Innovation and Technology at Xerox PARC, with Stephen Hoover, CEO
The Xerox Palo Alto Research Center is among the most venerable institutions in Silicon Valley.
Watch on YouTube
47:40 
The Xerox Palo Alto Research Center is among the most venerable institutions in Silicon Valley. Founded in 1970, PARC has led breakthroughs in areas as diverse as laser printing and user experience ergonomics. Wikipedia describes these accomplishments:
Xerox PARC has been the inventor and incubator of many elements of modern computing in the contemporary office work place:
- Laser printers
- Computer-generated bitmap graphics
- The graphical user interface, featuring windows and icons, operated with a mouse
- The WYSIWYG text editor
- Interpress, a resolution-independent graphical page-description language and the precursor to PostScript
- Ethernet as a local-area computer network
- Fully formed object-oriented programming in the Smalltalk programming language and integrated development environment.
- Model–view–controller software architecture
Our guest on this special episode is Stephen Hoover, CEO of PARC, a Xerox company. Hoover joined PARC in 2011 and oversees PARC’s work for clients in diverse focus areas and competencies including networking, novel electronics, human-centered innovation services, cleantech, intelligent systems, contextual intelligence, and more.
As Vice President of the Xerox Research Center of Webster (New York), Hoover supported core and next-generation research and development. He later led the global organization responsible for the company's software and electronics development. In these roles, Hoover was responsible for multi-million dollar R&D investments and product strategy encompassing several platforms and market offerings. He has led long-term technology investments in grid and cloud computing, nanotech, mobile, the Future of Work, and advanced printing and mass customization technologies.
Dr. Hoover earned his Ph.D. and M.S. in Mechanical Engineering from Carnegie Mellon University, B.S. from Cornell University, and 1 of 10 national fellowships at AT&T Bell Labs. He has 7 patents. Hoover has served on the Board of Directors for the Rochester Museum and Science Center, including leading its K-12 STEM Education Task Force; and is a regional Board Member of FIRST Robotics, an organization which inspires young science, technology, and engineering leaders through mentor-based programs.
LinkedIn: https://www.linkedin.com/in/steve-hoover-04376b3
Twitter: https://twitter.com/HooverSteve
Michael Krigsman:
(00:02) Welcome to episode number 146 of CXOTalk, I’m Michael Krigsman and today I’m so excited because I’m speaking with Stephen Hoover, who is the CEO of the Xerox Palo Alto Research Center. Xerox PARC is one of the most venerable research organizations in the world, and I’ll let Stephen describe all the incredible (Lost video 00:32 – 00:35) that were originally developed at PARC and so Stephen welcome to CXOTalk.
Stephen Hoover:
(00:42) Thank you Michael it’s good to be here.
Michael Krigsman:
(00:45) So Stephen just to begin very briefly tell us about your background. So you’re the CEO of Xerox PARC but how did you get there but very briefly.
Stephen Hoover:
(00:55) Sure, I have my PH.D from Carnegie Mellon and I’ve always been very interested in technology, from a technology side, the kind of the confluence of the physical world and computation. So my research area, technical research background is in applications of artificial intelligence to design tools, to kind of advanced methods of computer-aided engineering and then robotics and mechatronics.
(01:29) And when I finished my PH.D, I joined Xerox quite a while ago, actually 21 years ago and have been in a variety of innovation positions since then. Moving from the research side and into product development and delivery, and back into the research side in some technology strategy role. So it’s been a really interesting opportunity to kind of live across the whole spectrum of innovation from you know, crazy early ideas that you don’t know to work, to supporting you know 10 million customers with a new product release.
Michael Krigsman:
(02:04) So you’re a scientist who’s now managing scientists and researchers and I’m assuming some engineer people, different kinds of backgrounds. Tell us about Xerox PARC.
Stephen Hoover:
(02:18) So it’s a really interesting place and we have as you said earlier, a really storied history of founded in 1970 by Xerox, specifically chartered to create the office of the future. And we did some of the foundational work in personal computing, the creation of Ethernet, which we actually spun out here and in 3Com. To the creation of the graphical user interphase and the integration of the mouse into that to what you see is what you get editing, to the invention of laser printing and the commercialization of that.
