>>Live from Las Vegas. It's the cube covering UI path forward Americas 2019 brought to you by UI path. >>Welcome back to the Bellagio in Las Vegas. Everybody, this is Dave Vellante and we're here day two of UI path forward three. The third North American event is the cubes, second year covering UI path. The rocket ship that is UI path. PDC is here, he's the vice president of AI at UI path and Rudy Coon who is the chief marketing officer and co founder of process gold UI path. Just announced this week, the acquisition of process gold. So Rudy, congratulations and you may as well PD. Thank you. So that's cool. Um, process gold is focused on process mining. You guys may or may not know about them, but really maybe, maybe you cofounded the company. Why did you co-found you and your founders process gold and tell us a little bit about the problems that you're solving. Yeah, right. You know, um, many years ago I started my career with IBM and I used to be a business consultant. >>And typically if you try to implement any kind of technology like RPA, but back then we didn't have the LPA. But if you try to figure out what the real process and the company are and you ask people, please tell me how does the process where it looks like. Usually people cannot tell you. They say yes we have a documentation but it's outdated the moment you print it. So the idea was um, actually I came across process mining more than 10 years ago and I met the guy in, at the university of and he had this bright idea to reconstruct business processes solely based on digital footprints from any kind of it system. I mean, think about it. You, you use SAP, you use any kind of other it systems and you take the data that is left behind after the execution or the support of a process. >>You take it, you push the magic button and you see what the process really is, like an extra races and from business processes. But we, we saw that in the demo at the a analyst event. I thought it was like magic. I mean I think it's actually, I think of a small company like ours easement even though the number of processes we have and the relative complexity and by the way, half the time people aren't following them and but you were able to visualize them. So. So first of all, why did you acquire process gold? What was the thinking there? So you know, just to pop one level up the stack, what exactly are we trying to do as a company? And you are about as we are building this whole new set of platform capabilities, right? We used to have product lines in studio, orchestra and robot, but now when we look at the whole customer journey and all the elements that need to be there in that customer journey, we essentially have to weld something, what I call the operating system called a self improving enterprise. >>And what that means is that our three elements you need to combine. You need to have a measurement system in place, which can quantify the ROI of your automations. Of course you need a really solid RPA platform like ours to do the automation itself, you have to be able to bring in pieces for doing complex stuff, cognitive stuff using AI. And then you need a scientific way of planning those automations using tools like process board because you have to do process mining. Once you complete this, watch your cycle, you can keep doing more and more of the automation. Essentially you're feeding the beast of efficiency in your organizations. So essentially the way this worked, we can't do, don't, don't have the means to do the demo here, but you essentially pointed your system at a process and it visually showed me the steps and laid them out and in great detail. >>Um, and I said, wow, that's like magic. Um, but this stuff actually works. You got no real customers using this if you do. Yeah. Okay. >> So you know, we worked for companies like, like portion Germany, maybe you have heard about them. They, they build cars and they are using process code for part of the production process. Today in today's world, every process, no matter how offensive is a physical process like production or purchasing or whatever it's used or it's supported by it and at least a lot of data behind. And this is exactly that, the goldmine for us. So we extract this data and again, you know, we have a lot of algorithms in the, in the software. It's, it's sort of magic as it is a lot of mathematics, which is magic for me. But um, it works. Yeah, just take the data, you pushed a button and just see the process with all the details. >>As you mentioned, like stupid times, bottlenecks, compliance issues and this three, the, the, the source, you know, if he wants to see the process, you can then decide is it, is this process now suitable for automation or maybe should we first optimize the process and then vote for automation. And this is key for, for RPA. >> Well, I think, you know, I'm talking a lot of customers this week and last year offline as well. A lot of times we'll tell us the mistakes they made is they'll, they'll automate a crappy process. Yup. This presumably allows me to sort of highlight the shine a light on some of the weaknesses and the weak links in the chain. >> So process optimization is a big deal, right? Both in the pre automation phase and in the post automation phase. Once you automated a process, you need to know what are the bad things that are happening there, what are the blockers, what are the nonconforming steps that you're taking? >>So that's in the post automation but also in the pre automation phase where you haven't