Cognitive Computing and Digital Transformation

Michael Krigsman: Welcome to Episode #225 of CxOTalk. I’m Michael Krigsman, industry analyst and the host of CxOTalk. And today, we’re going to be talking about cognitive computing, and digital transformation. Before we dive in, I just want to say “thank you” to our video streaming partner Livestream, who is amazing, and they help us make CxOTalk. And, if you go to Livestream.com/cxotalk, they will give you a discount.

So, our guests today are two people who … two really, really smart folks. And, Guest Number One; we have a panel; Guest Number One is Anurag Harsh, who is a senior executive at Ziff Davis. And, Guest Number Two; I shouldn’t say Guest Number One and Guest Number Two; our second guest is Shanker Ramamurthy, who is a bigwig at IBM. Gentlemen, how are you?

Anurag Harsh: Wonderful.

Shanker Ramamurthy: Pleasure to be here, Michael.

Anurag Harsh: Thank you for having us, Michael.

Michael Krigsman: So, Anurag, let’s start with you and why do you care? Why do you care about digital transformation? You’ve written a book on this topic, you have another book coming out, why do you care?

Anurag Harsh: Why do I care? Because I think we should all care. I think, you know, if you don’t care, that’s the next revolution that’s happening. I mean, it’s sort of where our entire economy and industry is headed. And, every 30 or 40 years, there’s a major change in the way that industry changes and the way our modus operandi of doing business, and operating and living; and now, it’s happening every 10-15 years. So, that’s why I care. You know, I’m seeing a lot of companies getting disrupted. Look, you could look at the Fortune 500. Fifty percent of them actually ceased to exist now. And, half of those have essentially gone belly-up in the last twenty years. So, that actually tells you the rate of disruption is here. You know, if you don’t innovate, you disappear. So, you’ve got to do that. And so, that’s one of the reasons I really care, and I think a lot of people care. So, I just wrote a book to tell them how to think about this.

Michael Krigsman: You care enough to have written a book, and you have a second book that is coming out. And, I am putting the book on the screen - everybody can see that - and it’s called “Thinking Tech.”

And, Shanker, you are a senior exec at IBM. What do you do at IBM?

Shanker Ramamurthy: So, I’m responsible for strategy and market development and part of the practice of what we call “industry platforms.” My background is consulting over twenty-five years. I’ve worked in all the six continents over thirty countries, and digitization is being actually put on steroids by some of these new technologies like cognitive and blockchain. And, I feel like I’m at the most exciting part of my career because I can really see how these technologies are fundamentally digitizing the global economy and changing every aspect of how we live and how we do commerce together.

Michael Krigsman: So, you are very involved with cognitive computing, and maybe give us a brief explanation. When you talk about a cognitive system, what does that mean?

Shanker Ramamurthy: Sure. So, we are at that stage in the evolution of computing capability where machines are able to understand, reason, learn and interact with us. And, the way they're able to do that is quite different to traditional ways of interaction. So, historically, you had to train a computer using programming. So, you program a computer using a series of [...] rules, and those rules would create a small amount of data, and the data would become the system of truth.

Now, we live in a world where there’s virtually a finite amount of data, and computers are at that stage in their evolution where they can actually be trained to look at the data and then discern patterns and understand insights. And it’s not just structured data. We’re talking about textual data, video, voice, and other kinds of sound; all sorts of interesting information. And, if you can apply machine learning to that information, then you move to a paradigm where data creates rules as opposed to computer programs creating data. And, when you are on that type of model, fundamentally new applications and fundamentally new ways of doing business emerge from that capability.

Anurag Harsh: Yeah, that's absolutely true. I mean, you know, cognitive computing which is a term that IBM has coined, to lead on from where Shanker was talking about, is really about outcomes. And it's about, "What can we do with this technology," right? How can we impact society? How can we … What is the price of not knowing the cure for cancer? What is the price of not knowing whether a human being is going to make it until the end of the next five years if he's suffering from a disease? And, cognitive computing and machine learning, and a lot of this technology, is just at the, you know, at the core of all of this, right? It's … Or, what is the cost of drilling a well or finding out where you can drill? Or, you know, what is the cost to society of the child whose vocabulary at the age of, you know, let's say three or four years old, there's a direct correlation between the child's vocabulary set and his or her ability or capabilities in the far future. And so, if there was a way for understanding that and for a system to be able to interject and really make that change so that the child's vocabulary increases … You know, we're talking about a whole different level of society and benefits, right? And so, a healthier society, a more educated society … So I think that cognitive computing and this whole area has some profound, profound effects coming up in the next several years.