(03:01) And have continued in that vein developing innovations for Xerox’s businesses today which is much broader than a printing business but is in fact half of Xerox’s is in the services business.
(03:16) We do transportation services for cities and other local governments, ranging from things like the EasyPass toll systems, and some of your users may be use EasyPass to fare collection systems. We work for healthcare companies in analytics and information management. We do customer care, we answer $2.5 million customer interactions every day that Xerox does.
(03:50) So at PARC, we are set up as a whole subsidiary to support Xerox by developing early-stage technologies and innovation to drive into their business, but what’s interesting we’re set up as a subsidiary also because we’re a P&L, in the sense that we’re in the open innovation business, so we do that not only for Xerox, but for other clients and customers in this world. We stepped into this model in the early 2000s, focused on open innovation as a business practice, actually before Henry Chesbrough coined the term open innovation in 2003, and it makes this a really interesting place, as well as the you know the business model and that focus on not only on some long term innovation and research but also how do I drive the current business results from that, putting those two together and then as I said doing it in this open innovation business model.
Michael Krigsman:
(04:53) So what’s the relationship that you have with Xerox, it’s really interesting that you’re an R&D organization that also has a P&L. so as you talk about the relationship with Xerox what does that P&L responsibility imply for you?
Stephen Hoover:
(05:10) So we have two fundamental roles for the company, one again is to do early-stage innovation and research for Xerox for itself, so we had a relationship that then looks like what any, not any but innovation for organizations and large companies that do that well, which means you are you know closely working with the business units on their strategies, you’re developing technologies and new offerings that extend their current business and take them into you know, radically new places. You’re working with corporate strategy, and again to ensure that you are both being influenced by the strategy but also influencing the strategy about where it needs to go, what technology to enable, and business opportunities come out of that
(06:03) And in addition we have this role in open innovation where we again, many of the early-stage technologies that we work on are very relevant to people outside of Xerox. So how do we go and find those customers and clients to help them in their business, and then even find relationships where we may do an early-stage innovation work that we start to work with external companies, and then we bring that company in as a partner for Xerox to extend its offerings. So this open innovation is a very interesting model.
Michael Krigsman:
(06:42) So is the reason for having a P&L is so you can self-fund the organization or are there other reasons beyond that.
Stephen Hoover:
(06:51) One of the reasons is absolutely to be able to look at a broader reach. One of the roles is if you are a good innovation organization or a parent company, then what you are doing is developing innovations were they're occurring in businesses, but also innovating in new businesses that they are not in today, to give them some options to expand beyond their current markets.
(07:27) Now, it may turn out that some of those things aren't the right things for the parent company to take up for a variety of reasons, and then what do you do with those. They may be great ideas but the parent company isn’t the right place to do that. Or it may not be clear in early-stage technology or what the real opportunity is, so how do you go and explore that? When you go explore it by creating those opportunities yourself as an innovation center and finding other companies to do it. So it isn’t one aspect is yes, it allows us to have a broader set of innovation exploration areas, but doing it in this way it forces you to do it kind of in the crucible of a market still, even if it isn’t clear that the parent companies the right company to take that up, or in what form.
(08:17) So it kind of keeps you honest in some sense in that regard because you are forced to go out and compete you know in that way and survive. So yeah, part of it is in the funding to be in broader areas, but part of it is you know, it’s a great crucible to test out new ideas. And you know, in my experience and a long history in innovation there are three major reasons why significant innovations fail.
(08:49) One of them is the technology isn’t right, it’s not ready and it doesn’t work. But another is that you have understood the market wrong. You’re not actually solving the problem you know, in a way that customers will adopt and the third is that you don’t have the right business model. That while you are creating value in the world for those customers, you’re not able to appropriate value along across the ecosystem to the right people in the right ways, so that the whole value chain makes money right. In the end, businesses are going to need to make money to deliver a product success to a customer over the long term. And so, if you’re not exploring those other two aspects; the market and the business model, then I would argue that you’re actually not doing a good job of innovation. So this gives us a much broader way to explore those areas.
