last 12 to 24 months, it's completely changing the landscape of what IA is, how it's put to good use, how it's consumed, et cetera.

Pascal Bornet: The question that comes to my mind is, so when when will industry [inaudible 00:12:27]? And I remember we've seen industry 4.0 kicked off by the book by Klaus Schwab. It was maybe in 2017, '18, something like that. So it's four or five years ago. Do you think the cycles will become shorter and shorter in the future?

Lee Coulter: If you look at industry 1.0, 2.0, 3.0, 4.0, they have relatively long periods of time and believe it or not industry 1.0 involved agriculture. And this was probably the most fundamental innovation in the history of humanity for the one very specific reason that it enabled specialization in cultures and groups in cities and villages, whatever unit you want to talk about. Before everyone participated in subsistence nutrition, right? What they gathered and killed that day they ate. And once agriculture was around now, all of a sudden you can have a black smith and you can have a furrier and you can have an animal husbandry person. And you can have these people who specialize. And if you look at the whole rest of evolution, irrigation and agriculture are the foundational industry 1.0 innovations. You could argue fire also among them.

So some of these cycles we could say are 15 to 50,000 years. If you get to industry 3.0 and 4.0, we're talking 70 to 90 years. Industry 4.0 probably more like 50 years. And then here we are looking at 5.0, and I've made this statement before, that only by somebody sitting in the future and looking at the data, doing analysis, will be able to say, "Here's more or less when we started this industry 5.0 thing." Because really in my mind, most people don't realize that tablets, they were debuted to the world in July 2010. And tablets really are an example of industry 5.0, chip fabrication and telecommunications and display technology and battery technology. All of these individual technologies had to converge in order for us to get a tablet. And then you have software innovation, like an application platform and so on and so forth and standards for them to work.

So there are pieces of evidence of 5.0, pretty much wherever you look. The question is always about adoption, the speed of adoption. And of course we know from the adoption curve, right? We have the innovators, the early majority, late majority, the laggards and different groups will adopt in different periods of time. What we've seen very recently here though, in fact, you and I are a part of it right now, we're doing this fireside chat using remote technology.

Well, there's a billion five, a billion seven human beings that 13 months ago could not have imagined performing their job right remotely. And the pandemic forced radical adoption rates. Well, what does that do? Well now everybody has this base level of capability in their home, their kitchen, their library, their basement, wherever, and what that was all allow for in terms of the further separation between workplace and work space. I'm pretty excited about it. And I'm happy to diverge there, but that's another case of where, when we see the pace of things, adoption is usually the long pole in the tent. It is the thing that slows the whole process of innovation down.

Pascal Bornet: Just to transition more into the topic of work and workplace. Do you think the pandemic and the context in which we are has accelerated this transformation in bringing us closer to industry 5.0? Especially in the workplace?

Lee Coulter: Yeah, for sure. The acceleration, and if you look at the flow of VC dollars and PE dollars into, and I'm going to use a word that Cisco used to use for a product, but I'm going to use it as a general term, which is telepresence....the airline industry it's going to die, there will always be some amount of actual physical travel, but for a great part of the need to collaborate and to work together, we've been prying apart the workplace and workspace probably since telephony was really introduced into the workplace. So probably back in the '50s and '60s, when two people could collaborate together over at least to telephone connection, right? That was probably the first separation of the workplace and workspace. And of course, now we have virtual reality, mixed reality. And I don't know if you've had a chance, Pascal, to try the Magic Leap or the Hollow Lens. I have had the good fortune to try these technologies. And they made the hairs on the back of my stand up. The Magic Leap technology did.

...the level of reality for the interaction with that projected presence was so uncannily real that you can see just... Some of this stuff's available now, right? So for those of us who come out of the shared services world, they have the robot, which has wheels and a screen on a stick, and somebody can literally drive it around and into a conference room and be participating in meetings. That's early 5.0 in the telepresence stuff. Now you can imagine a full-blown avatar and I need to attend a meeting in Tokyo. And so I simply rent time on an avatar there and connect my rig in my home office and fully project myself into that area. I was blown away, Pascal. Home Advisor, which is a website for people who need help with home repairs and upgrades and remodeling.

They now have a section on home telepresence and home VR rooms. They have whole section, and this is a light bulb for me that says, this is hitting so mainstream that something is mundane, where you go to look for a good plumber or electrician is now offering advisory services to set up your VR room. How to be safe, you know where to put the harnesses, the hooks in the ceiling, how many cameras do you need in your room? What does the lighting need to be? What does the color need to be? What does the total dimension need to be? And how do you overlay with a projector, the boundaries of whatever current virtual environment that you're in and so on and so forth.

Pascal Bornet:  I agree with you. We see this more and more in the shops. And as you say, it's a signal that something is coming. But we don't see yet large adoption. And maybe the price point is an issue. And maybe the mentality is, I don't know whats your view on that.

Lee Coulter: Well, do you remember, and I'm going to say this is just five, seven years ago. So Cisco had a product called Telepresence and it was super high quality audio and video collaboration for companies and each room cost about $250,000 to install.

