Hello there. This is Smart Talks with IBM, a podcast from Pushkin Industries. I Heart Media and IBM about what it means to look at today's most challenging problems in a new way. I'm Malcolm Glapo. Today I'm talking with not one, but two bright folk from two very important companies. We have Sweeny Calla Paula, Vice President of Global Technology Strategy and Network Cloud at Verizon, as well as Steve Kannapa, Global General Manager and Managing Director of IBM's Communications Sector.
Sweeney Calla Paula leads Verizon's technology strategy and works closely with partners like IBM on emerging connectivity technologies. He's passionate about a world where people, machines and systems will one day interact seamlessly. I like me. Sweeney is extremely excited about the technological progress we've made during the pandemic. The digital transformations took off in the last one year because of the core ridden because we were all kept apart
and we're still needed to communicate. We cill need to get things done. Steve Kanappa heads IBMS Communications Sector, where he works with the telecommunications and media firms all over the globe to help them modernize their networks. Steve is extremely excited about the future of AI and machine learning, artificial intelligence, the ability to put machine learning capabilities where processes get smarter continuously and more they execute, the smarter
they get. Right after the break my conversation with Sweeney and Steve, the world of technological collaboration has never been more fascinating. What the two of you ascribe the nature of the challenge that the What is it you're trying to address in the work that you do. So in
a word, um, what we're focused on is innovation. So with at IBM, we work with clients in all industries retail, manufacturing, banking, media and entertainment, TA, our communications, government agencies, and we're helping them provide better services to our customers or their constituents by helping them modernize the way that they bring technology and applications to all of us. And we work in collaboration with the TA, our communications companies and helping
them bring those new those new capabilities. So should he tell me start by describing the nature of your partnership with IBM. What is it that you when you work with IBM, What are they bringing to the table. One of the things we do is that drive reliable connectivity. And now the need of reliable connectivity if you look back to it and and twenty years ago, the needs were different than the needs that are there today to
where we're going to be integers from now. So you know, ten twenty years ago, you know, hey, I want to move around, but I want to talk to people. Right then that evolved into messaging, That evolved into you know four G world where not only you are moving around, but you're able to do things with your phone and others.
But as we look forward, and especially if you notice what happened in the last one year, right the digital transformation just took off in the last one year because of the COVID and because we were all kept apart and we still needed to communicate. We still need to get things done. Now, the nature of connect really goes
beyond humans. It now goes into we need to deliver connectivity, highly reliable connect rety where machines can communicate with each other and machines can actually do things, you know, simple things like remote health care. Right. So where IBM kind of fits on how we collaborate is that we see that the future is about machine mobility. So where our networks now have to deliver connectivity two things that move around, things that are connected, which require a lot more relevel
connectivity than humans. Humans we can adapt to errors and changes at a at a you know, second level and all that. But machines they are designed to be perfect, and that means they require high level connectivity. We variazon. We build highly reliable, high performance networks. And when you're trying to develover these capabilities to let's say, automate industries or automate you know, transportation sector and others, IBM understands
those sector as well. They have been digitizing those sectors and and they understand that the main that particular domain and what sort of solutions can actually uh you know, can be incorporated. So when you bring these two expertise together, now we're able to deliver this automation, this decidation in more effective way. That's how, you know, that's how we
both you know, kind of collaborate and work together. When we talk about how communication between machines has to have a higher standard than communication between humans, what does that mean? Are you just talking about reliability? Are you talking about the size of the pipe, Are you talking about the scale of things you want to do. So I'm going to take an example of a let's say a car that is connected right and it's a call it a
semi autonomous car. This car is traveling at hund miles per hour and it's capturing, uh, you know, lots of information and it may have to understand what are the other things that are in that area. When you're traveling at hundred miles per hour um, the information you collect and the decisions that you need are going to be in the order of milliseconds. Because if you have to break at a you know, at a speed, if you take a second, you're already you know, causing an accident.