Michael Krigsman:
(09:47) So as you say into the crucible of market relevance. Steve we’ve lost your video, so make sure it’s plugged in, if not we keep going because we have your audio, but it would be nice to see you as we saw you earlier.
Stephen Hoover:
(10:06) Google Hangouts says I’m on and I see myself, so sorry I’m not sure what the problem is.
Michael Krigsman:
(10:12) Okay there’s nothing to be done about that. So tell us about research at Xerox, the basic research versus applied or product research.
Stephen Hoover:
(10:36) If your listeners are interested and kind of want to understand this business of applied research versus basic research, and how to think about that in a model, I can refer them, there is an interesting book and the book is entitled Pasteur’s Quadrant, and it’s trying to capture you know in MBA school we learned how to describe the world in a series of 2 x 2 matrices, right. And this 2 x 2 matrices, in this case is articulating one axis is about knowledge for knowledge’s sake or is it about solving a problem.
(11:24) And the other axis is basically about the time horizon of which you are operating, and Pasteur’s Quadrant is this upper right quadrant where the idea is it’s not so much about knowledge for knowledge’s sake, but it is about generating new knowledge and new science because there is the focus area in which you know you need new knowledge and new science to understand the world, and then to create things in the world which deliver the results that you need, and that’s the idea. You know, Louis Pasteur, he was very focused on not just new knowledge, but okay, how do microorganisms work, right. What is going on in this space, and knew that knowledge was needed to then lead to things like antibiotics and antiseptics etc.
(12:15) And so we live in that quadrant and you know the one below it which is okay, with that new knowledge now how do I actually turn that into a product or an offering. So, so I actually described that left axis wrong, it’s about a product or an offering and about knowledge for knowledge sake, or is it about application.
(12:39) So we live more in that right-hand side of that, and balance between the upper right and the lower right. So we do care about understanding the knowledge generation, generally it is in a focused area where we believe that we can both contribute to the quality of the knowledge and convert over time that knowledge into an innovation, right. And that’s one of the distinctions between innovation and invention, innovation is an invention that solving a problem right that people care about. And in the end we are about innovation, right, but to do that we do good science in those focused areas and that’s the balancing act that I think you’ve got to walk, and a lot of the more pure basic science, understanding the world right. You know, string theory, those types of things, that is more in an academic environment.
Michael Krigsman:
(13:37) And what are research areas or innovation areas that PARC is focused on right now
Stephen Hoover:
(13:45) We’ve got five core areas. One of them is there are a lot of people focused on big data analytics and we certainly are as well, because the amount of data in the world and the sense that you can make out of it, it is absolutely one of the forefronts.
(14:08) One of the things that we try to do is bring a unique perspective to that, which actually ties back to our history and so I called human centered big data. And what I mean there, and I’m going to start with an analogy right, so one of the things that we described earlier at PARC in our foundational days is we looked at computing technology, and we made it so that people, average people can use it; the personal workstation. That you don’t need to be a computer programmer or an expert, and we were able to do that because we understood how work got done in an office. Right, and this is again one of the values that Xerox brings to the market today, Xerox talks about work and work better. Very focused on the job that the person is doing.
(14:57) And our role then is to understand the jobs that people are doing and how technology can help make them get that job done, and in a way again that is easy for the person to adapt. So for example, one of the things that we’re doing today is in the support of Xerox, Xerox as I articulated earlier has a significant business in customer care. Right so, 2.5 million customer interactions daily, well how can we learn over all of those interactions, which is a more likely problem to solve, more likely and answer to a problem a customer has.
(15:38) and And how can we do that in a way that today you have got capabilities where computers – you know, automated chat-bots, right. You’re chatting over the computer and on the other end it’s not a person initially but a computer agent. Well, it turns out that computer agents are pretty capable, but there’s a place where they run out of knowledge right, and what you don’t want to do is tick off the customer on the other end because what starts to become apparent is this stupid agent asking them stupid questions and not really resolving their problem.
(16:16) So we are developing technologies to actually have dialogues and monitor that dialogue. Is the person getting frustrated? Does the conversation seem to be progressing, and if not, how do you seamlessly hand that call over to a live agent in a way that nobody even knows or can detect. That all of a sudden there is a person on the other end answering the question because the computational agent has run out of its ability to solve the problem.
