 Live from San Francisco, it's theCUBE. Covering Google Cloud Next 2018, brought to you by Google Cloud and its ecosystem partners. Hello everyone, welcome back. This is theCUBE's live coverage. We're here in San Francisco, Musconi West for Google Cloud's big conference called Next 2018. The hashtag is Google Next 18. I'm John Furrier, Dave Vellante. Our next guest is Tracy Gusher, Principal Data and Analytics at KPMG. Great to have you on. Thanks for joining us today. Yeah, thanks for having me. So we love bringing on the big system, global system integrators. You guys have great domain expertise. You also work with customers. You have all the best stories. You work with the best tech. Google Cloud is like a kid in the candy store when it comes to tech. So my first question is, obviously AI is super important to Google. Huge scale. They bring out all the goodies to the party. Spanner, big cable, big query. I mean, they got a lot of good stuff. TensorFlow, I mean, all this open source goodness. Pretty impressive, right? The past couple of years of what they've done. How are you guys partnering with Google? Because now that's out there, they need help. They've been acknowledging it for a couple of years. They're building an ecosystem and they want to help end user customers. Yeah, yeah. Yeah, I mean, we've been working with Google for quite some time, but we actually just formalized our partnership with Google in May of this year. So from our perspective, all of the good work that we have done, we're ready to hit the accelerator on and really move forward fast. And some of the things that were announced this week, I think our prime examples of areas where we see opportunity for us to hit the accelerator on. So something like what was announced this week with their new contact center API suite launched by the Advanced Solutions Lab, we had early access to test some of that and really we're able to witness just how accelerated some of these things can help us be when we're building end-to-end solutions for clients. It's a shortcut to the solutions because with cloud, the time to value is so much faster. So it's almost an innovator's dilemma. The longer deployments probably meant more billings. For a lot of integrators, we've heard people saying, hey, we've gone the old days of eight months to eight weeks to eight minutes and some of these techs. So the engagements have changed. At the end of the day, there's still huge demand for architectural shift. How has the delivery piece of tech helped you guys serve your customers? Because I think that's now a conversation that we're hearing is that, look, I can move faster, but I don't want to break anything. The old Facebook move fast break stuff. No, that doesn't fly in the enterprise. No, it doesn't. So I want to move fast, but I need to have some support there. What are some of the things that you're seeing that are impacting the delivery from integrators? Yeah, well, you know, some of the technology that's come that has reduced the length of time to deliver, we see and a lot of our customers see as opportunity to do the next thing, right? If you can implement a solution to a problem quicker, better, faster, then you can move on to the next problem and implement that one quicker, better, faster. So I think that the first impact is just being able to solve more problems, just being able to really, you know, apply some benefits in a lot more areas. The second thing is that we're looking at problems differently. You know, the way that problems used to be solved is changing and that's, you know, most powerfully noted, you know, as we see at this conference, by what's happening with artificial intelligence and with all of the accelerators that are being released in machine learning and the like. And so there's a big difference in just how we're solving the problems that impacts it. What are some of the problems that you guys are attacking now? Obviously, AI's got a lot of goodness to it. What are some of the challenges that you're attacking for customers? What are some of the examples? So, you know, our customers have varying problems as they're looking to capitalize on artificial intelligence. So, you know, one of the big problems is where do I start, right? And often you'll have a big hype cycle where people are really interested. Executives are really interested in, you know, I want to use AI. I want to be an AI enabled company, but they're not really sure where to start. So, one of the areas that we're really helping a lot of our customers do is identify where the low hanging fruit is to get immediate value. And at the same time plan for longer strategic types of opportunities. The second area is that one of the faults that we're seeing or failure points that we're seeing in using artificial intelligence is failure to launch. And what I mean by that is there's a lot of great modeling, a lot of great prototyping and experimentation happening in the lab as it relates to applying AI to different problems and opportunities. But they're staying in the lab. They're not making it into production. They're not making it into BAU, business as usual processes, inside organizations. So, a big area that we're helping our clients in is actually bridging that gap. And that's actually how I refer to it. I