You're now talking in the ranges of ten early seconds, ranges of area. So that's one thing we talk about result you know, highly latency sensitive networking, whether the humans let's say you're you're browsing something that are trying to pull a website, it takes a couple of seconds longer you don't actually sense that field that maybe five stand seconds you feel that. Whereas when you connect in any kind of automated thing, you know, dron't trying to deliver
a robotic vehicle within a factory environment. These things are moving at a pace where you can perceive a millisecond delay, and and so the networks now have to operate at that level is a much higher order of you know, you know latencies. Now, the machines that we're trying to automate, they're collecting all of this data, lots of data. We're talking about hundreds of magnitudes higher than what you used
to get collecting. Now, if you try to send all the data to a far away cloud, that means that you require massive amount of networks all the way, which which don't exist today. That's where edge computer comes from. The picture where you collect all of this data, you hand it over to an age cloud, but that's very close to the user. Let's say you know a few miles from the user. Then you process the data rumor a lot of data that you collect is more of
what we call it's not information, it is data. Right, You take the data, you process it locally, you get information out of it, and then you use that to both coming and get back to the object you're trying to control as well as you know, send it to where or you need to send it to. So in These are the kind of you know, key ingredients that you require as you look start looking into the future. Last week, I was in Phoenix and I took a ride on one of those autonomous vehicles. I ordered the
taxi on my app. It showed up. You know, no no human being to be found. The digital backbone, the
communication systems. This is based on better be good because if you know, if we drop coverage going at forty miles down the road trying to navigate a you know, then I'm in trouble, right, there's no backup here exactly, And if you think about what's actually were happening in those scenarios, it's it's about experience, so changing the experience, it's about personalization, and oftentimes it's about delivering the insights in real time about how you how how a process
is happening. I'll bring it back to the manufacturing shop floor as an example to kind of tie together a couple of the points that Strening was making. Now, we've been instrumenting with IoT devices and manufacturing shop floors for quite a while hundreds if not thousands, but now think about taking that to tens of thousands of sensors on everything that's happening in real time in that manufacturing floor.
Think about how videos changed all of our lives over the last few years, as it's become ubiquitous in the way we work and the way we you know, get entertainment, news, information, the way we communicate with each other. Or video is
essentially just rich data and it and and so. Now instead of just having swers on that shop floor, we could have video cameras watching everything is happening, watching workers to make sure they're in safe zones, watching whatever is being produced coming down the factory line, understanding if it's being done exactly the way it's supposed to be done, Watching the machinery itself and the way it's performing to
notice the slightest changes. And the second thing that we're applying in is artificial intelligence, the ability to put machine learning capabilities where processes get smarter continuously and more they execute,
the smarter they get. So now when you combine those two forces together, the ability to use high fidelity data like video, and the ability to interrogate and analyze that data in real time and do it right where things are occurring, like on the shop floor, and then have that backed by the kind of connectivity and the power that the Verizon can bring with their their edge computing platforms.
We have the opportunity now to add tremendous value to those processes, whether that's making sure people are safe, making sure that those lines are as productive and as efficient as possible, making sure the product quality is as high as it should be. A tremendous uh, you know, amount of value can be created by being able to apply that that intelligence, um that that high bandwidth capability, and to make sure that that network is there and ready,
as Shrine was describing, to constantly serve up those insights. Yeah, imagine that. I'm a uh, I have a you know, mid sized manufacturing company, you know, founded by my grandfather. You know, I've been keeping reasonably keeping pace with innovation. I think I'm pretty competitive. I have not done any of the things you've described. The two of you show up at my front door, and what kind of promise can you make me? What like, for example, what when
you talk about productivity improvements are we talking about? What are we talking about your two five percent, ten percent? Give me kind of more tangible examples of what you can deliver to a customer like that. So the way industries have been evolving is that they have a lot of these sensors, and the sensors have been more analog and and and what I mean is that more each one is designed to do a particular thing. I can tell you that like temperature sense of that basically says
if you cross certain temperature, you slve the machine. You have let's say some windspeed sens or something else, something else. What industries are realizing is that these sensors have been doing a certain function by now I'm going to use the term called virtualizing, by basically making them more lean and smarter and connect to the edge. You can now get all this information and you can actually make decisions based on collective intelligence intelligence of all the sensors than
each sensor doing what other things. And to do that you need a highly releveled character. By bringing the entire collective intelligence and and and in making these sensors more virtual. A couple of things you're doing. One, you're going to bring in latest capabilities in UM called product quality and manufacturing and others, meaning that you would improve the radios downtimes, improve you know, the amount of productivity, improve quality control issues.