even decided what exactly are you going to automate. It's really hard to quantify what are the high ROI processes, right? I can go in our bottle, automate something which is not useful at all for the users, right. And so we want our users to a wide making those mistakes. And that's why we are exposing these powerful, powerful set of tools where you can use all these tools to easily document your processes, manage your processes, use process mining to look deeper into how our people and the different entities in your organizations working together. You know, historically if you look at stuff like all of in all of human history, there have been certain processes, but as computers came on and stuff, you look at it on in, in scifi movies, everyone has always, as Rudy says, the X way for the enterprise. >>You always wanted to have this Uber system that can understand everything that we are doing and tell us, you know, how can we improve stuff? Or what can we do better? Because as a species that fuels our evolution. And so this is, it's, it's, it's fundamental to a lot of things that people do in every day and almost in every action that they did. >> So the in the secret sauce is math, right? So again, please, the secret sauce. Yeah, it's math, but you've got to have some kind of discovery engine as well. I mean this is, it's a system. So maybe can you give us a little bit more idea as to what's under the covers? Well, you know, it all starts with data and the data we need in the beginning, it's very, very simple. We need only three different attributes. The first attribute is what we call the case ID. >>So the case ID is a unique identifier for a case and it depends on the process. If we talk, for example, a very simple invoice approved process in the case that it would be the invoice number. When we talk about claims management or with a claims number or a purchase number, whatever the second attribute we need is the timestamp. And every time we find the timestamp in a system like SAP or lock file or database, this time subsume a timestamp actually represents some sort of activity. So we need a case ID, timestamp and activity and solely based on this data we can already show you how the process looks like. And then we enrich this data with other attributes like let's say supplier or invoice amount to give you some more ideas and some statistics. So this is the data we need. We, you know, we transformed this data, we access directly the database. >>So there is no, there's no need to extract the data. We directly access to data and we transform it and then it will be represented in our application. So you get rid of full transparency of what's going on. So when you were a consultant, you mentioned you're a consultant at IBM, you would sit down with a pen and paper and talk to people about what they did. Maybe time and motion studies and studies, you know, you know, this process mapping workshops, everybody comes out and just allows it. So you sit together with people in the room and at the end of the day you have more processes than you have people there. And everybody's telling you a different story and you know exactly that. Not everything is totally true. So a lot of gray area. Yeah. And the maps that you had to build and people simply don't know what the processes are. >>It's not that they don't want to tell you, they simply don't know. Or as I said before, different people have different processes and they don't follow those. There's no standard to follow. She's pretty, what's the vision for how, how process gold fits into UI path. So as a problem was talking about in his keynote, and Daniel talked about this too, um, a lot of our customers came to us, uh, to automate the processes that they already know about for the processes that they don't know about. We have this whole set of tools, the Explorer set of rules that we are releasing. Process world is a part of that. But essentially now you don't need to know what processes to automate. You can use an automated set of tools to do that process scored, as Rudy was talking about, can go in and look at these log files, uh, ordered logs that are generated by your systems of record. >>Um, and then be able to visualize, optimize our process. But the technologies are really complimentary because these guys, uh, used to work in the backend systems. That's why, you know, that's where most of the process mining works works in the back end looking at the audit logs, but you have as has, you know, we have really strong background in understanding the gooey in the front end, uh, understanding of apps, controls and the control flows that the users have using our computer vision technology. When you combine these technologies, there's a magical effect that happens. Like if your backend does not contain the audit, log off some actions that people are taking in the front end. Let's say it's a small application which does not generate that are the, once you combine these two data points, this is one of the first in the industry on the wonderful kind system that can look across all the different spectrum of applications and be able to understand the processes at a deeper level. >>Technically when you make an acquisition, you obviously looking at the technology and how it's going to integrate, how challenging will it be for you to integrate? What have you done any sort of, when you did the due diligence, you know, a lot of companies are really dogmatic about integration. Others frankly aren't that let's buy the company up by