Shanker Ramamurthy: Also, cognitive computing is here. So, machine learning is not tomorrow's technology, it's today's technology. And, interestingly, with every new type of technology, it takes society a couple of decades to actually figure out how to completely take advantage of that capability. And, we are at that point where cognitive computing is here, it’s being implemented by the early adopters really broadly and very widely. Anurag talked about, for example, the health care industry where cognitive is being applied. For example, IBM has been working with Memorial Sloan-Kettering and it’s got a whole bunch of its own data. And together, IBM, along with some of the smartest brains in the world, are looking at how to solve oncology.

And, this is about … Cognitive technology is all about augmenting human intelligence. It's not … You know, there was a book written by a couple of MIT professors which was titled "Race Against the Machine." We think of cognitive computing as "race with the machine." How do you this […] of computing capability and human capability to solve some of the most complex problems that we are dealing with in society and in business?

Anurag Harsh: In fact, that’s a fantastic statement you made, Shanker. I would like to … I’m going to read up a quote from Thomas Watson, Jr. You know, that’s who Watson was named after, and this was obviously a few decades ago. And he said, “Computing will never rob man or woman of his initiative or replace the need for creative thinking. By freeing man from the more menial and repetitive forms of thinking, computers will actually increase the opportunities for the full use of human reason.”

Now think about how profound that statement is. It’s man and machine, which is all … really what it’s about. It’s IBM’s philosophy. It’s not man or machine. And, you know, Shanker was talking about data. I mean, data is at the core of this, right? I mean, we all know about big data, and now Elon Musk is obviously talking about connecting the human brain to a computer. And a lot of people think it's about … Cognitive computing and machine learning and AI are about trying to figure out how the human brain works and replicating that. It's not. That’s not what it’s about. You know, we only use a certain percentage of our brain. And so, the idea is to be able to figure out and taxonomize, and make sense of all the data that’s on the internet.

Now here' the thing: the majority of the data. Right now, 80%; in the next three years could be close to 90%; of all data that's out there is on the dark web. It's deep web. It's not accessible. It's inside firewalls, it's, you know, everything that people are talking about within your email networks, and it's where the government operates, where academia operates; it's inaccessible. And, to be able to get a hold of that data, to be able to, in a manner, to then make sense of it, and understand, and use that data to inform a supercomputer like Watson to then learn from it; and every single iteration it gets better, and better, and better; that's what this is about, right? It's a new Moore's Law that is going to be written in the next several years. And so, I think that it’s about changing outcomes, you know? That’s what this whole thing is about.

Shanker Ramamurthy: Indeed. And, you know, even if the data is not dark data, even if it was available, it hasn't been computable until recently. Computers were typically working with structured data; with numbers. But now, with machine learning and cognitive, we are in that era where computers can actually read a set of text and understand, and not just the syntax, but the semantics of what is meant as well.

So, imagine if you are a legal professional helping financial institutions with regulatory compliance. There are over three hundred million pages worth of regulations around the world that a multi-jurisdictional financial institution would have to comply with. I mean, no human being, or even a combination of human beings, can understand all those regulations to understand what your obligations are. Well, Watson can actually ingest that and actually figure out what your obligations are. So, if you are dealing in foreign exchange in Italy, what are the regulatory obligations you need to comply with, as an example?

You know, more recently, H&R Block has been using Watson to actually learn the U.S. tax code and help its tax professionals be more effective. You know, cognitive is being applied to augment capability in a broad range of areas.