refer to it as mind the gap. As a great example, I hear this all the time, classic. Is it, what's the reasons? Just group think, I'm nervous, there's no process. What's holding that back from the failure to launch? There's a few things. So, the first is that a lot of traditional IT organizations embedded in enterprises don't necessarily have all of the skills and capabilities or the depth of skills and capabilities they need to deploy these models into production. There's even just basic programming types of gaps where a lot of models are being constructed using things like Python and a lot of traditional IT organizations are Java shops. And they're saying, what do I do now? Do I convert? Do I learn? Do I use different talent? There's technology areas that prove to be challenging. The other area is in the people. And I actually spoke with an analyst this morning about this very topic. And there's a lot of organizations that have started productionalizing some of these systems and some of these applications. And they're a little bit discouraged that they're not seeing the kind of lift and the kind of benefits that they thought they would. And in most cases- Ooh, the customer's the analyst. The customer. Yeah, I was having a conversation with an analyst about it. But in most cases, it's not that the technology is falling short. It's not that the model isn't as accurate as you need it to be. It's that the workforce hasn't been transitioned to utilize it. The processes haven't been changed. Operationalizing it, yeah. The user interfaces aren't transitioning the workforce to a new type of model. They're not being retrained on how to utilize the new technology or the new insights coming from these models. That's a huge issue, I agree. Isn't there also, Tracy, some complacency in certain industries? I mean, you think about businesses that haven't yet totally transformed. I think of healthcare. I think of financial services as examples that are ripe for transformation, but really haven't yet. And you hear a lot of people say, well, you know, it's not really urgent for us. We're doing pretty well. You know, I'll be retired by then. There seems to be a sense of complacency in certain segments of enterprises. Do you see that? I do. And I'll say that we've seen a lot more movement in some of those complacent industries in the last six to 18 months than we have previously. I'll also say, you know, going back to that, where do I start element? There's a lot of organizations that have pressing business challenges, you know, those burning platforms. And that's where they're starting. And I'm not advocating against it. I'm actually advocating very much for that because that's how you can prove some real immediate value. So some organizations, particularly in life sciences or financial services, they're starting to use these technologies to solve their regulatory challenges. So how do I comply faster? How do I comply better? How do I avoid any type of compliance issues in the future? How do I avoid other challenges that could come in those areas? And the answer to a lot of those questions is if I use AI, I can do it quicker, more accurately, et cetera. Are you able to help them get ancillary value out of that? Or is it just sort of, you know, compliance a lot of times it's like insurance, right? And if I don't do it, I get in trouble or I get fined. But are you able to, and this is like the holy grail of compliance and governance, are you able to get additional value out of that when you sort of apply machine intelligence to solve those problems? Yeah, yeah, that's always the goal. Solving the regulatory problem is certainly, what I would say is are the table stakes, right? The must have. But the ability to gain insight that can actually drive value in the organization, that's where your aim really is. And in fact, we've worked with a lot of organizations, take life sciences. We've worked with some life sciences organizations that are trying to solve some compliance issues. And what we've found is that many times in helping them solve these compliance issues, we're actually gathering insights that significantly increase the capability of their sales organization. Because the insights are giving them real, real information about their customers, their customers buying patterns, how they're buying where they might be buying in properly. And it's not the table stake of what we're trying to do. The table stake was maybe contract compliance, but the value that they're actually getting out of it is not only the compliance over their distributors or their pharmacies, but it's also over the impact that they're going to have on their sales organization. And for something like an internal audit department to have value to sales, that's like holy grail stuff. Yeah, right. What about the data challenges? I mean, even in a bank, who's essentially a data company, the data tends to be very siloed, maybe tucked away in different business units. How are you seeing organizations, how are you helping organizations deal with that data silo problem, specifically as it relates to AI? Yeah, it used to be that the devil was in the details, but now the devil's in the data, right? And there was a great Harvard Business Review article that came out, and I think Diane Greene