Now you and I can do this with our phones at any point in time, we can just flip the video and if we want to add more people we just start calling more people and add them into a video call. So already we've seen this dedicated technology that was the Cisco Telepresence rooms be made democratized and demonetized and available now to pretty much anybody who has bandwidth and a reasonable device. And I think we'll start to see that same evolution, I've seen some of the new display technologies, they're going with 3D display technologies, holographic display technologies.

The virtual environments that are built there's a real estate company. Their name will come to me in just a second, but they exist entirely virtually. Their office is a virtual environment and they all put on their VR rigs when they have team meetings, they currently hold the record for the most concurrent sessions in a VR environment. I think they had almost 800 people in a single VR environment and all of their interactions with their customers are in a virtual environment. As you can imagine, it's property so perfectly adaptable to high levels of 2d and 3d representations in a virtual space. But this is just an example.

Pascal Bornet: Virtual viewings and so on.

Lee Coulter: Exactly.

Pascal Bornet:I see a big impact of this in, as you just mentioned, in commercial activities, sales activities, and collaboration generally in the work place, do you see an impact of this in the productivity?

Lee Coulter: WhenI was at Ascension, I had a skunkworks and we were doing all sorts of interesting experiments around workspace and workplace and the ability to augment an individual person's productivity. So as an example, how accurate are retinal trackers? Well, it turns out that retinal trackers are really, really accurate to about 0.6 millimeters at four feet. So you can actually tell if somebody is looking at the screen, you can actually tell what part of the screen that they're looking at, what are they reading and why is it relevant to the piece of work that they're doing? So when I talk about, and I think about -

Pascal Bornet: It can be like glasses as well?

Lee Coulter: Yes, absolutely.

Pascal Bornet: We are still on the same computer, but yeah.

Lee Coulter: Or actually you don't need anything. So we were looking at a little pin cameras arranged in a grid versus the glasses that have inward facing cameras. But here's a case where miniaturization battery and telecommunications and bandwidth and all these things have converged to make it possible for you to put on a pair of glasses and I can know exactly what you are reading on the screen. That is 5.0 delivered as a microservice in this world of automation. And our world of automation, I think continues to grow and expand as well, Pascal. So of course, most of our automation work for human history until about 30 years ago, 50 years ago was all around physical automation, automation of physical activities. And so those capabilities are being converged with advanced technology information related technologies that are changing those worlds.

And I referenced the farm, right? It's managed by six people who never actually visited the farm. Everything that they do is through a keyboard, a hundred thousand acre farm run by six people who never actually go to the farm. It's remarkable, there's a great case study out there on that. And so as we look at the office or the productivity of large enterprises, these technologies are... We went through a radical acceleration of adoption, which provides a foundation now for the introduction right on the heels of this additional disruptive technologies. And I don't know, I'll just stop there. There's a hundred places I could go with that line of thinking.

Pascal Bornet: I think you should write a book on that. There is so much insight in what you say. The workplace of the future and what surprises me is from the beginning of our conversation, we haven't talked about artificial intelligence, about machine learning. So that makes me think that it's a given, it's it became as Google CEO would say, like fire or electricity, it's a basic, it's a given point. Now we're building on top of it. It's amazing.

Lee Coulter: So if you look at fundamental innovations throughout the eons of human existence, a few standout, we mentioned a couple, right? Irrigation, fire, plastic. The discovery of plastic has changed the world, right? In so many different ways. The invention of refrigeration, the invention of obviously a transistor and these are fundamental innovations that have fueled humanity over time. And if you look at the starting points for when each of those things happen and you overlay your exponential graph. It gives you a good indication of why we're seeing so many things moving at just ridiculous rate. Four times the capability at half the cost every 18 months. I mean, this is happening in real life, real time for us right now, because so many things are moving so fast.

And if I look at this separation of workplace and workspace, it introduces for an enterprise, a whole new set of questions. How should I structure workflow and workspace and workplace? And you mentioned AI and you and I had this conversation. I try to stay away from that term because it's so ill-defined, I prefer augmented intelligence or machine learning. The singularity, the point at which the experts judge that we will have machines that demonstrate true intelligence is somewhere between 8 and 13 years from now. So depending on who you like, somewhere in the late 2020s to mid 2030s, that this going to happen, this is not a question of, "Oh my gosh. I wonder if it will. I hope it will." It's guaranteed. It's coming. And so data and data management are in this world.

When you look at it in isolation, something like Cloudera is fantastic on its own. Now bring cloud and containerization and microservices and AI to the game. And what's possible in terms of how we reimagine the work that we do in our enterprises. There's something called zero base design, which is one of the big consulting businesses likes that term.

And it really just means clean sheet of paper it. I mean, it's a fancy, $1000 an hour consultant speak for, if you were going to build it today, how would you build it regardless of the constraints of your current reality? If you had a fresh, clean sheet of paper, how would you build it? And this is where things like Rocket Mortgage come from. Rocket Mortgages consistently giving, concluding mortgages in 24 hours or less. Something that has for the entire prior time on earth, been a six plus week process. So when you now bring clean sheet of paper thinking in, and you have industry 5.0 converged technology capabilities together, then our ability to re-imagine what portion of the work should be performed by who or what and then does when and where matter.

