>> Narrator: Live from Washington D.C., it's theCUBE! Covering Boomi World 19, brought to you by Boomi. >> Welcome to theCUBE. Lisa Martin, with John Furrier. We are in Washington D.C. at Boomi 19. This is our second day of coverage and John and I are very excited to welcome one of Boomi's customers, we have Olive Perrins, head of In Home Experience at Sky. Olive, it's great to have you here. >> Hi, lovely to be here, it's a fantastic event. >> It is, we saw you onstage yesterday so we're very pleased to have you join us on theCUBE, so I think a lot of folks know about Sky. Everybody, I shouldn't say everybody on the planet, but most of us have an ISP. We have cable services. So, we're all customers of Sky or some of Sky's peers across the globe, so, we all kind of understand that. You guys have built something very cool with Boomi, ‘Future Assurance View’ tool, and when you taught that, when you showed me this before we went live, (exclaims) bring that to the U.S. because whenever there is a problem with our internet, I mean people, we just stop, right? >> Yeah. >> So, talk to us about what Sky has built with Boomi and some of the great things that it is enabling. >> Sure, so I think we always had amazing diagnostic information and we had lots of data. What we never did was connect at that and did data-driven decisioning. So, for us, Boomi was there to connect all of the sources together with the, over six million routers out in the field, and live on demand for a customer, check everything, all of the, telemetry data from their hubs, from their line and make sure that line is connected, it's fast enough, it compares well to their neighbors, it's stable, it's not retraining, it's as good as the line can be, and the wifi to every device in the home is good. If not, it simply decides which engineer it needs to fix this and dispatches the job. >> And you started this initially in a reactive mode to start, okay, there's faults here. Talk to us about that migration, we'll say transformation, since we're here, the transformation from reactive to proactive and then unveiling what you guys are doing with predictive. >> Yeah so, I think when we started, we set ourselves this big game of getting to 69% digital first. We were around 25% before we lunged, and to be honest, most of that 25%, it was to find the telephone number in digital rather than do anything. We're now at 87% and as you can imagine, the amount of data logs that creates, about 300,000 customers a week running the tool, has now led us to know which outcome is most reliable and really optimize our decisions. So, then we started to think, "Okay, well it's great that we're fixing "these issues, but we probably have "a lot of customers in pain." Who we're not getting to because they're not calling us or visiting the tool. Why don't we go proactive and then go predictive. Find who's going to be faulty tomorrow and intervene before it happens. So, we've taken all of the intelligence in Boomi and codified it into an algorithm, and every night, it runs and predicts who'll be healthy or unhealthy a signal tomorrow, and then, anyone who needs an engineer, we dispatch it and it just fixes it free of charge before the customer even knows it's broken. >> And was this, I'm just envisioning of the recent issues I've had with ISPs, ah, I need this. Was this driven by, you said initially, just a couple of years ago, only a quarter of your customers were, only a quarter of them were starting their search digitally and now it's up to 87% in just a two year period. What you've done to be, to go from reactive to proactive to predictive, was that driven by customer demand saying, we want, I don't even want to have to call in. I want to be able to get to you from any channel, or was it more driven by you guys suddenly having a massive increase in data, and saying, "We've got a lot more information. "If we can connect it together and unlock the value, "the services we deliver can become predictive." >> I think it was a blend of both, truthfully. So, once you ultimately master the cost per consumer, you've got a really good data model that says, given this fault, then send this engineer and we know, we'll fix it and they'll be happy. I think at that point, you start to say well, where are the other costs to the business? And ultimately that comes from churn and attracting new customers. So, it just feels right to spend more upfront on engineers to save churn later and keep a really healthy and happy base. >> You know, one of the great things about In Home Experience is obviously wifi 'cause it goes down whenever you're on screens, calls. So, the operational side to totally get the efficiencies and the savings that probably comes with that, but people are working at home more, you've seen virtual, so, there's a real need for reliability at home, but also it brings up the data and the security questions 'cause now you got wifi light-bulbs, you got, everything's wifi. So, you know, the In Home Experience now has people maybe working at home. >> Yeah. >> Home and pleasure, security, malware, all these things are cutting-edge data problems. How do you guys view that? What's the internal thinking around how to protect the home and. >> So, I guess the first thing that we needed to be really clear on is traditionally, in an ISP world, you are risk averse and you said our demarkation is where the line enters the home. That's no longer acceptable in today's age. Every time Facebook goes down, our help contacts increase by 30% and so, we know that our demarkation isn't the device, it's not the application on the device, it's the consumer themselves. It's their understanding and as an ISP, it's our job to educate, support and handhold. So, everything that we can do to make our hub smart enough that they're plug and play, and everything that we can do to predict what customers need in IOT and security, and build that in its source, it's the right thing to do. You'll have healthier happier customers in the longterm. >> And parents also want to turn the wifi off when the kids aren't doing their homework. You know, these policy kind of user experienced things are kind of, I mean, as an example, we have kids, but you know. >> We just launched a remote control for the internet, so you can control what your kids have access to anywhere, in or out of the home on any device. >> And you guys have just in this last couple of years, where mostly it's been going from reactive to proactive, you said predictive was launched recently, but even in that two years, your NPS Net Promoter score has gone up 20%, so, can you imagine in the next year or probably last, the impact that you're going to have because customers are getting what they want, and they probably, some of them don't even know it, if they don't know they have a problem, but Sky has identified it. I can only imagine that the churn numbers will go down and the NPS will even continue to rise. >> Exactly, and that's precisely what this is about, it's the happier