From the heart of Silicon Valley, it's theCUBE covering Comcast Innovation Day. (smooth music) Brought to you by Comcast. Hey, welcome back everybody. Jeff Frick here with theCUBE. We're at the Comcast Silicon Valley Innovation Center here in Sunnyvale. Very cool facility right off the runway from Moffett. They got a ton of cool toys downstairs which I get to go play with, which I'm looking forward to, but today the conversation was all about CX, customer experience, and you know, Comcast is there. A lot of people like to watch their TVs, interacting with their cable systems for a long, long time, but there's a whole range of new and innovative things that are coming out from Comcast, and we're excited to have an engineer who's kind of down in the bowels here in the engine room building all this stuff. So like to welcome Mike Fine. He's a cable software architect for Comcast. Mike, great to see you. >> Likewise. So you had a really cool demo earlier, which is not a demo, right? I think this thing is-- >> Production. Now in production, it's called the X1 Eye Control. I think most people know what X1 is. What's X1 Eye Control? >> Yeah. X1 Eye Control is a web application that integrates with off-the-shelf accessibility hardware, so that could be a Tobii eye gaze rig, it could be something called a sip-and-puff, which let's users use their inhalation and exhalation to control the application, or any other off-the-shelf accessibility hardware that can mimic a mouse to a piece of software. Too, it's-- >> Yeah. The goal of the project was pretty simple. It was to let people with ALS and other conditions control their TVs independently. That's amazing, and you showed a great video. The gentleman on the video is using I think an eye gaze method, but you said you've got integrations to a number of different, you know, kind of ADA-approved interface devices. That's right, the journey that this project has taken has been interesting. We started with just the ALS use case, which was the eye gaze, but it turned out that one of our early users had control over his voice, which is somewhat unusual for ALS patients, and so he asked whether he could control it with his voice, so we did that work through he had Dragon NaturallySpeaking, which was nice, so we did that work, and then of course given that we have the voice remote we decided could we make voice work for everybody, which we did, so now the application is on par with a physical remote, and then we even went further and let people type in voice commands, so in case somebody who's perhaps mute or had a speech impediment, or some sort of speech pathology issue that prevented them from using a voice, they could do that as well. >> It's really interesting, I mean you guys have so many kind of interface points to an ecosystem broader than simply what's available at Comcast, whether it's on the front end, as you said, with some of these interfaces with ADA devices, or on the backend if I want to watch my Netflix or I want to watch YouTube, or I want to watch, you know, a different service. You guys have really taken, you know, kind of an open, integrated approach to all these, one might argue, competitive threats to really bring it in as the customer wants to experience. Why did you do that, what's kind of the philosophy driving that? Yeah, well, the first thought that comes to mind is that none of it's possible without the right cloud APIs, so somebody very visionary years ago made the decision that everything you can do on your TV or on the mobile app you can do through the cloud, and so a project like this couldn't happen unless it was possible for a piece of software that somebody invented well after the fact to cause a TV to change channels unless there was that underpinning, so like any other piece of software it's a bit of an iceberg. There's a lot of stuff underneath that you don't realize as a user-- Right. >> But it's there and that's what makes it possible. Right, I'm just curious about some of the challenges in terms of moving UI and UX forward into places that people are not familiar with. And I've joked about it on a number of these interviews that, you know, I still get an email, not only from Comcast, but from Google and from Alexa, suggesting to me ways in which I might use voice. You know, as you sit back from a technologist what are some of the challenges you guys, you know, kind of anticipate, what are some of the ones you didn't anticipate, and how do you help us old people, you know, find new ways to interact with the technology? Yeah, it's a great question. I mean there's a lot of us here that spend our days solving that exact problem, right? Part of it is is notifying you of interesting things through SMS or through mobile push, or the messages on the TV, so your team is playing in a game that you want to see, a movie that you've declared interest in has become cheaper, become free, or maybe even buyable if you wanted to do that. Obviously there's lots of AI and ML in terms of putting recommendations in front of you based on your viewing habits, based on broader trends across, you know, because you watch this, other people watch this, so we know this is probably a good solution for you as well, but yeah, we're all, there's a large number of us trying to optimize what we call "time to joy," from the time you pick up your remote to think about what you want to watch to the time you're actually watching something you want to watch; make that as seamless as possible. Preston said you guys get like a billion voice commands, what was the period of time? A month. >> A month. A month, yeah. (chuckles) So obviously a big, giant new dataset for you guys now to have at your disposal. What are some of the things that you're learning from that inbound, what can you do with it, how do you, you know, now use this direct touch with the customer to, again, kind of recycle and have another iteration on improved experience? Right, so voice is a lot like a text chat, like a bot interface in that it's an experience where users are telling you exactly what they want to do, so if a user sits in front of a traditional web application or mobile application and has trouble finding what they want to do, they can't figure out what button to press, what screen to go to, you have no idea, right? You can't infer that they're having a problem, but with voice or somebody interacting with a bot, they type exactly what they mean, or they say exactly what they mean, so we can mine those voice commands and find the popular ones that we don't at that point have implemented, and if we can iterate on that cycle fast enough we can quickly introduce new voice commands that our users are literally asking for as quickly as possible. >> Right. What about the stuff that customers are not asking for, because right? There's one line of thought, which is the customer knows best, but the customer doesn't know-- That's right. >> What they don't know. So how do you guys continue to look for more kind of cutting edge stuff that isn't necessarily coming back through a