Anurag Harsh: Yeah, I mean there's another example right here. We're just talking about it before the camera switched on about a variety of other industries. And Shanker's an expert on a lot of these and I'm just going to ask him about it, just because he is here, which is talk about some of the industries where Watson is really helpful. For example, in oil and gas, right? And oil and gas: huge opportunity, right? Every single rig, that has all these sensors, there are hundreds and thousands of these sensors all around the world. And, to be able to figure out where to dig, right? Where can ExxonMobil as a company go and dig? Or, if they actually have a rig, what is the capacity of the reservoir and how is that going to change? Because all of that affects oil prices. So, it has a profound impact on the industry.

And there’s obviously a whole bunch of other areas. There’s retail, for example, as well. I don’t know if a lot of people know that, but IBM actually gets the firehose of the entire Twitter feed from Twitter every single day. And so, how do they actually make sense of that feed? What’s in there doing the behavioral analysis? I’d love to hear about that a little bit, Shanker.

Shanker Ramamurthy: Sure. So, I mean, if we step back and think about the evolution of how analytics are being used, historically, analytics were very simple. There was what we used to call management information systems 10-15 years ago. You took something that had already happened, and then you can […] it into a set of graphs and insights that executives could understand.

Then, they move to, with advanced analytics, to a select capability that they call “predictive.” So, you start to try and predict what’s going to happen. And then, with stochastic probabilistic modeling, we move to prescriptive. And there are a number of areas where computers are now able to prescribe based on vast analytic algorithms what is required to be done.

Cognitive is a next evolution beyond that, where you’re able to look at not just structured data, but unstructured data; voice, video, sound, text, and so on; to then discern insights. And, in the case of the oil industry, determining where to put the rig becomes a big decision. So, what if you’re able to go look at geological data and information going back a decade or two that’s sitting in papers and graphs and have a computer understand it, and then help the geologist determine where to drill the well and how to optimize using all the sensors. But, that’s a perfect use of a draft analytics and cognitive computing coming together.

Michael Krigsman: How does …

Anurag Harsh: […]

Michael Krigsman: How does …

Anurag Harsh: Sorry.

Michael Krigsman: How does all of this impact business models? There is a tremendous capability that's being created by the availability of data, as Anurag was saying earlier; for example, sensors in the oil and gas industry. And, this data enables greater capability in terms of product services, understanding customers; and so, how does all of this, then, feed into business model change?

Shanker Ramamurthy: Well, that actually, and interestingly, it’s dramatically digitizing every industry. I mean, the industry that I grew up working in financial services has historically been inherently digital. But, every industry is becoming digitized because of this power of technology. In fact, Ginni says, “Data is the new natural resource of the 21^st Century,” Ginni Rometty, the Chairman and CEO of IBM, because if you’re able to take these huge amounts of data, and kind of mine it, you work on a model where the data can determine business models; the data can determine the business roles. And, when you have the power of cognitive computing, with smart business models coming together, entire industry ecosystems get disrupted. Entirely new ways of creating value come to the fore. Existing business models are required to be re-factored in the light of this new capability.

And typically what happens in industry […] for industry, but digitization, is the industry becomes much more oligopolistic, where you move to much more of a winner-take-most model, where being third or fourth or fifth might mean having a very, very small part of the value that’s getting unlocked, which is forcing every industry and every enterprise to aggressively and proactively take advantage of these new capabilities, to rethink and refactor their business models.

Anurag Harsh: Yeah. And if you just look at healthcare that we’re talking about, which is somewhere that IBM has really excelled; I mean, you know, in our lifetime; in a human lifetime; one million gigabytes per person in a lifetime of data is going to be captured, right? Or is available. And, the cure for a lot of diseases in terms of genomics and medical imaging and gene sequence and whatnot; all of this data, healthcare, and how we actually approach that; and how do we solve these health problems around the world; is all hidden in that data. It’s that one million gigabytes per person, per lifetime. How do we take all of that, first of all? Capturing that is not easy, but how do we take all of that, make sense of that, and then how do we arm hospitals and doctors and physicians and healthcare practitioners around the world to be able to make that leap and address some of these problems?

See, education and healthcare, they're sort of very similar and I think about this often because cognitive computing and machine learning: a lot of this, at the end of the day, it's about helping society, right? It's about making us better. That's what this is about. And so, those two industries; and it's one of the reasons why I think IBM is so … is like, you know, married into the healthcare space, because of that. Because, in both of these industries, you know, we are optimizing for the law of averages. If you think about a graph, you know, you have a lot of students at either end who are not doing very well, right? They're sort of at the frame.