And you're actually going to enable the factory to continue to adapt to the new or digital capabilities that are going to come out. And each one of these capabilities, where as AI and whatever else, they're all going to add to your productivity, your quality, or you know, the key metrics that you're you're looking at. So for a hypothetical imagine which is saying, is I suppose I have
a machine, a very expensive piece of machinery. My my my cousin is the president of a company in Queens that makes Jamaican beef paddies and they get these machines from Italy. The cost just to say, in amounts of money that you know, take the beat patty from what and when the when something goes wrong, it's like a crisis because someone has to fly in from Italy. Right, So listening to I'm thinking, well, suppose we discovered that the machine starts to make lots of airs when humidity
is hits a certain point in the company. It's kind in the factory floor, the temperature is such when it's been running for X number of hours in a row. When the operator makes this kind of mistake. I mean, I can imagine ten different things that we could say might affect performance, and you could have sensors on all those ten things and do a calculation at any given time about what my chances are that the machine might have a malfunction and presumably warned me before it happens.
Is that exactly exactly? And what you've what you've just done, is you've crossed over from observation into prediction, which becomes really powerful so that you can begin to understand what's about to happen actually before it occurs, and you can take corrective actions to do it. As a perfect example of the kind of tremendous value that can be gained here. Adding to my little scenario from before, we could have sensors that tell me when my raw materials are getting low.
We could have an AI system which reminds me when variations in demand and says you actually, every year have a huge uptake in at easter, you're you know orders that means you've got to order X, Y and Z, right. I mean, this is all the kinds of things you might want to do. How hard if you walked in how hard? What is it to build that kind of system? Is this something you can do relatively quickly and efficiently? Or is it does it take months to kind of
customize a system for a customer like that. So for for you know, any different clients, some things can be basically prepackaged and ready to deploy UM. And you know, as as these solutions of all UM, you know, there will be some offerings that are essentially load and go as a service. Others will be more customized or tailored
based on what given client will want to do. But one of the things that I think really important in what we're working on and what Verizon is working on, is there are a set of standards that the marketplace is embracing. And those standards allow for an ecosystem of players to come together and to work seamlessly. And and for IBM, you may have heard, you know, the term
our open hybrid cloud approach. We're very very focused working in collaboration with the telecommunication providers in the marketplace to help bring these open standards because two really powerful things That one is we create an enormous pool of innovators that can contribute and accelerate the rate piece of innovation. So we're working very closely with Verizon on on some of those open standard architectures. And then Secondly, I'm really
creating an ecosystem of partners. So when we announced just at towards the end of last year, are IBM Cloud for Telco, we announced with UM over forty different partners that are coming together to be able to bring innovation UM to these kinds of solutions. And I think that's that's going to help us UM really accelerate the opportunity
to bring these kinds of solutions. Were describing treaty. Is there somebody out there, an industry, a sector that you think could benefit the most from the kind of services that I'd be having Verizon provide. I mean, who where could you make an extraordinary difference? Retail? Retail is a very interesting area because both with COVID you started getting into this ideas of touch free detail and you do things more customized and you users and others. But retail
itself is actually changing. UM. You know, you hear about these UH stores without any assistance that you're walking, you buy and walk out. In all of these cases, what you're seeing is automation really taking a bigger place. For example, you know one of the retailers in a decent sized store is looking to put about high definition cameras. Now the cameras work as call it computer vision sensors, and they capture the information and using that you can actually
understand what you know, individuals are. You can do lots of things with the computers and of what people are doing, how much stock you have in the store, and lots of other things. Now you take all that, you process that and uh and you could you know, if you introduce automation at the point where people can come and pick up what they want, they can walk out without having to even interact with anybody out there. Right, that's one extreme of it. I just imagine that I'm I'm
running a store, did a bunch of cameras. I got fifty stores across America. I've just put in my spring line, and I got a bunch of new dresses. So you could tell me that, Malcolm, you have this new orange dress which you put in all fifty stores. Not a single person has looked at that dress in the retail store in the last week. You can also tell me seventy of the people stopped in the thing lingered in front of the teal sweatshirt, and you're out of teal
sweatshirts as a result. Order more teal sweatshirts now and forget about the dress. It's not going anywhere. And not only not only that, I'm not a retail expert. By there,
I can tell you that. Not only that, we can also tell you that people who looked at something for how longer time, two seconds, two minutes, whatever, tend to buy things versus if they don't look at it, that means, you know, if they're just walking off, that means you know that you're not gonna So you can get such a deeper set of insights that both can help you as a business to figure out how do you, you know,
how do you interact with your customers. But at the same time, you can also use this this data to actually uh you know call it push your prior to the customer and you know, based on their personal preferences, choices and color circumstances. Right. So that's the kind of the data gives you so much and so much what you call ability to kind of you know, customize yourself
to to meet the customer demands. You know, in a way, what we're talking about here is you know, a using video as we talked about as a rich fidelity set of data but also applying AI or intelligence to that so that you can take action on it. But part of this is about humans and machines interacting, which makes the human in doing that role more effective. Another area where we both are excited about is industrial automation. So if you look at you know, malcome, what happened over
the last few years is the world has digitized. You know, everything can be digital. The back office is the manufacturing process and others they have not being disguised. So you know, in a number of situations, you put this nice digital store front and digital experience in the front, but your backup is still lagging with you know, older factories and
older sensors and whatnot. And the problem is that at some point the problem is going to catch up, and and that is you will not be able to keep your front digital experiences without really you know, knowing what is going on your factories or do you have the product? You can you shave the product on time? And if you come in to somebody, I'm going to deliver something
away tomorrow, can you really get it there? So the next ten years and there's gonna be a lot of industrial automation that's going to take place now for industrial automation, UM, a few things that are going to make a bigger, bigger impact. One, you do need a highly reliable connectivity within those environments. Whatever small number of sensors they have today, they're connected using wires. But the problem the virus is that it takes a long time to really go connectivity.