actually quoted this in one of her presentations, that companies that can't do analytics well can't do AI yet, right? And a lot of companies that can't do analytics well yet, it isn't because they don't have the analytical talent, it's not because they don't know the insights they want to drive, it's because the data isn't in the right format, isn't in usable to be able to gain value from it. And there's a few different ways that we're helping our clients deal with those things. Just at the very basic level is good data governance. Do you have data stewards that are owning data, that are making sure that data is being created and governed the right way? And that's a huge deal, I'm asking for change. Data quality, meta data, lineage of data, how it's transformed, being able to govern those things is just a imperative. It could be a database thing too, it's one of those things where there's so many areas that could be mistakes on the data side. So I want to give you thoughts on the point you said earlier which I thought was about technology not coming out and getting commercialized or operationalized for a variety of reasons, one of them being processes in place, and we hear this a lot. This is a big opportunity because the human side of these new jobs, whether you're operating now, the network, really they need help, customers need help. So I think you guys should do a great job there given the history. The other trend that came out of the keynote today I want to get your reaction to is, and there's a tweet here, I'll read it. It says, GCP Cloud will start serving managing services enterprise workloads, including Oracle, RAC, and Oracle Exadata, and SAP HANA through partners. Interesting mind shift again, talk about a mind shift. Partners are used to dealing with multi-vendors, but now as a managed service, we'll change the mechanism a bit on delivery because now it's like, okay, hey, you want to slink some APIs around, no problem. You want to manage it, we've got Kubernetes and Istio. You want a little Oracle with a little bit of HANA? So it brings up a much more diverse landscape of solutions which makes the partners like sous chefs. You can cut the solutions up any way you want to your point about going faster to the next challenge. Is that going to be the new normal, this kind of managed service dashboarding? Do you see that as a- I think it is, and I'll take it a step further beyond managed service and actually get a little more discreet. One of the things that we're doing increasingly more of is insights as a service, right? So if you think about managed service in the traditional sense of I've got a process and you're going to manage that process end-to-end for me, that technology end-to-end for me, I do think that that's going to slowly become more and more prevalent. And that has to happen with our movement to putting our applications in the cloud and our ERPs in the cloud. And I think it is going to become more of the norm than the last, but I also think that it's opening the door for a lot of other things as a service, including insights as a service. So organizations can't find the data science talent that they need to do the really complex types of analysis. You know, just your insights as a service comment just gave me an insightful, original idea. Thank you very much. You're welcome. And I'll put this in the wrap-up, Dave, when we talk about it. But think about insight as a service to make that happen with all the underpinning tech, whether it's Oracle or whatever. The insights are an abstraction layer on top of that. So if the job is to create great experiences or insights, it should be independent of that. So Google Cloud is bringing out a lot more of the concept of abstractions, Kubernetes, Istio. So this notion of an abstraction layer is not just technical, there's also business logic involved. Absolutely. This is going to be a dream scenario for KPMG for your part, for your customers, for other partners. Because now you can add value in those abstraction layers. Absolutely. By reducing the complexity. Well, Oracle, that's not my department. That's Hanna, that's SAP. Who does that? He is, she's the product lead over there. Gone. Insights as a service, completely horizontally. Latin's that. And you know, to that point, there's magic that happens when you bring different data together. And having data silos because there are datas in different systems, just that's the analytics of 1990. And organizations can't operate on that anymore. And real analytics comes when you are working at a layer above the systems and working with the data that's coming from those systems. And in fact, even creating signals from the data. So not even using the data anymore, creating a signal from the data as an input to a model. So I couldn't agree with you more. Whole new way of doing business. This is digital transformation. This is the magic of cloud. Tracy, great to have you on. Yeah, thanks for having me. It's going to be a whole new landscape change over a new way to do business. You guys doing a great job. KPMG, Tracy, you guys are here inside theCUBE talking about analytics, AI. If you can't do analytics good, why even go to AI? I love that line. theCUBE bringing you all the data here. We'll stay with us for more after this short break.