 theCUBE presents UiPath Forward 5, brought to you by UiPath. Hi everybody, welcome back to Las Vegas. You're watching theCUBE's coverage of UiPath Forward 5. We've reached cruising altitude on day two. Christina Seacrest is here. She's the process artificial intelligence and automation GPS automation leader at EY and Bob Pucci's executive director for intelligent automation for the state of Tennessee. Folks, welcome to theCUBE. Thank you for having us. Good to have you. Okay, I don't know if I messed up that title, Christina, but it's kind of interesting. You got process, you got AI, you got automation, you got GPS, what's your role? I have a lot of roles, so thank you for that, yes. So my focus is, first and foremost, automation. So how do you get things like UiPath into our clients? But also, I focus specifically on our government and public sector clients. So SLEZ specifically, so state local education. So that's why I'm here with the state of Tennessee. And then we also like to take it beyond automation. So how do you bring artificial intelligence and all the technologies that come with that? So really full end-to-end spectrum of automation. So Bob, when you think about the factors that are driving your organization, how did you describe those sort of external factors that inform your strategy? What are the catalysts for how you determine to deploy technology? Well, primarily that, Tennessee says we tend to provide good customer service, but we want to get to a great status, best in class, if you will. And we had an external advisory review where it said, hey, we could make automation to improve our customer experience. And so that was like a directive of the state leaders to go across the board and automate all processes statewide, starting with the 23 executive agencies. So where's the focus from that standpoint? Is it on just providing better interfaces to your constituents, your customers? Is it cutting costs so you actually have more budget to invest kind of a combination of those? So it's really both qualitative and quantitative, right? So quantitative is where we're able to reduce hours and therefore we can redirect people to more less mundane work, if you will. And then qualitative is where we're able to reduce the errors, improve data quality, reduce cycle time for our citizens when they're making requests, et cetera. So I think it's a combination of both of those, of quantitative and qualitative metrics that we are mandated and micromanaged quite frankly to make those numbers. It's not from Massachusetts. When I go to a mass.gov website, I say, oh, this was done in the 1990s. And you could just see where the different stovepipes were, but then every now and then you'll hit one and you say, wow, okay, this is up to, it's such a great experience. And then the flip side of that is you want your employees to be happy and not have to do all this mundane work so you can retain the best people, you don't have to. So you're living that in state and local. So where did you start your automation journey? What role did EY play? Let's get. Yeah, sure. So I think the thought for process automation was probably three or four years ago, but then we started the program about 18 months ago and there was a lot of, let's say, behind the scenes work before we could bring EY in, what resources was I going to have in the state that were going to help me address all of the agencies simultaneously. Cause normally you'll see a project that'll do, be more siloed across the state and say, we're going to do this agency, we're going to do this division. Well, you have 40 other agencies that are, the momentum is just going to fall on the wayside. So how we looked at it was, let's blanket it and go across all 23 agencies at the same time, identify common processes that are used across 40 divisions, for example. So what we basically did is we procured the software, did the contracts and then it was really about, I designed, I'm going to say a multi-stream approach where we could run multiple work streams independent to find all the architectures required, dev, tests, production, disaster recovery. At the same time, in parallel, develop the center of excellence, the operation model, the processes, methodologies. And the third one was, let's go out to a few divisions, business administration, human resources, and be able to do a process inventory to see what was there. And then based on that, there's all this theory of, let's do a proof of concept, let's do a proof of technology, this will apply it well. The bottom line is, RPA technology has been around for a long time. It's proven, there's nothing to prove. But really what was important to prove before we decided to go full tilt was develop a proof of perceived business value. Are we going to bring in the business value, the hours and the qualitative metrics that is expected by our executive team, the leadership? We were able to do that with the help of EY. We built out the prototypes and we got the green light to go forward, got EY to start, and then we just basically went pedal to the metal. We had our foundation already defined. We built up the architecture in less than one to two months. Now in a public sector or private sector, it's just not heard of. But we have a tendency with EY's technical team and myself, we look around the road around the rock and so the rock in the road. So we ended up coming up with a very unique, very easy to handle architecture that was very scalable and then we're able to hit the ground running and deploy in production by December, we're ahead of. Was EY involved in the whole DevTest production DR, Center of Excellence, the process inventory, or did you bring them in? Did you kind of do that internally and then bring EY in for the proof of business value? EY was actually awarded the contract for suit to nuts. Basically the first phase, which was those four work streams I told you about. And they worked with myself and the state of Tennessee infrastructure architecture teams. We needed to get these things defined and signed off the architecture so we could expedite getting them built out. And then they basically ran all four work streams. The process inventory, the prototype, the proof of perceived business value, the building out the Center of Excellence and working with myself. And this wasn't just us in a vacuum. We ended up having to, I mean, I could do the strategy, I could do the technology and I could do the roadmap and all the good stuff, but we had to actually meet with a lot of the state of the Tennessee organizations on change management. How do we end up putting this process or an automation in the middle of the normal traditional process, right? So there was a lot of interaction there and getting their feedback and then tweaking our operational model based on feedback from the state of Tennessee. It was all very collective, collaborative. I think that would be the key word is collaborative and then building out everything. So then we ended up going to the next way where they knew so much and we had such a tight timeframe that we continued with EY. So