>>From the Bellagio hotel in Las Vegas. It's the cube covering UI path forward for brought to you by UI path. >>Hello from Las Vegas, the cube is live at the Bellagio. We are here covering UI path forward for Lisa Martin with Dave a long time. We're very pleased to be joined next by the chief product officer of UI path, param colon per I'm. Welcome to the program. Thank you. Thanks for having me. It's great to be back here in person. Isn't it? It >>Is lovely to be back >>Here in person, Dave and I got to see a little bit of your keynote this morning, and when we had to leave to come to set, we turned around and it was standing room only. It was amazing to see how many customers, partners, prospects, UI path has brought here. Clearly you've got over 8,000 customers now, tremendous growth. The IPO just six months ago. Last time you were on the cube at forward three was two years ago, 2019, where you unveiled this vision for a fully automated platform. Talk to us about what's transpired since >>It's been, it's been fascinating, been more than 9,000 customers now. Um, you know, when we were two years ago, um, we, you know, we shared a vision with our customers. The vision was all about having an end to end platform. The end to end platform can help companies automate simple processes, complex processes. It can help them automate anything from point in time tasks to long running into and processes. And we imagine that our automations will impact every worker within the enterprise. And, you know, you'll augment the capacity to build automations through citizen development initiatives. Um, we didn't have all the components then. So we worked hard since then to make sure that we can make this end to end platform available that can do everything from discovering automation opportunities to helping companies easily build automation. Whether that automation is for simple processes that are, you know, doing a predictable pattern of work or whether it's using more complex, you know, machine learning and AI algorithms and extracting information from documents to do that kind of stuff. >>And then also helping citizen developers build automation. So build pillar is what differentiates our time to value in the market. Helping companies realize quick valid for that. Then we, the next pillar we look at is manage and manage is all about making sure that our it stakeholders that are imagined managing the deployment of the entire platform from our server components like orchestrator to two components, like robots that run the workflows can, can easily be managed and low cost of ownership and can effectively manage the, meet the, the requirements of it, stakeholders like governance and security and audit and all those kinds of things. And then we have, you know, the next pillar is run, which makes sure that our robots can run on machines. And we're very excited actually in this release announced that our robots can run at cross-platform. So sonically our, you know, we were a Microsoft windows.net based, uh, code base. >>Uh, it only ran on Microsoft windows machines, and now we can actually run those robots on Linux-based machines as well. So you can deploy those robots in Lennox containers. Um, so that's, that's the run pillar. And then we have engaged, which is all about making sure that the software that we're releasing can be used by every user within the enterprise. So it's an engaging experience. Automations are available to business users where they need them, right? If they're working in a line of business application like slack or Microsoft teams, robots are available there, if they're working in a productivity application, so I'd look they're available there. And then we also have a curated set of, you know, engaging interface for people to use UiPath robots and something. We call it UI path assistant. So super excited about this sort of end to end platform. I think it's delivering a lot of value to our customers. Yeah. >>Tweeted out during your keynote, the, all the features under discover a build, manage, run, engage, and could barely fit it in the screen. It was an eye test in terms of just the number of features that you guys developed. My point is, was that, you know, the innovation is the lifeblood of life source of tech companies, you know, demand gen too. But I mean, they're real this, the product, right? And, and so, and the, the indicator is the pace of those innovations. And so you, it seems like you're picking up the pace, uh, as you scale as a company, is that true? Um, in terms of just the velocity of the features that you're rolling out, you guys have been busy since the pandemic hit. Yeah. >>At the, at the expense of, uh, you know, are not, not trying to annoy anybody within the company. We are a very product centric. You know, company it's company was founded by an engineers and, and ran it truly as a, you know, engineering centric company for many, many years. So as we've scaled, I think we haven't lost that attribute for culture to make sure that we are continuing to innovate, continuing to make sure that we are delivering capabilities based on what our customers are demanding of us. And that's been core part of our strategy as well, with all this said that we're going to build what our customers ask for. And we're going to learn from our customers. And it's very expanding our installed base of customers. We're learning more things from them, and we're continuing to respond to those customers as fast as we can and helps the fact that where helps that we're a bigger company now, so have more resources at our disposal so we can do more. Uh, but I'd say, you know, if you look historically at some of our releases over the last two, three years, we've done a lot of innovation in, in every one of these releases. And we do it twice a year. Um, these, these big releases, and this >>Is released 2021 dot 10, >>We're talking about that's right. This is released 20, 21 10. And it's the one of two major releases in a year. So we do a release in the first half of the year, and we do a release in the second