another one. What's your philosophy? It >>was kind of a match made in heaven. I remember the first time I talked to Rudy on the phone and uh, you know, are at the end of the day our philosophies aligned like almost a hundred percent because at the end of the day process goal and UI bad is all about that customer obsession, delivering the value to our customers. And the values are saying we want our customers to get out of this mundane tasks to automate the tasks as optimally as possible. And so both the companies, the, the, the outcomes aligned pretty well. Now the mechanics of the integration, um, I think both do. Both the companies are, these aren't you know, dot com era companies where you know, somebody came over the an idea and did this take Rudy and the team had been working in this area for 10 years. They have organization knowledge, they have the expertise and so does you have adults. >>And so we will take what I'm, what I call a loosely coupled approach where we can choose common customers, we can choose comments that are features that we are going to work on and that's how we will integrate. But again, the focus of all this is to deliver the value to our customers. Not think about the mechanics of what the integration would look like. I think one of the most exciting things that I'm hearing is this notion of the processes that are not known. Um, because so many processes today are unknown, especially as we go into this new digital world. We used to know what processes we want to automate your point, some technology at it. Okay great. We're going to automate now with this digital disruption that's going on. You actually may have no idea. You may be making processes up on the fly, so you need a way to identify those processes quickly and then those ones that are driving our ROI. >>Um, I'm interested in your thoughts on AI and ROI and how to measure that, how those things fit together. So, you know, AI, this is I think the biggest problem in the AI right now. There's a lot of hype in this space. We are tracking close to 3000 different AI startups in the world and uh, nobody can actually put a number to the revenues or the valuation, the real valuation because of this ROI quantification problem, right? Um, let's say I have a company, we'd say, Oh, we are the best in class. And understanding faces short, how is it going to be useful to an enterprise if you cannot measure what well you official recognition system is adding to your enterprise, it's not good enough for the business people. Because at the end of the day, my, I can have the world's brightest PhDs telling me I have the state of the art model in the world, which does law, but in fact cannot translate it into business value. >>It doesn't really work. And so that's why ROI quantification is so in parking and you have to make sure you align them econometrics of the AI, uh, measures and the business KPIs so that if, for example, so your data science team should be able to know what metrics they have to improve in order to get a better ROI for the business. So you have to align those two things. And that is part of research that is not really prevalent in academic circles. Interesting. I mean, you've seen some narrow successes in I'll call AI, you know, things like a infrastructure optimization. Okay, great. Makes sense. What I'm hearing from you is identify the KPIs that are going to drive your voice of the customer defines value first to take away, identify what those KPIs are. And this every business has thousands of KPIs, but there's really like three or four that matter, right? >>So identify those top ones and then you're saying measure on a continuous basis how your system affects those metrics. So in economics this is called the treatment effect. Uh, so for example, if you water my term sales and marketing processes, the KPIs that matter to you is what is your conversion rate from when the leads hit your system to when the revenue is realized or what is the total revenue that you're making? Right? As you said, there's only two or three top level gave you as that really matter. And now if for example you put an AI system in place that treats your leads differently, you should see an increase and uptick in revenue. And so that's what I mean by the Ottawa quantification. So if you instrumented the system properly, put it in the right quantification measurement system in place and have the auto optimization mechanism, that's how things should work. >>You know, with with cross mining we can even add additional KPIs to the picture KPIs you usually don't have because if you ask a company, nobody can tell you how many different variations of the process you actually have. And with process mining we can exactly measure how many variations there are. So if you are up to streamlining to simplifying the process to speed it up, we can actually tell you if your optimization effort is successful or not because we can show you how the number of very our variations is going down over time. Even if we, you know, we can also measure the, the success of RPA implementation. So it really pros we use process code and pro money not only for identification of processes but also for the monitoring of processes after an successful RPA implementation. I can see so many use cases for this. >>I mean it's like my mind is just racing. I mean sales guys in one region and sales gals in the other region doing things differently. You've got different country management doing things differently. If I understand you correctly, you can identify the differences in those processes, document them, visualize them and identify the ones that are actually optimized or help people optimize and then standardized across the organization to drive those metrics that matter. It's very powerful. It is really powerful. You know, as I said, we are living in the golden age of this system that can self-improve your companies. I mean this, this was the Holy grail of all of computer science work with technologies like process score with RPA, with AI. I think we are at that inflection point where we can realize that. So we got to go. But I'll, I'll give you guys sort of the last, last word, each of you. >>So actually first of all, Rudy question, how large can you tell me how large the process gold team is? How many people? We have grown with 60 people. 