Same way, you’ve got a lot of doctors. Like, think about the healthcare industry: it’s a seven trillion dollar industry, half of it in the United States; a seven trillion dollar industry; nearly a third of it, almost 40% of it, is waste. This is waste because of bad outcomes, doctors not knowing. Why do they not know? Because they don’t have anything in the past that they can rely on. They don’t know because they saw a person with a kind of cancer that might have been, let’s say, from the Indian subcontinent, and they’ve never seen that before, right?

And so, how do we actually forecast or know is this person going to live the next five years or not? That’s what this is about. And that’s where understanding this data and making sense of it comes. So, the idea is, you know, when Shanker was talking about “man and machine,” it’s making human beings more intelligent via the use of machines. It’s taking those people in the fringes, those outliers on the bar graph, and the bell chart, and making them more intelligent; actually making them as intelligent as humanly possible, right?

Michael Krigsman: So, Shanker, how do we do this? How do we gather that data, and use it in such a way as to make the people who are in business trying to solve these big problems like healthcare?

Shanker Ramamurthy: Sure, sure. So, a perfect example would be cancer. So, staying with the healthcare notion. So, what's in this learning oncology? So, what does it mean when I say, “What’s in this learning oncology?” Well, it’s actually looking at all the published research, you know, several million pages of literature and work that’s been done around the world in oncology; but then what it’s doing is, it’s got the smartest minds in oncology training Watson. So, IBM is working with Memorial Sloan-Kettering, and Memorial Sloan-Kettering’s best physicians have spent tens of thousands of hours training Watson, which already has ingested in under a […] millions of pages of material on oncology. You put those two together, this incredible capability from a cognitive standpoint, with the best that medicine’s got off of from Memorial Sloan-Kettering; now you’ve got a capability that can dramatically augment the capability of a typical physician.

So, that combined man-machine capability, which is now part of Watson, is being used in cancer hospitals around the world: Bumrungrad in Thailand; in Manipal Hospital in India. If you're able to take the average physician and then bring this incredible capability that's a combination of Memorial Sloan-Kettering and Watson, and bring the practice of medicine in the case of oncology to as close to the best possible around the world; I mean, that's a perfect example of the man-machine meld. And, there are a number of such examples in multiple industries where the combination of Watson plus human capability is actually changing the practice of business.

Anurag Harsh: I mean, I have a question, and it's about business models because Michael, he actually brought it up; and it's about platforms and business models. And I was thinking about the whole b2b and b2c space, and in machine learning and cognitive computing, there's been a lot of b2c, like you look at Airbnb, Google, Apple, Amazon, they're b2c companies, right? And this platform has been really, really great in the b2c space. But, other than Salesforce.com and they’re sort of app universe, there are very few examples of effective use as yet in the b2b space.

And so, you know, what’s your point of view, Shanker, on platforms? Not only in terms of building the platform business for yourself, but also, let’s say allowing other industries or other companies, or your clients, let’s say, to use or open up your APIs so they can actually use it?

Shanker Ramamurthy: Sure, sure. So, just staying with this oncology example here: So, you know, when it comes to machine learning and business models, and platforms, the platform we created on oncology is such that every cancer hospital around the world over time we would want [is] leveraging this capability.

It’s not just a black box of capability that’s available to everyone. Watson is now being opened up through a set of APIs. APIs, for those of you who are not kind of deep in technology, is simply “Application Programming Interfaces.” It enables programmers to reach into a system and use the capability simply and elegantly.

So, let me give you a couple of examples of several of the APIs that Watson … that any of us can use. So, one of the APIs that I would encourage each of you to go and experiment with, is something called “personality insight.” So, what Watson is able to do, because it’s being trained by sociologists and psychologists, is it can look at a few thousand words that an individual has written or created, you know, email or text that an individual may have written, and by reading that, discern the personality traits of the individual. Now, that becomes an interesting API, because, you know, if you’re meeting a senior executive who has written stuff before going into the meeting, you can actually try and understand the personality of the individual.