(16:44) So this idea of, again human centered big data, here the human is you know both sides of that equation, the end customer, so really understanding their state of mind and what they’re doing. Then the call centre agent and how are you helping them take over that call, and then as I said, actually using that as a learning opportunity. Now, we watch what the customer care agent, the human does to solve the problem, we take that data and through machine learning, make our diagnostic algorithms better for the next time.
(17:21) So I personally think there is a kind of a really big idea in here that is there is a lot of focus on pure automation and that’s good and it will happen, but as computers become more and more capable, you’re reaching this point where we're going to have human computer teams solving problems. And how do you interact together to come up with a better answer, and there is a lot of science in that, there’s a lot of technology in that. I can give your listeners an example in this, it used to be that humans were best at chess in the world, and then artificial intelligence, technologies came along and computers beat the best humans in the world. Big surprise you know just over a decade ago too many people and too many chess players.
(18:14) But who are the best computer chess players today? Human computer teams, literally you take a person and a computer and they work together because computers are really good at deep fast search, evaluating all of these alternatives, but they are not the best at what’s the high-level strategy. So, you may be thinking three or four strategies, the computer investigates those three or four strategies and tells the human the likely outcomes, the human therefore picks the next strategy and they play back and forth in the cycle. That’s what the best chess player is today, and we’re going to see work happen more and more in that way, so we have a significant investment in this whole area of machine learning and empathetic computing, understanding again what’s people’s intent and behaviors and how do I help them.
Michael Krigsman:
(19:07) So a core foundation of what you’re doing is this notion of simplicity and making the problems applicable so that people in ordinary offices can make use of these technologies
Stephen Hoover:
(19:22) Exactly, work can work better and it’s interesting you talked earlier you know at PARC we have a large number of scientists and engineers in the physical sciences and electrical engineering and chemical engineering, but we also have social scientists. Cognitive psychologists, who are experts at conversational analysis, ethnographers who go out and understand and observe and again as people are doing a job. What is the real job they're doing and what is the things that are hard and easy about it and what can help them best.
(20:00) So recognizing again, you got it dead right, this is about how do we do this simply and in ways that stick with people that actually they can use, requires that kind of deep understanding as well as you know, design thinking and those types of things. It’s more than that, there's an actual science into that. And we do work in areas again of modelling cognitive processes of people right, how do people think.
(20:32) We have some very interesting work in those areas particularly around human behavior. You know one of the hardest things to do is to change yourself right, be it I want to eat less, I want to exercise more, whatever that is, there’s an actual science of how change happens. And a lot of people are saying you know, we’re going to build apps to monitor your food and be the nanny on your shoulder telling you what to do. That’s not enough, you actually have to, again understand how people do change behaviors and then build those artificial intelligent agents and chat-bots to provide the suggestions and advice to people in a way that actually sticks.
(21:19) So for example we’ve gone out and it turns out one of the best ways for people to make changes in their lifestyle is a coach, and we’ve done studies of how coaching works, why does that work and how could you provide the same, and coaches are very sensitive and very dynamic right, watching how you’re behaving, how do you respond and then giving you the right nudge and are you the right person. You know, some people respond more to you know, ‘hey why don’t you try harder’ and some people respond more to ‘hey that was a great job’ right, and a good coach knows what you are and picks that up and tunes that message and styling to you. Again, how do we have computations to determine your personality profile and address you in that way, and there’s a lot of science and research we’re doing in those kinds of areas.
Michael Krigsman:
(22:10) I spoke yesterday with a woman named Julia Hu, who is CEO of a company called Lark Technologies which is actually working precisely in this area you were describing of coaching. So they take sensor data from phones and then they take human experts and they deliver the coaching through an app, so it sounds like this is a similar kind of area.
Stephen Hoover:
(22:38) Yes a very similar area and then how do you engage you know people’s social network right, into this in a fruitful way, and again there is some science there about what makes good teams, right, how to build a social circle in this case that is again, the right mix of people and the right kinds of things and yes how do you create you know, AI agents who can perform some of the coaching capabilities. That’s dead right, what’s interesting we’re doing this for a variety of reasons. I mean it’s actually very relevant as I said for Xerox as well as some of our customers outside of Xerox.