half of the year. And that's how we've been doing or this, this is how on-premise customers consume sort of our software, the cloud customers get these capabilities more frequently. That's every two weeks, but the on-premise customers give them, uh, get these capabilities every twice a year >>Time, do the on-prem customers have to get to the new release. I mean, you can't just make it, you know, and minus 30. So what do you give him a couple of cycles? Three, four cycles, two, three. >>We give them so six cycles of three years to upgrade to the next version of the software. Um, most customers upgrade much sooner than that, because if you looked at our software three years ago and you compared it to what we have today, you basically said, there's no reason why I should be paying you for that capability when I can get so much. So generally we see customers upgrading more frequently. We've also because we run the same code base for our cloud customers that we run for our on-premise customers and Cod customers. As we update their service every two weeks, we've built enough resiliency to make sure that the upgrades are seamless for customers. So it's, you know, it doesn't cause them a lot of pain to go from one version to the next. So, you know, we, we expect most customers upgrade within 12 to 18 months of deploying a software. >>Let's talk about acceleration the last 18 months. We've seen a massive acceleration in automation as a mandate for every industry to first survive what was happening in the world. And then to really thrive, how has the pandemic influenced the roadmap, maybe what's in 2021 dot 10. What, what is, and how have you helped customers accelerate their need to automate, to stay in business and to be successful disruptors? >>Yeah. So what, that's a great question. What we've noticed is that, you know, in, in our customer base, there are certain aspects of business processes that have just required a lot more demand to be able to accelerate, you know, certain parts of our business, whether it's a, um, a, a retailer trying to meet demand for a certain set of product category, or whether it's a healthcare organization, trying to make sure they can provide care to more people more, more frequently. Um, the, the truth is that we didn't have to make fundamental shifts in our technology to be able to meet that demand because we had built the core underpinning of our technology was built on the premise that you are allowing a machine to mimic human behavior. And to do that. And if you can throw more compute at those machine, doesn't have more robots. >>You can do more work at that. Um, so the pandemic has definitely accelerated the demand for automation. It hasn't changed our roadmap to go in a different direction as we go there. We've just made sure that we can continue to meet the demands of our customers. Um, the, the, the one thing that it has impacted us, it it's brought automation to the top of the priority list for it stakeholders, right? In the past, we sold a lot and our value and messaging resonated a lot with line of business owners. But as, as the pandemic, as they started using more and more automation, the it stakeholders are now a lot more interested in sort of what we're doing. And we've done a lot of work in ensuring that the requirements that our it stakeholders have from running mission, critical workloads from compliance, from governance, from security, are all met as part of the platform, you know, basically out of the box, as opposed to having to, to bolt on >>One of your competitors, uh, just got bought out by private equity and the PE firms going to mash them together with an integration company. You guys bought cloud elements. I look at it as a nice clean integration play. I wonder if you could talk about the importance of cloud elements, how that fits into your product portfolio and in the marketing. >>Yeah, it's a, it's a fascinating time to be in the, in the automation business, the slots going on, you know, we, when we saw cloud elements earlier this year, what we realized was, as companies are trying to automate processes, there are certain systems where you can get to information. We can get to actions you want to take through API, and it's maybe better to access them through API because the user interface doesn't provide the same set of capability in the same way. And there's other systems where you want to have UI based automation because API APIs don't exist or too hard to use. Historically, there were different products that people had to use either have to use a, an integration platform type product to be able to get to all the API APIs. And it was a very developer driven workflow to put that process in place. >>Or you could use something like UI path where you can connect systems through UI automation. What, and when we looked at that problem, we said, the problem is, is to connect, break the silos that exist across processes and apps in the enterprise. And we wanted to provide one single design environment in which developers can use API or UI automation, or even machine learning based predictions and construct an end to end workflow where you can blend together a set of ML skills, UI automation, API automation, to, uh, you know, to compose that into an automation. So that was our thesis behind that the market definitely seems to be paying attention. And it seems that, you know, other people are trying to follow the same, same path as well as some industry analysts have written about sort of the consolidation in, in categories happening. >>Yeah. You guys started a trend for sure. Let's talk about security. You guys announced a partnership with CrowdStrike endpoint security, clearly the leader there, uh, they're, they're really amazing company. What what's that partnership all about how to, how to come about and, you know, maybe give us some color there. >>Yeah. So we're very excited about this partnership. And as I was mentioning earlier, you know, one of the things that pandemic has done is elevated the importance of automation to the boardroom, to the CIO, the