Now, there's another API that converts voice to text. So, if you can put these two APIs together, any voice dialer you might have in a call center construct can be converted to text, then the next API can pick it up, read it, and then discern the personality of the individual, and now you've got very interesting things that you can do to understand the personality of the person that you're interacting with at the end of the telephone to provide the right kind of services.

There are other kinds of APIs, for example, that recognize objects. […] a Watson hackathon, a 24-hour hackathon; and one of the more interesting use-cases was one of the individuals actually used Watson to play Pokémon Go. And what he did was he actually had Watson, through the cell phone, all he needed to do was drive a car, and have your cell phone pointing out the window, and Watson was sitting right there recognizing all these Pokémon images coming up and then creating to everyone saying, “You know, here’s a Pokémon at this particular intersection that you may want to go and capture.”

So, you can now meld the real world with the computer world with image recognition, with all sorts of interesting… So, you’re able to use the power of cognitive computing without having to own all the power of Watson in the way it was four or five years ago, where you had to buy the whole system to take advantage of the capabilities. And that’s where the world is going increasingly.

Michael Krigsman: You know, by breaking it down that way, you have really helped sort of explode the black box because it's not just magic happening, but there's a sequence of steps that are practical steps that, taken together, change the outcomes very dramatically as Anurag was saying earlier.

Shanker Ramamurthy: Absolutely. And, there’s no magic behind cognitive computing. I mean, if you cannot deconstruct what it actually does behind the curtain, as it were, there are a combination of powerful techniques that we have built up over time. So, there’s one technique called “deep learning,” which is nothing more than what we used to call “neural networks” a decade or two ago, where the computer mimics the neural network of a human brain and it’s able to discern patterns. But neural networks cannot tell you, in an if-then-else way, the logic by which it came to a particular conclusion. So it’s perfectly useful in object recognition, in looking at maps and looking at images, and then discerning patterns from that. But, you wouldn’t apply it for credit decision where you need to be able to articulate to a regulator why you said no to a particular credit request, as an example.

But there are other techniques like genetic algorithms, which actually evolve mathematically to figure out the most optimal algorithm in a particular situation. There are traditional Bayesian mathematics techniques you can use that inductive-deductive logic. You put all these capabilities together, you get what I call an emergent property, meaning you start seeing the whole computing capability actually providing insights that start to mimic or mirror real intelligence. And, that's where we are today.

Michael Krigsman: Anurag, so is this, then … Connect this now to digital transformation, because in effect, this is the digital transformation.

Anurag Harsh: Well, it is. I mean, look. Digital transformation, as I keep saying, is still in its infancy. And this is cognitive layered on top of the digital transformation. Digital transformation is about, you know, obviously talked about in the past where it’s not just about your SCO and your web strategy and your customer-facing and all of that, but it’s also about digitizing the operational processes in the back-end, right? Simple things like expense reports and all that thing. The majority of the companies still don’t do it electronically. And now, they’re trying to automate a lot of this.

So, as you think about digital transformation, a lot of this is happening at the same time as cognitive disruption, right? So, what is the key thing with digital? It's the capture of data, right? It's the taking of data, it's the harvesting it, it's taxonomizing it, making sense of it, and then as this data starts to get more, and more, and more, and more in our lives and in industries that we're in; I'll give an example: New York City, you know, the traffic in New York City captures about 500 gigabytes of data every single day, right? That's a lot of data. So, how do you actually take all of that information, as a city, and make sense of it so you can avoid traffic accidents, right? Or you can say, "Okay, well we need to put or we need to repair this traffic signal."

Those are the kinds of things that are key to discerning or deciding how digital transformation is going to take place at a city level as well as a company level, right? So it’s fundamentally about data; taking all of that information, figuring out where we can get that information, which is the data. It’s very easy to just google an app and companies like Amazon, which are in the space itself because they get this firehose of consumer data; but for the rest of us, we've got to figure out how do we get access to that data in order to be able to …

And that's key to digital transformation. Of course. Then, it's a matter of making sense of it and actually doing something with it, right? Digitizing operation processes, as well as from a leadership perspective, to have the buy-in to then make those changes. Because a lot of companies are not doing it, and that's the problem. So I think the cognitive layer on top of the digital layer; it's fundamentally the same thing. When we talk about digital transformation, what we're saying is there are aspects of digitization which involve machine learning, which involve AI, which involves the use of artificial intelligence and neural networks, to be able to … It's the automation of automation, right? It's robotic thinking that allows a system or a computer or a process to learn with each iteration, and then optimize and make itself better, and better, and better, to the point where it can be potentially, arguably, theoretically, at 100% capacity for that particular process.