(23:26) We do a lot of work today with Healthcare companies in helping them kind of manage their information processes and more and more information is about their patients. You know, what are they doing and how are they behaving, so we see moving from not only kind of managing the digital flow of information inside the enterprise, but helping enterprise organizations serve their clients more directly. I mean as everybody becomes more and more connected I think there are fewer and fewer B2B companies and everybody in some sense becomes a B2B2C connected customer.
(24:13) Again, Xerox is more of traditionally a B2B business, but as our customers are business customers are connected to their customers in real time, which we all are; we’re a connected world, we all become a B2B2C company, so that’s part of the progression that we’re on.
Michael Krigsman:
(24:27) So as you’re dealing with these future of work issues, you talking about technologies but at the same time there’s just as much a cultural and human and social set of dimensions so you must be addressing those as well.
Stephen Hoover:
(24:44) Yes and there’s kind of two sides to that and one is inside our own innovation environment right. so one of the things that we really value at PARC, we think a lot of innovation comes from multidisciplinary work, which kind of makes sense because okay, there’s a well understood area or field, a silo, a technology and you make progress in that field, then there’s another one. And if you can connect the two in interesting ways then you know that’s a white space; it’s a new opportunity.
(25:29) And so I talked about there’s a lot of focus on big data and on what computers can do and we’re saying, but wait, it’s also about how people learn and how they think and you know, marrying those two together.
(25:38) So, one of the cultural things we try to do is to leave space for people to explore those white spaces because I think that’s a key piece of innovation. The other pieces is you know, a lot of ideas when you start out it’s not clear where they’re going or that they make sense but some of them will. But if you try to kind of narrow those down and control those too early before the people got the time to experiment and figure that out, then you’re going to miss a lot of innovation opportunities.
(26:17) On the other hand if you just keep pursuing a bad idea (26:21 – 26:23 Lost audio), well you’re wasting everybody’s time and money. So it’s a real balancing act on how do you give people enough freedom to kind of explore something at the beginning that it’s not clear that it makes sense, but they’ve got to have the ability to go and figure that out. And you’ve got to let people be – you know, you’ve got smart people who can think out of the box, you know you’ve got to let them be a little crazy every now and then. And if you try to turn this too much into a tops down process you’re going to end up in a place that’s not very innovative.
(27:02) So there’s this real balancing act between how do you set some large strategic goals, and then how do you give people the white space to innovate, but how do you devine the good ideas from the bad ones and kind of invest more in those, so there’s a whole art to that.
Michael Krigsman:
(27:22) So how do you herd the genius set of cats inside that crucible of the market of reality to produce something that’s absolutely great.
Stephen Hoover:
(27:37) Well I’ll push on you a little bit is yeah that’s a piece of it but how do you let them herd you. You know it isn’t one way. I mean I don’t know where the right places to go, all right I’ve gotten really excited. You heard me earlier about the idea of human computer teams, and you know it does start to resonate with me as I said, I see a strong value in this idea of connecting humans and technology and solving their problems together, and again that’s just who PARC is.
(28:14) But framing this idea of new human computer teams, right, that’s not something I came up with. That’s the cats who herded me right and they start to talk that way, that’s an interesting metaphor. Right, on some levels it’s a metaphor, but metaphors are powerful because they give you different ways to think about things. And so when you start to think about it that way that’s a different way. So it’s not only how do you herd those cats but how do you get herded.
(28:42) That’s a really important piece, and if you want people who are innovative you’ve got to be open to that. I mean or they’re going to go somewhere else where people will stop and think differently right. So if I can’t think differently after some of the people who are working in here who have new ideas, then you better fire me because I’m not going to get the job done. So it’s a two-way street.
Michael Krigsman:
(29:07) So you yourself being adaptable in terms of what you’re doing and I’m assuming also in terms the strategy of the organization to some extent.
Stephen Hoover:
(29:18) The strategy of the organization, what are the you know the technical directions, as I said this whole framing around human computer teams it’s a relatively new framing. And I think again one it’s differentiated and valuable which is what we ought to be doing.