When and where matter, and what technology tools can I bring to bear to improve the experience of the people involved. Those performing the tasks and those experiencing the service, or the work being done. The accuracy and consistency, speed throughput. All of these things that when we think about large enterprises. These are big topics we wrestle with. Big companies have big stuff. You have 50,000 vendors and you have 3000 offices and you have a quarter of a million people. These are big numbers, and there's a lot of work that goes on that's open for transformation.

And what has happened to, and I'm going to use another term and you and I, we've talked about this as well, RPA, which I worked really, really hard to get that redefined differently. But RPA, when it began was a specific tool, and it had a set of use cases that were specifically tied to RPA. And then there was RDA and they had a set of use cases. What happened was [crosstalk 00:34:29].

Pascal Bornet: RPA for robotic process automation, RDA for robotic desktop automation, which is UI-based automation.

Lee Coulter: Yes. Human in the loop, interactive automation, RDA. So, that shifted materially in the last 18 to 24 months. So the, the conversation about, "Oh, are you doing RPA? Are you doing RDA?" Even, "Are you doing intelligent automation?" Those were questions about what was thought of as a very specific thing, your specific family of technologies that does a very specific thing. And what's happened is, we're now bringing that set of tools to the workbench as we re-imagine work. And it's now being combined with very sophisticated... I'm going to use a term, "AI tools," that can be now purchased as a microservice for 10 thousandths of a penny via the cloud. And I always use the example of OCR, because OCR is an incredibly complicated thing. And It requires a ton of computing power.

And now for pennies, fractions, tiny fractions of a penny per word, you can rent these services. We were doing some videos like this, and in one hour, we can upload this video and have both closed caption, we could have everything that we said under our names, with a 100% accuracy, returned to us in an hour. And we would pay about 10 bucks per minute. This is an example of just how dramatically different our workplace can be.

If you've run in a global organization and you need to communicate across borders, across languages and cultures, you could imagine having a near real-time service, which maybe you record a meeting, an important meeting. And then within an hour, you've had it subtitled and translated into the seven languages you need and distributed to the teams that need to interact with it, all within a couple of hours. And it changes, it just changes the nature of the work that we should be focused on, and opens up fresh thinking to all of the work that we currently do in an enterprise.

Pascal Bornet: That's amazing. Thank you for sharing this vision. And I share it with you.

Lee Coulter: I know you do.

Pascal Bornet: All the benefits we can get from those technologies for the future of work. And I really believe that it will help our work and our work place to be more human. And so I'm conscious of the time. We want this video to be short in time, [inaudible 00:38:16]. Last word from you, how would you end this discussion? What are the key thoughts that come to you for finalizing this?

Lee Coulter: Couple of key thoughts. One, pay attention. Pay really close attention. And what do I, what do I mean by that? I mean, actually start different information feeds. If you have a news feed, go in and learn about what's happening in quantum computing, learn what's happening in 3D printing, or drones. Each of these spaces Is literally an industry 5.0 field itself. And the explosive changes occurring there, and the use cases of all of them are just extraordinary.

So, there are less than a dozen technologies that are 5.0 technologies that are going to shape the next two decades. And they're each changing so fast that if you do not make a deliberate effort, and go seek the information, you will not believe how much changes each 12 months. And I share this with you, having been a student of this for five years, through my affiliation with Abundance360. And it really impressed on me the importance that I have to go in search of this information. Because it's advancing so fast that if I just stopped watching it, if you didn't watch what happened to Blockchain420 for months in 2018 and '19, you wouldn't believe what happened from this arcane technology, to all of a sudden the Bitcoin millionaires. So, I think that's probably number one.

And the second, final thought I have for people is, as you are thinking about innovating, and re-imagining the way work has done, and what work is done, hold your constraints off in a separate pile. Put them over in a pile. You can bring them back in, but to free your thinking, don't allow your constraints to kill the ideation that can come from tackling a problem with a clean sheet of paper, when you have the tools that we have today. Those are my two points.

Pascal Bornet: Yes, yes. So your point is about, don't, don't think about the constraints of current technologies, but think beyond. And don't try to reproduce, or imitates whatever is being done today and the way we do it today, but really think out of the box and recreate, redesign.

Lee Coulter: Reimagine it. And it's not just the constraints of the technology, but sometimes we think that we have an operational constraint in our company. "Oh, well, we have this regulation, or we have that policy. And we often use those as crutches to limit how much free thinking we can do about how to reimagine that work. So, my suggestion is whether technology-based, cost-based, policy-based, regulation... What the source of those constraints might be, hold them at bay until you're done with your thinking about re-imagining the work.

Pascal Bornet: Excellent. Thank you for those final words. Lee, it was really a pleasure to have you with us today. An amazing discussion, delightful vision of future of work and industry 5.0. So if you want to follow Lee, we'll flash on the screen how you can connect with him. And stay tuned for the next Fireside Chat's, on the automation and the future of business. Thank you to the automation community purely. Thank you, Lee. Thank you all.