So, I mean, that’s what I think digital transformation is about. It’s about taking some of these, or archaic and legacy systems, and applying some of these new principles and new thinking, and new machine learning, to those legacy principles, so as to be able to make those processes more intelligent, but at the same time, be cognizant of the fact that the human being is actually running those. So it’s not about taking our jobs, but it’s about having them become more intelligent in applying the intuition and the thinking to these tools to make them more and more better.

Michael Krigsman: So, and …

Shanker Ramamurthy: And digitization is actually changing business models here today. I mean, the classic example is […] becoming bits increasingly. So you're thinking about … Michael, I'll give you an example. My daughter, a year ago, dropped her iPhone and cracked it, and called me really sorry and saying "Daddy I broke my phone," and I said, " You realize, you destroyed a trillion dollar device." And she freaked out! She could not even compute was a trillion dollars was, but I explained to her that when the man landed on the moon, Apollo 11, at that time, the phone that she carries in her pocket would have cost a trillion dollars to manufacture. Not that we could have manufactured it, because two to the power of 30 is a billion, and in the last 45 years, computer capacity has improved, capabilities have improved by a factor of a billion.

When that kind of improvement happens with no end in sight, the model is no longer the same. The model’s different. And digitization is forming in the real world, right? You no longer want to buy a video camera or a camera; you know, everything is all; or books, or music CDs; all of that is now encapsulated in your cell phone, and increasingly, with the cognitive, human intelligence augmentation is encapsulated into digitization as well. And, over time, there are fundamentally new technologies like blockchain that are going to get overlaid on top of digitization and cognitive that has even more profound implications for eliminating the middleman.

In fact, if you think about how society has created belts over the last two thousand years, the rules haven't changed. It's about specialization of labor. It's about putting in place the rules and regulations that enabled us to move beyond being tribal to actually cross a set of rules and a set of intermediaries to actually conduct commerce. What are those intermediaries can actually become software code; and that's really what blockchain's about; what is the amount of friction you can eliminate? What is the amount of reconciliation, confirmation, settlement-style activity that you can eliminate? How does society take advantage of these powerful capabilities? That's really what blockchain's about. So, you take digitization, you apply cognitive, you overlay blockchain; there are more and more and more powerful tools to actually make us all profoundly more productive.

Anurag Harsh: Yeah. I mean, I'm really glad you brought up blockchain, because, you know, I was really thinking about this on my drive this morning, and I was thinking, "How is blockchain" … Well first of all, I think we should explain to the listeners and viewers kind of what blockchain does, and what it is. I'll kick it off in some ways: think about Wikipedia versus Encyclopedia Brittanica. I'm going to sort of take the 300 thousand feet view and then we'll work our way down so people understand when people talk about blockchain, how profound it really is and its effect. So, back in the day, you had Encyclopedia Brittanica, and you had these bunch of guys who would essentially go and research and write up this content, and we, as consumers, will use that content. And now, and eventually, we've gone to the Wikipedia, which is sort of crowdsourced. And now, you can just go to Wikipedia, or wherever, and you can actually get that information. And you trust it because it's a self-correcting algorithm, right? Other people will correct it, and you trust it.

The blockchain is a very similar concept. Think about your ledger. Think about your bank account, right? Basically, money going in, money coming out, but there's only one guy, your bank, or let's say it's PayPal, who's governing that. But, what if you had forty different ledgers, right? Every single bank or every single institution had their own ledger. Money coming in and own, and they’re sort of talking with each other: How do you trust, how does information get transferred and trusted and entered in that sort of distributed ledger setting?