Michael Krigsman:
(29:34) What about some other areas, such as Internet of Things or automation and robotics. I know you’re doing work in areas such as that so talk about that a little bit.
Stephen Hoover:
(29:47) So the couple of different ways into that. So certainly the Internet of Things we see as the analogy that I make for people, it’s a really important trend for people who haven’t kind of thought about it deeply but have heard about it I think it’s a useful analogy.
(30:12) If you think about it you know one of the things that Google and search has done for us is it has infinitely expanded the capacity of the human memory, right. I don’t have to memorize all the facts in the world. I can go out and look them up and find it if I want to learn about reinforcement learning. You know, I go and Google it and I find the Wikipedia article and I learn about it.
(30:41) So it’s expanded my brain, my memory to be you know to nearly infinite capacity right, that I can quickly and easily find that new information and integrate it when it’s relevant to me on a preference.
(30:53) The analogy I want to make for you and your listeners about the Internet of Things is it’s about Googling reality. It’s about right now my body right, my body is the sensor. I see things, I hear things, I sense the world around me, well why does that have to be geosynchronous and why does it have to be synchronous in time and space from where I’m at, right which is what my censors normally are right.
(31:21) You know, I sense something at the time it occurs and it has to be near my body, well, the Internet of Things says no it doesn’t. I want to understand the state of pollution in Beijing; I go and find it on the internet, right.
(31:33) So this idea of kind of expanding your body right, so now take that idea and that’s for an individual not for an organization right. I can instrument and understand what my customers are doing with my products across the world now. I can understand are those devices starting to fail. I can understand the environment they’re in and adapt their behavior to be responsive to the local environment, right.
(32:05) You know, GE and their jet engines, when a plane’s running into a headwind you know, run the jet engine differently because I know I’m in that situation. So this idea of the Internet of Things and then you couple it with the data analytics and the machine learning so that I can make sense of all that data, right, to get a job done, what’s the problem you’re trying to resolve. And in our case Xerox we put – again transportation systems, cameras, where our cars are going. We as I talked about earlier about EasyPass, that’s the Internet of Things. You put a little transponder in your car, it senses as it goes by, it knows it’s you, it charges your credit card. But now, why do I have to put the transponder in the car, I’m going to have a camera up there to take a picture of your license plate because computers can see.
(33:00) And now I can see how many cars go by and I can help cities understand how to manage traffic better. You know there are cameras across cities for a whole bunch of reasons now. I can be looking at parking and helping people find where’s the open parking space.
(33:17) Connected cars, cars drive by, you know, how many cars today and I think it’s a regulation either today or next year that all cars will have a backup camera. Well, you’re driving down the road and your camera is seeing where there’s empty parking spots, how do I get that information and direct people to the open parking spots.
(33:38) To optimize again the capabilities for people to get the job done they’re doing, and so in our view that’s the power of kind of the Internet of Things and going to capture that power and then from a technology viewpoint where we’re working, whether it’s analytics to kind of optimize outcomes and behaviors.
(33:56) Another, we have a significant investment I talked about our foundational work with the creation of Ethernet, we have a totally radical new network and technology called Content-Centric Networking that really deals with two things; the explosion of information that is occurring and is just going to radically increase with the Internet of Things. Because now if I can Google reality, imagine all that data. And the reality is current internet protocols isn’t designed to handle that quantity of data. And the other piece as we all know is security.
(34:30) It’s not built into the network, it’s an add-on so Contract- Centric Network that builds security inherently into the network and therefore is much more secure.
(34:42) And then the third area for us is the Internet of Things, I talked about a lot of it is it’s sensing the world. How do I do low cost, highly distributed sensing if I’m going to put the things on the internet and if it takes $100 to put a smart computer on a bottle of vaccine to measure it’s temperature during to the shipping, well I won’t do that. but we’re working on technologies in areas like printed electronics, how do I make very low-cost electronics that are smart enough.
