>> Voiceover: Live from Las Vegas. It's theCUBE covering UiPath Forward Americas 2019. Brought to you by UiPath. >> Welcome back to the Bellagio, everybody. You are watching theCUBE, the leader in live tech coverage, this is Day 2 of UiPath's Forward III Conference and Kourtney Bradbeary is here R&D specialist at American Fidelity. She's joined by Michael Setticasi, who's the senior director of business development at Boston-based DataRobot, but Michael's from Seattle. Guys, welcome to theCUBE. >> Kourtney B.: Thank you. >> Kourtney, let's start with you. I know you guys, you kind of do benefit solutions, but maybe talk a little a bit about the company and some of the big trends that are driving what you guys are doing. >> Kourtney B: Absolutely. So I work with American Fidelity, it's an insurance company based out of Oklahoma, but our main focus is providing solutions to our customer pain points. So we're a niche-based organization that focuses mainly on education, so the public sector, so education in municipalities in providing solutions and benefits to our employers and our employees that we work with. >> Cool, and Michael, you guys, obviously data science is your thing, but describe a little bit more about what you guys do. >> Yeah, we're an AI enterprise company. What we're really trying to do is democratize the use of AI machine learning within organizations, and we really appeal to both data scientists and business users that understand their business and data and want to do more. >> So Kourtney, you're title is really interesting. R&D special projects, so you got this little sandbox that you get to play with, RPA is on the hype cycle and now it's in the trough of disillusionment, but it's kind of an early play around with things. How did you get in to RPA? Where you guys at? What's this R&D thing going on? Right, so with research and development, I guess there's a lot of space to work with emerging technologies, and AI, and RPA, and how those two things come together and anything new that we see and exciting we're able to apply that technology. It's one thing to think, "Oh, AI, that's cool. Let's do that." But if it doesn't benefit your customer at the end of the day, if it's not driving decisions in your organization, then we don't want to do AI just 'cause it's cool. We really want to do AI because it's what benefits our customer. So we got into RPA because when we saw a demo, and it was like, whoa. If that's real, if that's what we think it's going to be, that's a game changer. So you have RPA, and you have AI kind of coming up at the same time and whenever it was, first coming out a few years ago, they're silo, they're separate. What we've started to do recently is to bring the two industries together and really bring together the RPA component and the AI component to really become IPA, or Intelligent Process Automation, so that way we can really start to transform businesses. >> So this is interesting to me, Michael, because as Kourtney was saying, most people think of these things as separate and more aspirational down the road. You guys are AI experts, what are you seeing in terms of these two domains coming together? >> You hear about intelligent automation everywhere, right? We are pushing it hard, and we're seeing a lot of customers and potential prospects look at it, but I have to give credit to American Fidelity. They are ahead of the curve. They're combining this ability to use an RPA process and a machine learning model to really automate things and provide better customer service and get to the endpoint faster and more efficiently. So I think they're ahead of the curve, but you're going to see more and more of this in the marketplace. >> So Kourtney, a lot of the customers that we talk to, this is kind of my observation, is they're automating obviously mundane processes but frankly really crappy processes. They're really screwed up in a lot of ways. And they're throwing RPA at the problem, it sounds like you have a little different philosophy around how to apply automation. Can you explain that? >> Right, so you don't want to automate something that's bad because then it's going to break a lot, and it's just not a good idea. So what we've tried to do is whenever we get request in the door, there's always a stopping, if somebody has to make a decision, in the past, it's been "Okay, well we can automate the first part and the last part", but it's kind of have to stop in the middle for you to make a decision. And what DataRobot has allowed us to do, in the past, it was really hard to actually apply machine learning, 'cause you had to have these data scientists and they'd have to spend months trying to figure out what model for the data, and is it, you know, retraining a model is really difficult. DataRobot makes a data scientist's job so much easier and actually applicable to the workplace where you could scale, enable scaling, because without DataRobot or without a service like that, it's impossible to scale. So it allows us to implement AI with our RPA to then not just automate the mundane processes, but the small decisions that we make everyday, just 'cause we do our jobs everyday and we know how to do our jobs, AI enables us to automate those processes, as well. >> And you're doing that in an unattended way, or is it an attended automation? >> Both, both. So there's some processes that we have