 Live from Orlando, Florida, it's theCUBE. Covering Microsoft Ignite. Brought to you by Cohesity and theCUBE's ecosystem partners. Welcome back everyone to theCUBE's live coverage of Microsoft Ignite. I'm your host, Rebecca Knight, along with my co-host, Stu Miniman. We are joined by Mike Flasco. He is the principal group product manager here at Microsoft. Thanks so much for returning to theCUBE. You are a CUBE alum, Mike. I am, yeah, thanks for having me back. I appreciate it. So you oversee a portfolio of products. Can you let our viewers know what are you working on right now? Sure, yeah. I work in the area of data integration and governance at Microsoft. So everything around data integration, data acquisition, transformation, and then pushing into the governance angles of, once you acquire data and analyze it, are you handling it properly as per industry guidelines or enterprise initiatives as you might have? You mentioned the magic word transformation. I would love to have you define, it's become a real buzzword in this industry. How do you define digital transformation? Sure, I think it's a great discussion because we're talking about this all the time, but what does that really mean? And for us, the way I see it is starting to make more and more data-driven decisions all the time. And so it's not like a light switch where you weren't and then you were. Typically what happens is, as we start working with customers, they find new and interesting ways to use more data to help them make a more informed decision. And it starts from a big project or a small project, and then just kind of takes off throughout the company. And so really, I think it boils down to using more data and having that guide a lot of the decisions you're making. And typically that starts with tapping into a bunch of data that you may already have that just hasn't been part of your kind of traditional data warehousing or BI loop and thinking about how you can do that. Mike, bring us inside the portfolio a little bit. Everybody knows Microsoft, we think about our daily usage of all the Microsoft product that my business data runs through, but when you talk about your products, they're specific around the data. They help us walk through that a little bit. Sure, yeah. So we have a few kind of flagship products in the space, if you will. The first is something called Azure Data Factory. And the purpose of that product is fairly simple. It's really for data professionals. They might be integrators or warehousing professionals, et cetera. And it's to facilitate making it really easy to acquire data from wherever it is. Your business data on-prem from other clouds, SaaS applications, and allow a really easy experience to kind of bring data into your cloud, into our cloud for analytics, and then build data processing pipelines that take that raw data and transform it into something useful whatever your business domain requires, whether that's training a machine learning model or populating your warehouse based on more data than you've had before. So first one, Data Factory, all about data integration, kind of a modern take on it, built for the cloud, but fundamentally supports hybrid scenarios. And then other products we've got are things like Azure Data Catalog, which are more in the realm of aiding the discovery and governance of data. So once you start acquiring all this data and using it more productively, you start to have a lot. And how do you connect those who want to consume data with the data professionals or data scientists that are producing these rich data sets? So how do you connect your information workers with your data scientists or your data engineers that are producing data sets? Data catalog's kind of the glue between the two. I wonder if you could help connect the dots to some of the waves we've been seeing. There was additional kind of BI and data warehousing, and we went through a kind of big data, the volumes of data and how can I, even if I'm not some multinational or global company, take advantage of our data, now there's machine intelligence, machine learning, AI, and all these pieces. What's the same and what's different about the trend and the products today? Sure. I think the first thing that I've learned through this process and being in our data space for a while and working on our big data projects is that for a while we used to talk about them as different things. Like you do data warehousing and now that kind of has an old connotation feeling to it. It's kind of old feel to it, right? And then we talk about big data and you have a big data project. And I think the realization that we've got is it's really those two things starting to come together. And if you think about it, everybody has been doing some form of analytics and warehousing for a while. And if we start to think about what the big data technologies has brought is a couple of things in my opinion that kind of bring these two things together is with big data we started to be able to acquire data of significantly larger size and varying shape, right? But at the end of the day, the task is often acquire that data, shape that data into something useful and then connect it up to our business decision makers that need to leverage that data from a day-to-day basis. We've been doing that process in warehousing forever. It's really about how easily can we marry big data processing with the traditional data warehousing processes so that our warehouses, our decision making can kind of scale to large data in different shapes of data. And so probably what you'll see actually at Ignite Conference in a lot of our sessions, you'll hear our speakers talking about something called modern data warehousing. And it really doesn't matter what the label is associated with it, but it's really about how do you use big data technologies like Spark and Databricks naturally alongside warehousing technologies and integration technologies. So they really form the modern data warehouse that does naturally handle big data, that does naturally bring in data of all shapes and sizes and provides kind of an experimentation ground as well for data science. I think that's the last one that kind of comes in is once you've got big data and warehousing kind of working together to expand your analytics beyond kind of traditional approaches, the next is opening up some of that data earlier in its life cycle for experimentation by data science. It's kind of the new angle. And we think about this notion of kind of modern data warehousing as almost one thing supporting them all going forward. I think the challenge we've had is when we tried to separate these into kind of net new deliverables, net new projects where we're starting to kind of bifurcate if you will the data platform to some degree. And things were getting a little too complex. And so I think what we're seeing is that people are learning what these tools are good at and what they're not good at and now how to bring them together to really get back some of the productivity that we've had in the past. I want to ask you about those business decision makers that you referenced. I mean, there's an assumption that every organization wants to become more data driven. And I think that most companies would probably say yes, but then there's another set of managers who really want to go by their gut. I mean, have you found that being a conflict in terms of how you are positioning the products and services? Yeah, absolutely. In a number of customer engagements we've had where you start to bring in more data, you start to evolve kind of the analytics practice, there is a lot of resistance at times that we've done it this way for 20 years, business is pretty good. What are we really fixing here? So what we've found is the best path through this and in a lot of cases the required path has been show people the art of the possible, run experiments, show them side by side examples, and typically with that comes a comfort level in what's possible. Sometimes