(35:21) Silicon has got cheaper and cheaper to put on more and more transistors. You know, my iPhone is the power of a crazed super computer from 1996 and I can buy it for 200 bucks with my 24 month plan. But if I want to put a temperature sensor on a bottle of vaccine as I said for 50 Cents, the reality is I don’t need a whole lot of intelligence, and it’s gotten cheaper with silicon to cram more and more intelligence, but what I want is a whole other price down at that dollar, smart label, sensor that I can put out and sense those things. So we have a whole focus area on printed electronics because we think that’s the Internet of everyday things. And we have a set of partners in that area and we’re also working with Xerox you know to take some of those initial technologies to market.
Michael Krigsman:
(36:06) How does this actually work? So do you have projects with commercial companies, what are the mechanics of that?
Stephen Hoover:
(36:14) So we have our sort of open innovation business. So the mechanics of this are, yes, we have four classes of open innovation models. I mean, one is a pure kind of collaboration, right. We’ll work in an area and we’ll go out and find a university or a commercial company who has got capabilities in those areas and we form an agreement to team and kind of I’ll solve this part of the problem and you solve that part of the problem, and we’ll share and compare information. And that’s a kind of a classic way to a lot of open innovation that happens, and by and large at most companies it means universities.
(36:58) The different thing that we do is we also go out and we do a fair amount of early stage, high risk research under Government funding, because it turns out the Government is willing to fund and has a process to fund a fair amount of potentially high value, but high value risk, longer term research and technology and development.
(37:23) So we go get research contracts with DARPA if you’re familiar with them, the crazy wing of the defense department research right. we don’t do any classified work but a lot that they do isn’t classified because they’re trying to again just uncover new areas. We do work with the National Institutes of Health; we do work with the Department of Energy in those areas. And then we also work with commercial clients where we are helping them innovate in new spaces and that’ll take two kind of forms.
(37:59) Sometimes it’s a problem that they have and we co-ideate with them, how could you solve that problem, we’ll generate ideas; generate a product out of that. we have a whole process for this. Again, kind of our history in innovation, there’s a way to approach that initial stage of ideation. And we work with them and develop that and then, if we find a good idea they’ll partner with us, fund us to do the forward looking research and innovation in those areas.
(38:31) And sometimes it’s again innovation that we’ve already developed, be it on government funding or other sources where we’re partnering with them to take that to market. I can give some examples if that would be useful for you.
Michael Krigsman:
(38:46) Well you know, we’ve only got about five minutes left, and I wish we had another hour but maybe tell us about automation and robotics, what are you doing with respect to that?
Stephen Hoover:
(39:07) Our focus comes in two ways and one is ties very much to what I articulated today, kind of half of Xerox’s business is in a document and information intensive business processes. And so there is what we call robotic process automation, which basically is looking at current workflows and work streams that are not automated, and how do you automate them in a digital world, right. So this is physical world automation and this is digital automation, and it takes the characteristics of the analytics and things that I talked about. And again automating customer problem resolution, and so there’s a whole focus in that area.
(39:49) In terms of the automation in the physical world, we’re focused on – and this is one of those areas where we’re kind of in advance of where Xerox current business is, but I think there’s some interesting long-term opportunities for what we call ‘Systems of Systems’.
(40:12) So here we’re doing some interesting work with the defense department about – you know a passed example that I can talk about what we’ve done is satellite swarms, which is basically the idea is right now, we build one big satellite and send it up in space, and if that satellite is really expensive and if it fails you’re done.
(40:36) What if instead I could build a series of small satellites that were all individually re-deployable but can be controllable in some coordinated way. So you know, it’s a swarm of 50 satellites, small and cheap. Well if one dies that’s okay. When I want a lot of imagery on a certain place I’ll aim 50 of those at the same place, when I don’t I distribute them, and there’s a whole challenge of basically if you think about it you’ve got this complex system and you’re trying to redesign it constantly during its use because you’re trying to have it do different things. As pieces fail, and don’t fail, multitask again to look at different problems, to sense different things.
(41:18) So there’s a whole science around kind of AI planning and thinking how do you manage that system of systems. The other piece of it is, back to my human and computer team. In the end those systems are being tasked by a human. How does a human interact with that level of complexity and manage it and hope the system has local autonomy understand what the human that begat the task, what the human is trying to perform and coordinate and adapt together. And so we’re working in that space and I think about it as the Google autonomous car, well what do you do when you’ve got 1000s of autonomous cars on a road, and how do they behave together. And how do they behave with the humans that interact with them. so we’re at that system level of the problem, and we think that’s where the next wave of complexity as automation occurs it’s going to be systems of systems interacting and the science and again the interactions with people we think has interesting problem spaces.