Like, let's say you're sending money from the United States to Mexico, or from the United States to India, right? You don't know for a period of 48 hours or 72 hours where your money has gone. You don't; because there are like twenty banks in the middle who are processing your money and essentially just clearance taking place at every single level. So, if there was a way because there are separate ledgers, if there was a way there was a common ledger in the middle, the sending bank and the receiving bank both share and update it, the money could get transferred instantly, right?

So, that’s where blockchain comes in. Essentially there’s like five or six thousand computers that are competing with each other and they’re solving a mathematical problem, and whoever can solve this mathematical problem efficiently then has owns the bag and says, “Hey, I have solved this problem. Do you guys agree I’ve solved this problem?” “Yes, yes, yes.” “Okay, I can then make the entry into the ledger, alright? Because I’ve solved the problem.” That’s how this kind of works. And, you make the entry into the ledger, and all the other computers in the network, all the other ledgers, they will then make sure that what you’ve entered is accurate. So, it’s a fundamentally different and sort of distributed to technology.

So, my question for you, Shanker, is in a setting like that, which is; think about bitcoin, which is like completely distributed and shared and also accurate and trusted; in a setting like that, how does enterprise blockchain come into play? Because with enterprise blockchain, you have permission-based access, right? And, in a setting which requires permission, and sort of their own domains, and limited to their own domains, how does that come into play? And is it sort of against the whole modus operandi principle of open distributed ledgers?

Shanker Ramamurthy: Yeah, so permissioning is required when you apply blockchain technology in a commercial construct because you can’t operate in a truly anonymous fashion, you know? Bitcoin that sits on top of blockchain is completely anonymous, and there’s no way for regulators to understand flows. So, how do you know drug money is being moved around, and how do you ensure, from a know-your-customer standpoint in a banking context that you are not ending up banking illegal, black money, drug money, and so on? So, that’s a reason why you need a permissioning-based blockchain approach in a business context, where you know who the players are that you’re trusting.

And, there are lots of use-cases for blockchain in a permission business-to-business context. I'll give you some of the ones that we are working with at IBM to eliminate friction in a number of ecosystems. One of those is global trade. We are working with Maersk and a few governments to actually build this blockchain-based ecosystem that enables global trade to happen without all the frictions in between. So, if you think about the movement of containers and goods from one part of the world to another, the number of parties involved are extraordinarily large. Clearly, you’ve got the truckers, the shippers, you’ve got the porch, you’ve got the buyer, you’ve got the seller, you’ve got the bank that lends the money, the bank that’s going to collect the money, you’ve got insurance; you’ve got multiple ports and multiple tax regimes to deal with.

Now, a blockchain-based permissioned approach that has a combination of these actors, let’s say the Singapore government; let’s say Maersk, let’s say IBM; let’s the US tax authorities; let’s say a couple of financial institutions; that are involved in trade finance, come together and agree that they are going to have to share hyperledger, then you’re able to eliminated 60, 70, 80, 90% of the friction in all that commerce happening, and the societal impact of just that alone is in the tens, hundreds of billions of dollars.

Imagine, for example, and this is a wonderful use-case, in many emerging markets; in fact, even in the US; where, when you’ve got properties being bought and sold, you do not have a trusted intermediary in the middle, which is why, here in the US, you have insurance in the middle for title … you have title insurers, who underwrite that, you know … If you are selling me the property, then you’re the owner. In many parts of the world, you cannot even trust that intermediary who is a government official or a government authority because they might be corrupt.  A blockchain-based technology ensures that if there is a dictator or regime change, you can still authenticate ownership of property while dramatically reducing the cost of that.

In countries like India, the lack of trust in property transactions is a 1.5% tax on the GDP on an annual basis. Imagine eliminating that kind of friction on the back of blockchain?

Michael Krigsman: This is a fascinating conversation. We only have less than five minutes left, but we have a really interesting question from Twitter. And, on Twitter, Sal Rasa asks, "During IBM's re-engineering, there was a list of customer imperatives set forth. And so, how do customer imperatives relate to today's conversation? Where does the customer fit into all of this?