60 equals zero. We are based, our headquarter is in the, is in the, in from the Netherlands. Um, so this is where we are very close to university. This is where our developers basically are located. And uh, I'm based in Frankfurt in Germany, but for now, let's see what the future will be. So what's a home run for you with this marriage? The home run, you know, since we are in Las Vegas, I was wondering if you hit the jet park Jack photo, if we hit the jackpot. But I actually think of the customers, our customers get the Jaguar because this combination of, of your technology, of our technology, this is really, you know, good answer. So that as I was gonna ask you the same question PD is, I can't even tell you, um, almost every one of the UI path customers has expressed interest in process glow, right? >>Because right now we have a portfolio of products, but the interest that we are getting in process board with the process mining offerings is unparalleled. So Rudy is right. Our customers are the ones which are driving this inhibition and the integration. And I'll be able to actually acquire this solution. I forget, I have my notes with relatively near term, right? Yes. We are gonna make it available to our customers as soon as possible. Awesome guys, congratulations. Really great to have you on the cube. Thank you. All right, and thank you everybody for watching. We'll be back with our next guest right after this short break. You're watching the cube alive from the Bellagio UI path forward three. We were right back.

>> Announcer: Live from Las Vegas, it's The Cube. Covering AWS re:Invent, 2017. Presented by AWS, Intel, and our ecosystem of partners. >> Welcome back to Las Vegas. We are live here on The Cube continuing our coverage of re:Invent. The AWS, the big tent. As we were just talking about with our guest, Justin Moore and John Walls here. Your hosts here on The Cube and we're joined by Kiran Bhageshpur, who's the CEO of Igneous Systems and Kiran, thanks for being with us here on the Cube. Good to see you. >> Great to be here. >> Now we were talking about, you know, this is the big tent now. Didn't used to be that way, right? >> Nope, nope. >> It wasn't that long ago this was, I wouldn't say a specialty show, but you said this has certainly taken on a very different vibe, a very different feel. I mean, explain that a little bit before we get into Igneous and what you're doing here. >> Absolutely. I was first here in 2012, I believe it was the first year they had AWS at re:Invent and it was a very different feel, much smaller, maybe about 6,000 or so people. Mostly engineers, hardcore engineers who were discovering this new cool set of toys, if you will, or tools that was quite revolutionary and niche at that time. Fast forward now. It's much more of a mainstream show. It's much more corporate IT, lots and lots of large enterprises are present out here. There still is a lot of developers, but it's more the devops, more people who are operationalizing this rather than building on it for the very first time. So big change from early stage to very mainstream right now. >> And Justin, you made a comment. I mean, to the extent of a jacket, I've got a suit and tie, a jacket. We've all been to shows where maybe the wardrobe was maybe a little different, but this is illustrative of, again, of the maturation of the marketplace and expansion of the marketplace. >> Yeah, you go to some of the developer conferences and you see a lot more people with spiked purple hair and then utilikilts. I've yet to see a single utilikilt here at the show so it does feel, unlike at previous years where there's been, again, a lot more engineers and people are still here in hoodies and casual clothes, but there are a lot more suits. There's clearly a lot more money here and it's become a little more corporate. It'd be interesting to see how it transitions over the next couple of years whether Amazon or AWS is able to maintain that kind of developer vibe as all of these other companies come in and start to see actually, this is a pretty robust and mature ecosystem now. >> Yeah. >> Yeah. And obviously, the expansion reflects that. You're here exhibiting for the first time. >> Yes we are. >> Your booth back at K37 if you're here at the show. Kudos to Igneous. Let's talk a little bit about what you do and why are you here? Who are you trying to talk to this weekend and why does this week matter? >> That's great. So what we do, Igneous is an early stage company. We have launched our company a year ago. We have a bunch of customers right now sort of growing very nicely at this stage and what we do is enable businesses, enterprises with lots and lots of file data on premises as well as in