to have a human select things and make certain decisions along the way, or there's some processes that are completely unattended. With any automation, your goal is always to automate 100%, but in reality, you're usually going to get about 80% of a process automated. So what we try to do, we go for the hundred percent, rarely get that, but then you can take out the 20% for human review. And so maybe of the 20% that's not fully automated, maybe we can make stop points for human interaction there, but there have been some processes that we have been able to fully automate. >> So Michael, the data scientists complain that 80% of their time is spent in wrangling data and getting the data ready to actually build a model. I presume that's what you guys do, you solve that problem, right? >> We definitely solve some of that, right? If you get the data all in one place, DataRobot takes care of a lot of the data preparation that's involved in data science. We've also have ways to kind of manage the best places you store your data, so that if other people use the platform, they can see where to get it to. But overall, I would just say, when you look at UiPath and the way it's growing, it's such an exciting growing company like we heard Daniel yesterday mention their growth from customer from year to year, how they're the fastest enterprise software growing company out there. So you combine that RPA market with this growing machine learning market, and there's a ton of excitement. I mean, that's what you're seeing at the conference today. >> So you guys have data scientists on staff, is that right, or-- >> Correct! >> Okay, and so what does this mean for them? Does it mean you just need less of them, or they spend more of their time doing productive work? >> It means they spend more of their time doing productive work, instead of trying to figure out what model to fit, 'cause if you're a data scientist, or an actuary, or any, data analyst, or any of those things, you might know five models that you try to fit everything to. What DataRobot enables us to do is not be stuck to those five models that we know. It enables us to combine models, and choose models based on that data, so it really helps us with the modeling. >> Are you, I should've asked this before, are you still in R&D? Or are you in production? Or where are you at in terms of majority? >> Oh no, we're in production. We have two IPA processes in production today, and we're working on increasing that as we go. We have over a hundred an fifty RPA processes in production, as well as, many many just machine learning, so we're working on combining those now. So we have many machine learning, we have many RPA, and we're working on increasing our IPA. >> What have you seen as the business impact? Do we have enough data yet to sort of-- >> Absolutely. We don't try to focus on ROI. What we try to focus on is how is this impacting our customer, and how is this impacting employees' lives. There's obviously a lot of fear around automation but at American Fidelity, what we try to do is show how this is going to improve our employees' lives and we're by no means trying to cut jobs. We're actually going to have a net increase of jobs over the next five years. We're re skilling our workforce. We're really focusing on how it improves our employees, rather than focusing on ROI. >> So you're not on the ROI treadmill? So how did you get your CFO to sort of agree to all of this? >> So we do track ROI. It's not something we share publicly. But we focus more on our humans and our employees than our ROI. >> Is that because, I mean you're not, virtually every customer I've talked to says, "Well, we're not firing people. We're just getting more productive, or shifting them to more interesting tasks, et cetera, et cetera," and if you do the ROI calculations, you say "Oh, I don't need as many humans to do this anymore", and so you'd say, "Okay, FTE cost" and then you apply that, it's kind of a BS number, 'cause it's not like you're cutting people, so it's not a hard ROI. Is that why you don't focus on ROI? Or you just think it's worthless metric? >> No... >> Actually, I'm sorry. You said you do have it, you just don't share it publicly. >> Right, we just don't share our ROI publicly. And I don't think it's made up, or it's fake. I've never met an organization that says they have more people than they have work for people. There's always work. I really enjoy the first video opening of UiPath, it's, "since the beginning of time, humans have worked", and everyone thinks that automation is going to get rid of jobs, there's a lot of controversy over that, but realistically, if you think about the first industrial revolution, that was, after the first industrial revolution hit, that was the biggest economic upturn that had seen since that time. We're in that same space now. It's just hard to see it with where we're at. It's only going to increase, work is only going to increase. It's definitely going to change. I think it's naive to think that jobs won't change. And there will be jobs that will be eliminated, job functions, but I don't think there's elimination of humans needed, if that makes sense. >> Well yeah, it does. You guys sound like you're pretty visionary about how to apply technology to your business. And Michael, I mean, Kourtney's right, machines have always replaced humans, this is nothing