it exposes new capabilities and options, sometimes it also shows that there's some other ways to arrive at decisions, but we've certainly seen that almost like anything, you kind of have to start small, create a proving ground and be able to do it in a kind of side by side manner to show comparison as we go, but it's a conversation that I think is going to carry forward for the next little while, especially as some of the work in AI and machine learning is starting to make its way into business critical settings, right? Pricing your products, product placement, all of this stuff that directly affects bottom lines, you're starting to see these models do a really good job and I think what we found is it's all about experimentation. Mike, when we listen to Satya Nadella and talk about how things are developed inside Microsoft, usually hear things like open and extensible, you got to have APIs in any of these modern pieces. It was highlighted in the keynote on Monday talking about the open data initiative, got companies like Adobe and SAP out there and they have a lot of data. So the question is, of course, Microsoft has a lot of data that customers flow through, but there's also this very large ecosystem we see at this show. What's the philosophy, is it just, oh, I've got some APIs and people plug into it? How does all the data get so that the customers can use it? Yeah, it's a great question. That one I work a lot on, which is, and I think there's a couple of angles to it. One is, I think as big data's taken off, a lot of the integration technology that we've used in the past really wasn't made for this era. Where you've got data coming from everywhere, it's different shapes and it's different sizes. And so, at least within some of our products, we've been investing a lot into, how do we make it really easy to acquire all the data you need? Because, you know, like you hear in all these cases, you can have the best model in the world if you don't have the best data sets, it doesn't matter. Digital transformation starts with getting access to more data than you had before. And so, I think we've been really focused on this, we call it the ingestion of data, being able to really easily connect and acquire all of the data. And that's the starting point. The next thing that we've seen from companies have kind of gone down that journey with us is once you've acquired it all, you quickly have to understand it. And you have to be able to kind of search over it and understand it through the lens of potentially business terms. If you're a business user trying to understand what is all these data sets, what do they mean? And so, I think this is where you're starting to see the rise of data cataloging initiatives. Not necessarily master data, et cetera, of the past, but this idea of, wow, I'm acquiring all of this data. How do I make sense of it? How do I catalog it? How does all of my workers and my employees easily find what they need and search for the data through the lens that makes sense to them? Data scientists are going to search through a very technical lens. You know, your business users through, you know, business glossary, business domain terms in that way. And so, for me, it all starts with the acquisition. I think it's still far too hard and then becomes kind of a cataloging initiative. And then the last step is how do we start to get some form of standards or agreement around the semantics of the data itself? Like, you know, this is a customer, this is a place, this is what, you know, a rating. And I think with that, you're going to start to see a whole ecosystem of apps start to develop. And one of the things that we're pretty excited about with the Open Data Partnerships is how can we bring in data and to some degree auto-classify it into a set of terms that allow you to just get on with the business logic as opposed to spend all the time in the acquisition phase that most companies do today. You mentioned that AI is becoming increasingly important in mission critical or at least bottom line critical in business models. What are some of the most exciting new uses of AI that you're seeing and that you hope expands into the larger industry? Sure. It really does cross a number of domains. We work with a retailer, ASOS, who, you know, every time we get to chat with them it's a very interesting use on how they have completely customized the shopping experience. From how they lay out the page based on your interest and preference through to how the search terms come back based on seasonality of what you're looking at based on what they've learned about your purchase patterns over time, your sex, et cetera. And so I think this notion of intensely customized customer experiences is playing out everywhere. We've seen it on the other side in engine design and preventative maintenance where we've got certain customers now that are selling engine hours as opposed to engines themselves. And so if there's an engine hour that they can't provide that's a big deal and so they want to get ahead of any maintenance issue they can and they're using models to predict when a particular maintenance event is going to be required and getting ahead of that through to athletes and injury prevention. We're now seeing all the way down to connected clothing and athletic gear where all the way down not just at the professional level but it's starting to come down to the club level on athletes as they're playing starting to realize that something's not quite right. I want to get ahead of this before I have a more serious injury. And so we've seen it in a number of domains almost every new customer I'm talking with. I'm excited by what they're doing in this area. You bring up an interesting challenge is I've heard Microsoft is really, I guess, verticalizing around certain industries to put solutions together. One of the challenges we saw when they saw surveys of big data, the number one use case came back was always custom and it was like, oh, okay, well, how do I templatize and allow hundreds of customers to do this? Not every single project is some massive engagement. What are you seeing that we're learning from the past and it feels like we're getting over that hump a little bit faster now than we were a few years ago? Yeah, so if I heard you correctly, it's a little bit loud, but they're saying everything started at custom and how do we get past that? And I think it actually goes back to what we were talking about earlier with this notion of a common understanding of data. Because what was happening is everybody thought they had bespoke data or we had data that was speaking about the same domains and terms, but we didn't agree on anything. So we spent a ton of time in the bespoke or custom arena of integrating, cleaning, transforming before we could even get to model building or before we could get to any kind of innovation on the data itself. And so I think one of the things is realizing that in a lot of these domains, we're trying to solve similar problems. We all have similar data. The more we can get to a common understanding of the data that we have, the more you can see higher level reusable components being built saying, ah, I know how to work on customer data. I know how to work on sales data. I know how to work on, you know, oil and gas data, whatever it might be, you'll probably start to see things come up in industry verticals as well. And I think it's that motion. Like we had the same problem years ago when we talked about log files. Before there was logging standards, there was everything was a custom solution, right? Now we have very rich solutions for understanding IT infrastructure, et cetera, that usually became because we had a better baseline for the understanding of the data we had. Great, Mike, thank you so much for coming on theCUBE. It was a pleasure having you. Thank you for having me. I'm Rebecca Knight for Stu Miniman. We will have more of the CUBE's live coverage of Microsoft Ignite coming up just after this.