the base is, the more stable. In the end, you're going to spend more on engineers and less on churn. That is the perfect balance, it really is. >> And in terms of spend, let's talk about the cost savings, dramatic cost savings. The first year alone, you saved a six million pounds? And the second year? >> Six million pounds and on track for similar this year. >> That's transformative to the business. >> It absolutely is, yeah. I think what it has allowed us to do is really knuckle down to watch at our budget bate, and get stability around that. So, now we've given the business some controls and dials, and they know what they can pull to control costs. >> What's next? What are you guys working on next 'cause obviously, that's good in return, you're reinvesting, it's more data, it's more things to do, it's got remote control internet. What are other things you guys looking at operationally to get into to innovate on? >> So, I think there's a real need for speed, for us it's about investing in fiber. We're putting all of our customers on a high fiber diet right now (laughs). So, it's dark fiber, faster fiber, one gig connections, and then on the wifi side, it's giving guarantees, so, it's no longer acceptable to have a router squirting out wifi. What we're now doing is guaranteeing you will have wifi of the best quality anywhere in your home to support any device, and we're putting our money where our mouth is and sending wifi heat mapping engineers with pods to get your house up and sorted right first time. Beyond that, I think it's very much going into the world of IOT, smart sensors, cameras and with that, of course, data. It means IP storage, backup for your cameras. >> One of the interesting (mumbles) trends we've been covering is automation. You're saying RPA, for instance, is a hot sector, observability on the data side, so this Cloud has, but you've mentioned the demark has changed to the user, so, you got wearables, I mean, if you've got gamers in the house, they're going to look at ping times. The kids know what ping times are. So, you're going to have all the speed issues. So, what's that going to look like for you guys as you think about more speed, more data, more people wanting custom services. Is there automation involved? I mean, where do you guys see the automation low-hanging fruit and where's the vision go? >> So, for me, it's not necessarily about automation, it's about personalization. We already have that data. We already use that data. Is it relevant to every customer? I'm sure my mum wouldn't want to know about ping, she wants to know if it's broken. So, I think for us, it's matching what's your intent and have we serviced that in an outcome, and right now, that's exactly where we're going with conversational AI, and then, really starting to consider, have we achieved your goal? RPA has a place, but I think right now, it's less about the generic quality of service and more about targeting your individual consumer needs in the home. >> I love that personalization angle because I think we sometimes in this digital age, personalization is lost. Sometimes we do that of our own, if we're going you know, on Door Dash or something to, instead of going to a restaurant. We want, I think we want a mix of both, but that personalization where something like wifi comes into play, like you were saying, when Facebook goes down, 38% spike in people calling and going, "Hey, there's a problem here." Whether or not it's Sky's problem or not. So, when we look at this demand for personalization, peoples' levels of frustration with, if there is an issue, you guys have resolved that, obviously, but in terms of what Boomi and Excentra announced yesterday with conversational AI. >> Yes. >> Really really exciting stuff there. You guys said, you and I were chatting before we went live, that there was a purposeful decision at Sky to not start this digital transformation with AI. Now, you're ready to take this on. Tell us about that decision and how you're now, really have the foundation with which to actually do it, conversationally, and make it personal. >> Yeah, and I think so much time goes into training bots and I really think that it needs to be authentic. You don't need to feel like you're talking to a human. It's okay that you know you're talking to a virtual machine, but that first interaction needs to be meaningful and helpful or you'll quickly stop engaging with it. So, I think for us, it was about define what does good quality look like? What might be the things that go wrong with Broadband? Ultimately, it really is only slow, not working at all or dropping lots outside the home or inside the home, and really it's about saying, what might be the problems we know about? Eliminate those and there's only a finite number of alternative problems left that we can really start to train a model on our learnings to date. So, I think having excluded all of the weird wonderful edge cases and dispatches, there's less there to worry about, but it's higher value for the consumer, and I think on the personalization angle, the key for us is understanding, are you tech-avoidant? Are you tech-savvy? Where are you on that scale? And which channel should we serve you up those steps in and how complicated or handholding should those steps be? And I think that's, for us, where conversational AI comes in. It's personalization, the number of steps, the type of steps and the channel that it's best served in. There is no point having Siri guide you through really complicated hands and knees wiring stuff. That's best done with some images sent through WhatsApp, for example. >> So, you guys will have the data to be able to determine, not just maybe knowing, why is this person calling in or why are they engaging with a chat bot? But to understand, what's that persons' preferred method of communication. There's that whole consumerization effect and that demand of the consumer of, you know, your mom and my mom-- >> And my handholding. >> Exactly, would have different levels. So, you're going to have enough of that quality data to really deliver the personalized experience way beyond knowing what boxes I have installed, what routers I have, what version, but also, my level of technology understanding. That's pretty cool. >> Exactly that. Exactly that, that's the destination for this year, absolutely. >> Well, sign me up. Bring this over to the U.S. And before we go, I want to note that that Sky and Boomi together, your design won the Best Enterprise Project at the UK National Technology Awards recently. >> It did, it did. >> Congratulations. >> What an honor. Thank you, it was a great night. >> Exactly, well, Olive, it's been great having you on theCUBE sharing with us what Sky is doing to really deliver a personalized experience going from reactive to proactive to predictive, awesome stuff, thank you. >> Exactly. Thank you, pleasure. >> Ours, too. For John Furrier, I'm Lisa Martin and you're watching theCUBE from Boomi World 19. (electronic upbeat music)