You need a highly reliable wireless communication. Number Two, we're gonna have inordinate amount of sensors all over these environments because to your point about your cousin's uh Jamaican beef parties by it does sound very attractive. So they're good, They're really good. So so now you're gonna put sensors everywhere so that you can predict, you can understand exactly
what's going on. The third thing you're gonna do is that use all of the data that's getting generated and apply AI and machine learning tactics on that so that you can not only understand the the you know, the world is going on within the environment, but you can actually predict. You can forecast now in a much better way so that if your customers expecting X y Z on certain time, you have data to tell you that you can absolutely come into the X y Z to
be delivered to the customer on time. So it's that combination of sensors, programming, and intelligence and the AI and others. IBM understands a lot from that environment to the reliable connectory that we bring and and the edge computer that that we put to other those elements that you've just mentioned, Um, there they can't all be at the morning. If they're they're not all the same level of sophistication or development. Or is there one that you worry the most about?
Like for example, when you were talking, I was thinking, you know, the Internet in my house it's not Horizon, but the Internet in my house it's not that good. I mean, I have to reboot my router every two days, and I get a little I have to have massive backups because sometimes it's just cons I you know what I mean. Like when I hear this, I think this is really really wonderful, But the reality in the little town I live in is Interne's just not that good.
So if I was a company in botttle town, how would I do what you're doing? If my if my fundamental services so spotty. Yeah, by the first I wish we were your internet providers. You wouldn't be having those issues. That's my uh So, I think the infrastructure, the broader connectivity infrastructure UM will need to evolve. Again, I go back to COVID because then COVID truly pointed it out that you can operate at a normal pace if you have a good connectivity, but if you don't have a
good connectivity, you will get left behind. We certainly continue to deploy more and more connectivity UM. Where I see it's not a worry, but it's more of an optimism. Is that is five G is going to change that. Now five G brings in higher amount of throughputs in a wireless manner, so that you know, wired is a difficult proposition. Digging streets is not an easy thing, and then connectivity cost a lot if you go through that way.
But then on the other hand, a wireless high band with connectivity, reliable connectivity makes it easier because it's something we can deploy in a matter of hours within the factory. And what I actually worry about is the technology is the ingredients are evolving at a rapid pace. It is the people who operate those environments and it is the you know, the pace at which they adopt is what
has got to catch up. Um, what are you going to start seeing is that the tech is evolving so rapidly that the expectations um from the consumers in terms of you know, hey, when you guarantee something I wanted to be guaranteed because you should have had that. You know, you won't promise me something if you can't keep up, those sort of expectations are going to continue to increase.
And if the the the people who are actually managing and operating the factories and and working those factories cannot adopt and cannot evolve fast enough, you know, that's what we're worried about, is that the adoption and the evolution of those environments. I mean that's a question for both of you along those lines. You know, are are the two of you, your two companies big enough to handle
this challenge? And I say that not I'm not trying to provoke you, but you're talking about something where potentially every business in this country could you know, benefit extraordinarily from this revolution. We haven't even mentioned healthcare GDP. You know, we're talking about hundreds of thousands, millions of people and who are employed in that industry. Many of the industry standards in healthcare are straight out of the nineteenth century.