CSOs, and, and now they're looking at, you know, UI path automation platform as a key enabler of getting work done within the enterprise. So when they look at something that's gonna, you know, take like 20, 30% of the transactions in a, in a working environment and move it through that platform, they want to make sure that it has all of the security that they have heard about for all other application than end users are using as well. So CrowdStrike has done phenomenal work in security solutions to manage how employees are using based on their profiles, different applications and different API as well. What we've done is we've taken that the capability that they have built for the employee workforce and make sure that we can apply it to the robotic workforce that our customer is using as well. >>So the same policy based control that says, you know, a, this employee in this department isn't allowed to access the system, or it needs to be logged that they had access to this data. The same can be applied to robot as well. So you're no longer able to say, you know, I don't know what my robots are doing. I need to go through locks to find out he can maintain a policy. You can get alerts when robots do something that they're not allowed to do. And not only that you're able to pinpoint specifically to which workflow process that robot was executing when it tat, and when it touched a system that it wasn't allowed to. So it gives you that peace of mind, if you are a business owner or an it stakeholder at a company, to make sure that my robotic workflows isn't doing what it is supposed to do. >>So the point of integration for Crosstrek is a robot. Is that correct? It is, it is a robot. So it monitors what the robots are doing. And it, it makes it reports it back into the, their Falcon interface that basically says I've got a policy that when applied to my robot workflows and when, as they're executing things, I can monitor control and get alerted of different things that they're doing. And the robot today already does it's own identity access, right? Yeah. Right. Yeah. The robot has its own identity. So a robot generally historically, had to go log into many line of business systems to do the work. Right. And if the robot is running in, attended as it it's working with a side-by-side with a human, then it assumes the identity of the person that's working. So they can go access those systems. But if it's working unattended, then it's using a service account to make sure that it can go access the applications that it needs to access, to be able to, you know, pull information or update actions into those systems. >>So you mentioned more than 9,000 customers. Now, I know you've got more than 1200 with a hundred thousand or more ARR customer listening. I know is very important to you. I bet we just had the CIO and digital officer of Coca-Cola on talking about that was a big differentiator when it was, I said, who else did you look at, but what was it that really made you iPad stand out? And that's that customer listening the voice of the customer, the impact that it has on the technology in the company. Overall, talk to me about some of your favorite stories of customers that have really helped influence, especially in the last year and a half, the direction of the company, you know, IPO and six months ago, as I mentioned, how is that customer voice really critical to the innovation that your team is driving? >>Yeah, I know it's, it's very important. And that's something that we've done from like the beginning of, you know, creating this market and entering this category of RPA. You know, it was initially a customer that educated us about what RPA is, you know, brought us in and told us what kind of solution they wanted and to help us create the first solution. And we worked very closely. In fact, I joined the company about three and a half years ago. In my first month, I spent a lot of time with a customer in Japan to learn from what they thought about automation, very large bank in Japan, and we've continued that process. So we have, um, every quarter we eat with our customers around the world and customer advisory boards. We run a product advisory board every six months, very specific information at a feature level on what we're working on. And we ask our customers to ward on the feature list and we decide based on how much importance they put to different product areas to see what we're going to innovate on. Um, we feel this is the best way to build enterprise products. You know, it's, it's, it's helped us tremendously. Uh, and, and we're, we're extremely grateful to our customers to be able to, you know, help us shape the platform in this way. You know, >>When the whole discussion around AI first started AI and RPA, and there were some naysayers, and it seems like you're applying AI everywhere. It's hard to even understand it sometimes. So can you help us understand your strategy with respect to AI, where aren't you applying it, but maybe you could help us sort of shape that discussion. >>That's really a good question. So, um, I I'd say AI is a core intrinsic aspect of what we do in our platform. Um, and I'll tell you how it, it matters. It matters because what we're trying to do at the end of the day is mimic how people work, right. And, and make sure that, you know, we can process the systems that just like humans processes as well. So when you look at something on your computer screen, you're using your computer vision to be able to identify what is on the computer screen, whether you're looking at a browser or you're looking at a specific control that you can enter the customer name, or, you know, hit the button. So our technology essentially give our robots the same skills to be able to understand what is on the user interface in front of the screen. And if I'm going to give you an instruction to say, start a certain application and click the new button, and then enter, you know, this information on this form, the robot is able to process that exactly the way you would do it by using those eyes. >>And