the public cloud to better manage and have a handle on this. So our customers tend to be businesses with sort of literally billions of files, hundreds of petabyte or dozens of petabytes spread across a lot of systems traditional, legacy that hook, attach to a system on premises and what they are seeing in their growth is they're going from one data center to multiple data centers within their own infrastructure and now to multiple clouds and as this core asset, data continues to grow. They look to folks like us to help manage that better. So the very first thing we do is we enable them to back up and protect all this data on premises into public clouds like AWS so we literally have scalable solutions which go into their data center, talk to all of their filers as they're called, interrogate all of that data, and create a copy of that into AWS S30 glacier. >> Yeah. There's a lot of companies who are struggling with the idea. Two things really. One is being able to manage data everywhere because data has gravity as people like to say, but also this multi-cloud idea and being able to manage my data in multiple physical locations. Some of it will be on my own site. Some of it will be in Collo. Some of it will be in one or as you say multiple clouds. That really hybrid IT way of things. What are you seeing as the driver behind that need to have this data in multiple locations? >> Yeah, that's a great question. For the things that we see is, look, things have remained on premises. It's not gone away and things continue to grow on premises and Amazon recognizes that. That's why you see starting last year into this year a lot more push into hybrid clouds, if you will. You saw that with the big partnership with Vmware and so on. So that's continuing to grow, but in the same time, they're having new applications being born in the cloud or leveraging the cloud. So one thing which is very common for a lot of our customers is they have infrastructure on premises which is already paid for and continues to grow, but they want to leverage the public clouds, AWS, for its elasticity and its agility to be able to burst into it and use it as they see fit. Now to do that, you require agility of applications and data between on premises and the public clouds and say AWS. So that's kind of where, you know, we come in to go help them in that and the other thing we're also seeing is customers are not in a single cloud. Even if they started in one place, they're starting to exist in multiple different locations. Good example will be in, you know, most of our customers tell us that, say, a Google cloud has the advantage for things like AI and machine learning whereas Amazon has the more mature infrastructure. So they might quite have a lot of infrastructure and data on premises as well as on Amazon, but they might be running a bunch of new applications which are leveraging the machine-learning APIs and Google Cloud. But then how do you get the data from on premises Amazon cloud into Google Cloud, use it but not leave it around and triple pay for it all around so that's really the management challenge. >> Yeah so you mentioned a particular use case there that happened to use Google. So AI and machine learning is something and I'm hearing that in talking to customers myself that they like to use different cloud for different reasons. So what are some of the workloads that you're seeing from customers who are needing to put their data not just on site, but they say, you know what, I want to burst into the cloud, I want to use some of that elasticity that cloud is so great at. What are some of the workloads that you're seeing them use your product for? >> Yeah, I'll give you a great example. Let's take the word of the movie world, right? So lots of it is all digital right now. The data is created and you're gonna go create heavily CGI or computer generated effects using lots and lots of computer cores. What you come to is at the end of the movie, there's a crunch time where they need way more compute than they have available within their data centers. In fact, in the past, there used to be a vibrant side business where little boutique companies would rent you servers and they would literally carve that into your data center for six weeks and take it away again so now that's gone and you'd rather use the public cloud, you use Amazon and EC2 instances for that workload. That's a good example which everybody can relate to. Hey, it's crunch time, movie's coming up for release. I have a lot more work to do, but that pattern exists in pretty much every industry whether it's drug discovery or electronic design. Everywhere, there is a need to grow burst beyond what you have available and that kind of drives the adoption for workflows which already exist on premise to also adopt a cloud. >> Yeah. >> You got it. >> What manageability. I mean, talking about multi-cloud. >> Kiran: Yeah. >> And obviously as you parcel out your assets, you decide what data's gonna reside in what environment managing all that and then managing the cost of all that. I mean how do you keep up corale on that and also help your clients get a handle on where their data's going, 'cause yeah. I don't know, right? >> So that's what we exist to do which is help customers manage this data asset that they have across multiple locations no matter where it lives. The first thing we do in our journey with our customer is just back that stuff up which is all on premises into the cloud so it gets a copy of the dat into the public