new, first time ever that it's in cognitive function, so that scares people a little bit, but what else are you seeing in the marketplace that you can share with us? >> We're just seeing increased use of automation. So like, you might think when you talk DataRobot, you're using us for the top 1% things that a company might do, right? If you're a bank, you might use us to help out, figure out, how you can more efficiently lend customer's money, and make sure that you're making good investments, but what we're finding is, automation and machine learning models are being used everywhere. They're being used in marketing now, right? An example could be this show. We'll get leads from this show. Let's run some machine learning to understand what leads to follow up on first, because we'll get the best result. We're seeing machine learning in HR, right? Making sure their employees are happy, tracking employee churn through machine learning, so I think what we're seeing is it's being adopted more broadly, which means you need more people. We're not replacing people. >> So, why UiPath? >> Whenever we started the vendor process and started looking at several vendors, the UiPath product just was unmatched, frankly. There was a lot of vendors that had more code base, and there was then UiPath that anyone can learn. And that's what we really liked 'cause in American Fidelity, we've chosen to go with, we have a COE but we've also chosen to go with a democratized model where everyone in the organization will be able to build robots. We're training people to build robots. We have, each department has people that are dedicated. A certain portion of their time is building robots and UiPath really made that available with their products for anybody to be able to learn. >> So you have a COE. >> Kourtney B.: Yes. It was interesting, Craig LeClaire this morning, I don't know if you saw his keynote, but he kind of made this statement, it was sort of a off-handed statement, he said, "COE, maybe that's asking too much". He didn't use term tiger team, but I inferred, it's like, rather just kind of get a tiger team of some experts, but talk a little bit more about your COE. >> So, we kind of go with a hybrid model. If you think about, typical, it's weird because RPA is only a few years old, and we're thinking typical RPA, but people usually either go with a COE or completely democratized. We've really gone with a hybrid model, so we have a COE with governance where we've set a loose framework of what to follow, and we have code standards, when you say, follow these things. We have a knowledge library that we share. But we only have a handful of full-time RPA developers, and everyone help, those developers help, teach and help grow that knowledge throughout the organization, so that way we have people in every area that can also develop. So our developers are not our own key developers. Our developers are focused on the IPA, on the AI, whereas our other people throughout the organization are focused more on RPA so we can really make a big difference more quickly. >> Do you have a software robot that automates auditing and checks for compliance? >> Yes, so we have, one of our robots, the function that it does is audit one of our inputs, so we do have robots in almost every area that, yeah, we do have audit robots. >> Has it cut the auditing bill? Is that part of the ROI? You don't have to answer that. (giggles) >> Michael, our last question for you is where do you see this all going? This is very interesting to me because I've inferred from a lot of the conversations that, like that PepsiCo guy was up yesterday, talking about an AI fabric throughout the organization, not just tactical projects, and that kind of interested me, but I expected it's much further off. I'm hearing from Kourtney that it's actually real today. What's your sort of prediction or forecast for the adoption of this more advanced intelligent process automation? Is it kind of just starting now and it's going to explode? Or am I just missing the mark here? >> No, I think you're a hundred percent on. I mean, first off, I think, like I mentioned earlier, RPA and machine learning separately, are in these incredible growth stages. Right, and we think our message to customers now is if you're not thinking about how you're doing AI and machine learning, you're already behind 'cause your competition is. And so you better get thinking about it. I think we're going to get to that level with intelligent automation, with RPA plus machine learning very soon. I do think right now we're in that infancy stage where people are looking for used cases, and they want to hear great stories, and so I do think American Fidelity is ahead of the curve, but they're not going to be ahead of the curve for long. It's catching up. If you're not doing it, we're going to eventually get to that point where you'll have someone like Elon Musk or Masayoshi Son, say, if you're not thinking of intelligent automation, you're already going to be left behind. >> All right, congratulations on the work that you've done. >> Kourtney B.: Thank you. >> It's a really awesome story. Thanks so much for coming on theCUBE. >> Yeah, yeah, thanks for having us. >> Thanks for having us. >> All right, keep it right there, everybody. We'll be back from UiPath Forward day number 2. You're watching theCUBE. Be right back. (upbeat music)