that's what we know intrinsically. We build a computer vision capabilities. Now we've applied that, that ability to understand computer screens, to documents as well. And that's where we build machine learning bottle models for extracting information from documents. So if you, if you hired somebody in accounts payable department, and you showed them a piece of paper, and you said, Hey, this is an invoice. Please put it into this system. You know, they are able to look at that invoice and when they get the next document, they know this next document is also an invoice. And they're able to put that in, you know, extract information from it and put it into a line of business system. So we're teaching a robots, the same skills to be able to classify documents, a certain document types and using machine learning models, extract relevant information from those documents. Even if one invoice looks different from the second invoice, you're still able to, as, as human I reliably extract information from it, the robots have the exact same skills to be able to deal with. >>So, okay. So classification, it makes sense because you're using math and now you have enough processing power and enough data. And so you can apply that those algorithms does the AI help. Maybe this is a stupid question, but does it, does it give context as well? Right? It can, it, can it interpret context like a human, could we, we, at that point, >>Um, well, it depends on what you mean by context in general, you know, the applications of AI, um, broadly speaking are narrow, right? And they're not able to go broad and, and understand the entire context that's where humans are better, right? That's where humans are able to sort of truly apply the context, especially if it's, you know, a rare occurrence of that context. And that's why we fundamentally believe that, you know, AI is powerful only if you can apply together with humans. And so we've built capabilities to support like a human in the loop scenario where if a document was read by the robot and it extracted information, but extractive the low confidence or it didn't, wasn't able to find all the right fields of a supposed to, we can suspend the process, send the document and the extracted information to a human and the human can correct what they've extracted. And with that, you know, you're creating retraining model data for the model to behave appropriately. Next time it sees similar information in the next document. So we do believe that, you know, AI is not going to solve all the problems. Yeah. It's going to be able to solve the narrow path, but in, as you look at sort of broader and contextual specific things that are out of normal behavior, you would still need humans to be able to guide what robots are able to do. >>Chatbots can interpret contexts. I don't know if that's because it's brute force and he's just throwing a lot of repetitive sort of data. I don't know if you guys seen the, um, the, the very first AI, standup comedy. No, you haven't seen this. Yeah. >>It's, it's the >>First standup comedy routine written by machines. And it's a lot of, it's really stupid, but some of them, some of the jokes aren't bad, you know, and they have an audience and they're like this, the fake laughter coming in. But so it seems like it's amazing. You have to I'll, I'll send it to you. You have to check it out. But, but, so we're starting to get to the point where you could interpret context or create context. And like you say, human in the loop allows you to sort of verify >>That are correct and validate, you know, robot actions, actions. Yeah. >>Okay. So you'll just kind of ride that AI curve, wherever it takes you. Right. >>Well, I think yeah, is really important and they'll become increasingly important as we look at the future. We're also looking at, you know, in the future AI based development or AI based, helping of build building development script. So something that, you know, you'll hear Daniel talk a lot about semantic automation. So where, you know, investing a lot and ensuring that we can have semantic automation type capabilities into a product that helps, you know, the process of building automation simple enough, so that more citizen developers can build it and drives developer productivity. And at the end of the day, I have a specific use case. I want to show you offline and tell me if you can help. >>I got to ask you one more question. As here we are talking about humans and AI surrounded by humans. Finally, for the first time in a long time, what's just been some of the feedback that you've heard yes. Between yesterday and today in the last 30 seconds that we have here, >>The feedback about the product specifically, that the >>Technology, the direction of the company, what you're announcing. >>Yeah. Our customers are very excited about, you know, some of the things that we've announced today, um, you know, I'd say the excitement is specifically around some of the things we're doing around the discovery pillar. We've got a lot of excitement around what we've acquired through cloud elements and the integration service that's available as part of our platform. Uh, we've got customers very excited about running automations on Lennox machines that they can scale them, manage the total infrastructure, uh, around it. Um, we've also released a automation suite that gives our on-premise customers the same sort of manageability and deployment, uh, capabilities that are available to our SAS customers so that the admin experience has improved. So in general, the feedback around the innovation and things that we're doing has been very, very positive from our customers. >>The validation problem. Thank you for joining David me on the program today, talking about what's new, congratulations on that. And for holding a very successful in-person event, >>That's a big deal. Thank you so much for having me. We're happy to >>Be here for Dave Volante. I'm Lisa Martin. We've been coming to you all day from Las Vegas. We're going to see you tomorrow. Same channel from UI path forward for C then.