You could throw an army at that problem and you wouldn't put a dent in it. I mean, do both your companies have to scale up to meet the challenge of delivery of making this revolution reel. Well. For IBM, we've been working with each of these industry sectors for for many, many years, and so we have a pretty good understanding and a longstanding relationship with with many of the leaders in each of those different industries, and so we're working day by day with them as just trying
to adopt these kinds of innovations into their business. So as much as I think Trini would say and it as what I did, that we both have ambitions to to really bring a lot of value to all of these inns because we know there will be many other firms that will be doing the same. And what we want to be able to do is to create UM is open, efficient, and is automated a technology environment as possible that allows for that innovation to happen. M hmmm, yeah,
and Steve's summarized very well. I mean, Malcolm, the way to look at this is tech is evolving, so much, and on multiple fronts. Right, we talked toward sense as we talked about we talked about the connectivity, We talked about cloud and software and computing. There's no no one single company that can claim and say that they have mastered all the domains. That's why this is a you know, collaboration and kind of standards base interoperably across companies is critical.
The way I look at OS is we provided these what we call the code ingredients, connectivities of critical code ingredient, compute clouder code ingredients. We provide them these basic ingredients so that others can come on, innovate on top, and you know, delover these uh, these these kind of emerging the new services to the to the enterprises and costomers. Yeah, there's a couple of terms that I wanted to get more complete definitions of, and I wanted to start with
edge computing. I know, Steve, you had talked, you had given us a an initial definition, but imagine that I am a complete computer literate. Let's let's really dig into that term and what we mean when we use that term, and how that kind of differs from what might be a more conventional um computing model. Yeah, so let me use an example that maybe it could help. UM. You kind of draw a picture of what that might be. UM.
I Unfortunately, I'm based in Los Angeles. As you know, most of your listeners would know, California has been battling with fires, UM. You know, in the fall almost every year. Now, this last year very devastating. UM. In an edge connected world that Triny and I have been describing, a fire begins in northern California on Wednesday. Everything is fine on Thursday. On Thursday, we've got an unfolding disaster in you know,
some part of the state. An edge platform in proximity to where that fire is comes to life as a platform, and a set of applications come to life on that edge platform, and immediately drones are flying above that fire and taking video imaging and sending it down to that edge platform, and a bunch of AI tools are analyzing that video to analyze whereas the fire, what's the terrain look like, what are the things that are in the half of where that fire is likely to go, so
that we can begin to plan out how how to fight it. And then data coming perhaps from our weather company is going into those models on that edge platform, and it's infusing intelligence about wind patterns and moisture in the air and things that also will impact the way that that fire is likely to perform. And that data is being analyzed and then fed to the first responders.
And when the first responders show up on the scene, their their trucks, their cars, their equipment is placed in an optimal position based on the analysis has been done on how that fire is likely to perform and where in fact, you know, property or people are at risk because of it. And sensors we were talking earlier about sensors. Sensors are in the area that are measuring the amount of moisture in the foliage. That also is going to impact the way that the fight against that fire is
going to be done. The point is a lot of data coming from a number of different sources, from sensors, from video being analyzed in an environment it's very close to where that fires occurring, so you can get that real time response and then providing information. But the ultimate objective to save lives, save property, get that fire out as soon as possible, and when it's out, that edge platform can essentially wind down and and um you know those applications can then be kept and made ready for
the next fire wherever it may happen. That's the kind of automation and intelligence and real tie insights that could fundamentally be brought to bear in an edge environment. I think that hopefully that paints a little bit of a picture of how we see this edge environment creating tremendous value. Yeah, wonderful. Well, thank you so much to both the Ustraenian Steve. UM.
There's been Um, it's been really fascinating. I'm so grateful for the work that companies like Verizon and IBM are doing to help fight wildfires, and to Scrini and Steve for taking the time to chat with me about all that and more. And the sensors. We can't forget the importance of sensors. Sensors that infuse new intelligence. Whether they're monitoring where a fire is going or making sure equipment in retail factories is functioning safely. Sensors are crucial to
protecting people in so many different capacities. It's pretty amazing stuff. Smart Talks with IBM is produced by Emily Rosteck with Carl Migliori, edited by Karen Shakerji. Engineering by Martin Gonzalez, mixed and mastered by Jason Gambrel. Music by Gramoscope. Special thanks to Molly Sosha, Andy Kelly, Mia Label, Jacob Iceberg, Heather Fane, Eric Sandler, Maggie Taylor and everyone at eight Bar and IBM. Smart Talks with IBM is a production
of Pushkin Industries and iHeartMedia. You can find more Pushkin podcasts on the iHeart Radio app, Apple Podcasts, or wherever you like to listen. I'm Malcolm Gladwell. See you next time.