Shanker Ramamurthy: This is a wonderful question. And by the way, everything we do in IBM is driven by the customer. So, because we are a b2b company, we do not do things without a customer involved. So, every one of the examples that I’m talking about is something that we actually do with a customer who, in some instances, might be a partner. So, if you are setting, in this instance, if we are setting up a blockchain-based ecosystem, Maersk might not just be a customer, but Merck and IBM together might actually collaborate to create an ecosystem in which we share the rewards.

So, in fact, in IBM, we pride ourselves on solving some of the most complex business problems that our customers have. Think about them as clients rather than customers, because we actually tend to work with the Fortune-1000 around the world, from which we derive most of our revenues, and these are long-term relationships that […] in individual transactions that we have with them; and it's about solving their most complex problems, and increasingly, technology is able to be applied to the point that Anurag was on in a digitization and a business context in technologies like blockchain and cognitive, are actually infusing themselves, integrating themselves, into the strategy of our customers. So, it's not about … You determine the strategy and then you apply technology to solve this strategy, it's a fusion where strategy happens real-time by understanding technology capabilities and business imperatives and bringing them together. And, that is the intersection in which we are operating.

Anurag Harsh: You know, it’s interesting you talk about customers, and I was thinking about the whole “freemium” model of Google, for example, or Amazon. And think about Google Translate for a moment. We have a lot of clients here at Ziff, and a lot of technology companies; a lot of legal departments are starting to use Google Translate, and this is something we were talking about earlier, and their models are a freemium model, “We’re going to give it to you free, but if you want to use it on a super-user basis, then you’re going to have to pay for it a little bit,” right? Sort of like Gmail.

How does IBM think about that model, in terms of customer pricing in the b2b space, as well as if you’re thinking about going into the b2c space?

Shanker Ramamurthy: Hey, that’s a great question. So, the very simple answer I gave you is, IBM's model is based on clients owning their data while the b2c models that are typically based on the enterprise that provides a service, providing it free, but owning the data that they can mine. So, it's fundamentally different models, each has its merits. In a b2b context, we pride ourselves on having our clients own their data and the insights we might derive by leveraging cognitive or any of the other capabilities we bring to bear. Each has its role, but we have specialized in solving the most complex problems and ensuring that any problem we solve is owned by the customer rather than by anybody else.

Michael Krigsman: Okay, so customer at the center all the time. We have about a minute left, and I’ll ask each of the two of you to give a tweet-sized summary of your distilled wisdom on these issues. And Anurag, why don’t we start with you? So, a tweet-sized summary of your distilled wisdom of everything you know about this stuff.

Anurag Harsh: Everything I know about this stuff in one tweet-sized summary. Look, digital transformation, cognitive transformation is upon us, it's happening, autonomous cars are going to be part of our lives, it's going to change business models, transportation is eventually going to become a data-based industry, a services industry. A lot of product companies are moving into the services sector, and vice-versa. So, I think, embrace it; love data; capture it; make sense of it; make data your friend, and inform your business decisions based on the intelligence you receive from that data.

Michael Krigsman: I love that.

Shanker Ramamurthy: That was very, very succinct. I would say step back, recognize that these technologies are here, and they are going to change your business. And so, just ask yourself if you’re a business executive, “Is my strategy ambition enough, given what’s happening?” You know, am I able to learn fast enough, and am I on that virtuous cycle? And, do I have the people process technology capability to actually implement all these capabilities to take advantage because if you do not do that, your competitor’s going to eat your lunch.

Michael Krigsman: Okay. I love it! Wow, I wish we had a lot more time.

You have been watching Episode #225 of CxOTalk. And what an amazing discussion we just had on blockchain and cognitive computing, and digital transformation! We've been speaking with Anurag Harsh, who is a senior exec at Ziff Davis and has written multiple books on this topic. And, we have been speaking with Shanker Ramamurthy, who is a senior exec at IBM, and who spends his life immersed in these issues, with very, very large corporate clients.

Gentlemen, thank you so much for being here, and I hope you’ll come back and do it again another time!

Shanker Ramamurthy: Pleasure.

Anurag Harsh: Thank you.

Michael Krigsman: I'm Michael Krigsman and thank you for watching. Go to CxOTalk.com/episodes to see what's coming up. And, subscribe to us on YouTube. Thanks so much, everybody. Bye-bye!