feedback loop? Right, yeah, so it's an interesting question. So we're trying to add other non-TV use cases into the mix, right, so controlling your IoT devices at home, controlling your security, seeing your cameras through the Set-Top Box, and so on. So you know, until those use cases exist nobody's asking for them, and so you do have to be a bit visionary in terms of what you want to put out there as voice commands. You know, luckily we have people who, well, we're all customers of the platform generally, so we know what it means to be a user, but you know, we have people that talk with users and have a general sense of what they want to do, and then we figure out what the right commands are. Right, not voice specifically, but let's unpack a little bit deeper into the impact of IoT. You know, Nest probably was the first kind of broadly accepted kind of IoT device in the home, and now you got Ring, which everybody loves to take pictures of people stealing their boxes from the front porch, but that puts you guys with the internet connectivity in a very different place than simply providing a football game or the entertainment. So as you think of your role changing in the house, specifically with now these connected devices, how do you think about new opportunities, new challenges that being the person in the middle of that is different than just sending a TV signal? Yeah, there's a lot of talk about trying to be the home OS. Certainly we are in a unique position being in the home, both in terms of the router and the internet, but also, you know, often frankly you know when your system's setup a human being came in and helps you understand how to best position the physical devices in your house, and so on, that other companies don't have, right? Those vendors just don't have that builtin advantage. Clearly security has become a big thing for us. Home automation, I sit very close to that group. They're doing amazing things with automating rules like, you know, "Tell me when my door's been open too long," and these sort of things, and so more and more the use cases start to converge, that, for example, when you say, "Good morning," we have this idea of scenes, all right. So when your morning starts you not only want to tune the TV, but you also want to crank up the lights and unlock the door and open the windows, or whatever, and when you go to bed, so the actions that are involved in those use cases span not just TV and not just internet, but all of it. Right, it's just funny because I don't think Comcast would be the first name that people would say when they're talking about voice technology and the transformational impact of voice technology, right? They're probably going to say Siri was the first and Alexa's probably the most popular, and you know-- >> Right. Google's got Lord knows how many inputs they have, but you guys are really sitting at a central place, and I might argue it's one of the more used voice applications-- >> Absolutely. Out there, so from kind of a technology leadership perspective you guys have a bunch of really unique assets in terms of where you are, what you control, what you're sitting on in terms of that internet. You know, how does that really help you and the team think about Comcast as an innovation company, Comcast as a cool tech company, not necessarily Comcast as what used to be just a cable company? >> Right, right. Well you know, as somebody in the valley with friends in the valley it's always interesting to try to differentiate reality from the view that many people have. You know, this is definitely much more than your dad's cable company. It's a consumer and electronic company as much as anything else. We very much position ourselves with all the, you know, with the FAANG companies, et cetera, so you know, when we talked about CX it's no longer the case that whatever's passable for a stodgy cable company passes as CX anymore. Now you're being compared to a set of customers, companies that are providing fantastic user experiences for their customers, and you're being held to that standard, so you know, there's a lot of pressure on us, which is great; we like that. We want to produce fantastic products, and yeah, I don't know if I have a great answer in terms of how to move forward in terms of melding it all together, but we have a lot of smart people in the hallways making that happen. (chuckles) So last question is really the impact of AI, because you know, we cover a lot of tech events and a lot of talk about AI, but you know, I think those of us around know that really where AI shines is applied AI in specific applications for specific U cases. So how are you guys, you know, kind of implementing AI, where are some of the opportunities that you see that you can do in the future that you couldn't do the past, whether it be just with much better datasets, whether it be with much faster connectivity and much better compute so that you can ultimately deliver a better customer experience using some of these really modern tools? Right, so some of the work is just making what you already do or experience better, so for example showing you recommendations, right? Just make that algorithm better, and so there's a great deal of effort, as you might expect, at a company like this on that problem, but there's also work being done to just take any interactivity between you and the system out of the picture completely. We talked a little bit about this earlier, that, for example, we're working on technology that when you turn your TV on in the morning it should probably tune to the channel that you normally tune to in the morning. That's a pretty simple problem, in a sense, but you know, if I watch your viewing patterns and I see that you turn on a particular news show in the morning, why should you have to pick up the remote and change it from what you watched the night before to that channel? It should just happen. We talked about the Smart Resume stuff, that's obviously a fantastic use case for end users, so there's, you know, it's not surprising it's being used all over the technology set. It's in the home automation world. You know, it's in A/B testing, so trying to figure out the right cohorts to try different things in front of, so it's everywhere as you would expect. Right, right, it's pretty amazing. I mean there's just so many things going on, you know, kind of under the covers, some that we can see, some that we can't see where you guys are really kind of progressing, you know kind of the leading edge, cutting edge customer experience with something that people interact with every single day. That's right. >> Yeah, cool stuff. Well Mike, thanks for taking a few minutes. Congratulations on the Eye Control; really a cool story, and look forward to more publicity around that because that's a really important piece of technology. >> Thank you very much. It's been a pleasure. >> All right. He's Mike, I'm Jeff, you're watching theCUBE. We're at Comcast Silicon Valley Innovation Center. Thanks for watching, we'll see you next time. (smooth music)