cloud. Now that enables workflows like being able to use the cloud for disaster recovery or use the clouds for burst computing very well. But it's just beyond that. It's also how do you get the data, where it lives, which could be on premise, on a T01 filer to where it ends to be. Perhaps the public cloud for a back-up deal or a burst-use case or perhaps into a separate cloud for using machine learning and when you do this, how do you ensure you have one copy, one protected copy of the data, not three or four every place? In fact, if you look at the world today on premises, already customers will tell us they have hundreds of systems that it's not infrequent that hey, they have infrastructure say in Santa Clara as well as Israel and it's a same copy which exists in both places because they have no way of globally looking at this in one single way. >> That's kind of what we do is hey, what are your data assets, where do they live, how do we ensure you have one copy of it or n copies as you desire but not a proliferation of that dataset, three how do we get the data from where it lives to where it's needed in a programatic, systematic way that your end user can sort of you know, help themselves too rather than requiring an IT trouble ticket and somebody going through a manual process. So those are sort of good sets of early things we are helping customers out with. The other thing that goes into here and this is where the cloud comes in again is we had targeted customers who are looking at literally billions, tens of billions of files, hundreds of petabytes, tens of petabytes to 100 of petabytes of data spread across many locations and many hundreds of systems. How do you get your hand, your head around that? It's beyond human scale and it's only possible with software and sort of machine learning if you want to use the buzzword and that's the sort of next place where you come in and provide a human comprehensible structure for the sort of data which continues to grow and it's important because this is core assets for businesses today. >> Yeah we were discussing this earlier, both of us, actually. It's that idea of automation because humans don't scale. >> Yeah. >> So when you have these billions of files as you're talking about, that's just not trackable for humans to deal with. So what are some of the automation and autonomous systems capabilities that Igneous has? >> So the first thing we do is go ahead and ensure your automatically and at scale, being able to discover all of that data, right? So think of, you know, if you look in the consumer world, really what the web is goes and crawls every website and indexes all of the data. Well we do that except within the enterprise for the unstructured file data, which happens to live on a net app filer or a Dell cluster or maybe it's living in AWS inside S3 where we go crawl all over that, index, all of that and give you a view into that. That's the first level, simple way of doing that. But then the next level beyond that is if you can give a level of structure on that because it's not useful to just find it. You wanna know what you have or where you have it, how it's changing, who is accessing it, what applications are accessing your data? What applications are modifying your data? Today, that is an extremely manual process within businesses. >> Yeah. In order to make sense of that, again, you're trying to appeal to developers. What APIs and sort of programmatic aspect do you have for that rather than having to employ 1,900 humans who would all have to sit there and drive around with, going through interfaces? >> So since our customers tend to be sort of more on the business side of IT today who are trying to go understand about this data, the interfaces we provide them is clearly the higher level abstraction of what the data looks like of how they want to interact with that, but everything you do in the modern world is API enabled and the vision is clearly to go expose all of this through API such that customers, developers within their organization can go consume it. >> So before we let you go, I want to talk about your presence here, the decision to exhibit. It's not a light one, I know that. At the end of the day, when you walk out of here on Thursday, what do you want to accomplish and I guess from the, in terms of the kinds of audience that you're hoping to be exposed to, who would that be? >> So the customers, the prospects we talk to are typically businesses, enterprises with lots and lots of unstructured data so people in the media world, in the design world, any form of design that is electronic and automated to. You know, geospatial imagining. All of these folks and they are all present here this year. This is the show to be, you know. In the past, it used to be Microsoft PDC, it was RealWorld, it was Oracle World. Today it is AWS re:Invent and they're all here and for us, it's success if we walk out of this being exposed to a whole bunch of people. We as a smaller organization could not have had immediate access to without coming to this show and that's what I think we get out of here. >> Well, good luck on the next three days. It sounds like you're off to a great start in the right place at the right time. >> Yes, indeed. >> And we wish you all the best down the road. >> Thank you. >> Kiran thank you for being here. >> Thank you very much. >> Live on the Cube, you're watching us here at re:Invent where it's AWS's big show here in Las Vegas back with more live coverage in just a moment. (energetic music)