Michael Krigsman:
(42:28) So a core part of your activities it sounds like this notion of where the person and the computer come together, and how does the person accomplish something of value of using the computer, and having the computer be structured and the user experience be structured so that the human can manage that complexity.
Stephen Hoover:
(42:51) Exactly and again there’s a whole set of science around that, someone goes that’s design no it’s more than design, you know it’s how do I understand, how does a computer build a model of the world that’s a shared model with the human because there’s assumptions. Again, you and I are having a conversation and there’s a lot of implicit assumptions that we share and those evolve over the conversation. How do I do that? Again understanding are you a pessimist or an optimist, are you a realist and how do you take information in better.
(43:27) And what is the best way to split the tasks that again leverages the strengths of computers and people, as computers become more capable that becomes more important and harder, and how do you leverage that. So there’s a whole host of scientific and technical challenges in there that we’re investigating.
Michael Krigsman:
(43:47) Well it’s a fascinating conversation Steve, as we go, as we leave unfortunately our time is just about up, what advice do you have for organizations who want to innovate, maybe they’re stuck with their traditional processes, and having resistance and pushing back, folks in those organizations, how can they do a better job of pushing innovation.
Stephen Hoover:
(44:14) The two biggest things that I see that people struggle with are 1) Is this idea of risk. Exploring new areas and ideas is inherently risky, and so much of our business process where you’re about delivery of something is about risk avoidance, and in innovation you need to do the exact opposite. You need to embrace risk, not because risk is good but because the reward that comes with solving that problem is valuable, and as I said it’s not about risk avoidance. I said earlier there’s three kinds of risk: technical risk, market risk, business model risk. Wherever the biggest risk is, how do you create an experiment to uncover is that risk, risk means I can’t predict the future. To predict what is going to happen or to design a system to you know, so that that risk doesn’t come true.
(45:17) So there’s this weird mentality shift of embracing risk and then agile learning processes around that. the other piece I would say is the hardest thing for existing businesses to do is to innovate in a way that challenges your current business model.
(45:38) New technologies can radically change business models and if you’re not open to that you’re going to get caught behind the eight ball. So how in your organization are you leaving space and it can be very painful at first, I mean you know if you’re used to selling hardware and software and cloud computing comes along, you’re going to think you’re going to make less money because people are going to buy fewer computers and you don’t get as much money off of every unit sell in the cloud. They can buy it by the drink, they don’t have to buy it at peak capacity, they buy you know the dynamic capacity.
(46:20) So yeah, you’ll make less money on each transaction, but the consumers in the end will consume more of it, and that’s the opportunity you’re going to get but it’s really hard for existing businesses to see that. so set up space where you are innovating and exploring innovation in ways that hurt your own business. If you don’t do that and technology changes your business you will end up a victim and you don’t want to be there.
Michael Krigsman:
(46:42) Well, it sounds like what you’re describing is the philosophy that you said earlier that you aim to yourself, but you have to let the cats hurt you as well, which means be open to change, be adaptable and flexible.
Stephen Hoover:
(46:58) Yes very good.
Michael Krigsman:
(47:01) Well Stephen Hoover, CEO of Xerox Palo Alto Research Center, known as Xerox PARC thank you so much for taking the time. This conversation has gone by in a flash and it’s been fascinating. Thank you so much.
Stephen Hoover:
(47:18) Thank you, take care.
Michael Krigsman:
(47:22) And to everybody who has been watching we really appreciate it and come back next week where we will be talking with the CIO, the Chief Information Officer for the city of Palo Alto. Thanks so much everbody, have a great week, bye bye.
Mentions in today’s show:
DARPA: www.darpa.mil
GE: www.ge.com
Google: www.google.com
Lark Technologies: www.lark.com
Wikipedia: www.wikipedia.com
Xerox PARC: www.parc.com