Christina, Bob mentioned Center of Excellence a couple of times in the state of Tennessee, but then beyond state of Tennessee, other organizations you've worked with in this space. What's the relationship between Center of Excellence and this thing we've been hearing about over the last couple of days? The citizen developer. Has that been leveraged in the state of Tennessee, Bob? Have you seen that leveraged in other places, Christina? What's that relationship look like? Yeah, so we don't leverage that model yet. We have a centralized model and there's reasons for that so we don't end up having Mavericks run off. Yeah, run-offs have run-offs that have a UA path version but down this division or have another RPA tool and another division, right? So then all of a sudden we have a maintenance nightmare, a manageability nightmare. So we basically, I negotiated an ELA with UI path so therefore if anyone wants to go do another automation on another division or they would basically follow our model, our design, our COE, our quality gates, we're the gatekeepers of brain into production. Got it. Yeah, now Christina, what's your perspective because I can imagine Nashville and Memphis might have very different ideas about a lot of things. Little Tennessee reference there. But what about other places? Are you seeing the citizen developer leveraged in some kinds of places more than others? Not yet and that's part of because of the foundation we're building. So we laid, when Bob talks about the first phase of eight weeks, that was amazingly fast. Even when you spoke about it. That's ridiculous. To say you're going to lay these four foundations. I was excited. Like I was like, wow, this is a very serious client. They want to go fast and they want to get that momentum. But the AOM was laid out so we could propel ourselves. So we are at 40 automations right now. We're in the works of creating 80 more automations in this next year. We'll be at 120 really quickly. The AOM is critical and I will say at a client, I've worked with over 50 clients on automation programs. The way State of Tennessee treats the AOM and they abide by it, it is the living document of how you go and go fast. Got it. The one thing I would say is it's also allowed us to have such immense quality. So I always talk about you put in 40, you put in another 80. We're at 98% uptime on all our automations. Meaning they don't go down. And that's because of the AOM we set up. And the natural progression is going to be how do you take it to citizen developer? How do you take it to we call process automation plus. But methodically, methodically. It's methodically. Not just throwing it out at the beginning and hoping the chaos works. Exactly, exactly. In the ratio of bots to automations, is that one to one or you have automations? Oh no. A single bot is doing multiple, so how many bots are you talking about? We're doing, Bob you're going to answer this better than I will, but the efficiency is amazing. You've been pushing that. So our ratio now, because we have a high density architecture we put in, is four bots, excuse me, four processes, the one bot, and four bots to one virtual machine EC2 server. So it's four to one, four to one. Now what we're going to get by next summer we'll do more analysis, we'll probably get to six to one, six to one. That's made serious shrinkage of our footprint from machine management perspective. From 60 down to seven. Now we're going to the next chunk we add other Haiti automations in fiscal year 24, we're only going to add two more bots, two more servers. That's only 10 running on like close to 200 bots. And is doing this on-prem in the cloud? The architecture is fully cloud based. So we use UI past SAS model. So that handles the orchestrator, the attended bots, all the other tooling you need, automation hub, process miner, et cetera, et cetera. And then on the state side in AWS we use unattended bots. Sort of bots that have to go down into the legacy systems, et cetera. And they're sitting on EC2 instances. Was there a security knot hole that you had to get through internally? What was that like? No, actually we were lock and step with the security team on this. I mean, there were some standards and templates and what we had to follow. But they're doing an assessment. Every single release they do assessments on all the bots. What systems it's activating or accessing, et cetera. The data, you can see a FedRAMP data of FTI data in the public sector to make sure we're not touching that. Do you guys golf? I do. Not well, but yes. I mean, I like golf, but don't golf well. But so you know what a mulligan is. If you had a mulligan, right, for the state of Tennessee, what'd you learn? What would you do differently? You know, what are some of the gotchas you see? Maybe Christina and other customers. And then maybe specifically, state of Tennessee. Right, I would say, you know, it is the intangibles. So when we talk about our clients that go fast and go big like state of Tennessee, it's because that we call it phase zero that gets done that Bob did. It's about making sure you've got the sponsorship. So we've got executive sponsorship all the way up. You've got amazing stakeholder engagement. So you're communicating the values of what we're trying to do. And you're showing them the value. We have been really focused on the return on investment. And we'll talk a little bit about that, but it's how do you make sure that when you do, you know, states are different with those agencies. You have such an opportunity to maximize return on investment if you do it right. Because you're not talking about an automation in one agency. You're talking it across multiple agencies. We call that the multiplier effect. And that's huge. You understand that and how to actually apply that. The value you get is amazing. So, I can't say there's a mulligan here. Bob, you may think of some. I know on other clients, if you don't line up your stakeholders and you don't set the expectations early on, you meander and you may get five, six automations in over the year. You know, when I go to clients and say, we're doing 40, we're doing 80, they're like, wow. But that's the bottom line, gotcha, is if you want to have an operational impact and have multiple zeros, you got to go through that process that you set up front. Exactly. Anything you do differently, Bob? Well, what I do differently, I mean, I think, I mean, we did get executive sponsorship in one area, but we still have to go out to all the 23 agencies and get, and bring awareness and kind of like set the hook to bring them in, right? Bring them to the lake, right? And I think if it was more of a blanket top-down, getting every agency to agree to investigate automation, it would have been a lot easier. So we're getting it done. We've gone through 13 agencies already in less than a year. All of our releases are sprinkling across multiple agencies. So it's not like a siloed, I'll look at that age. Everyone at every agency is being impacted. So I think that's great. But I think our model now is just trying to make sure we have enough backlog to do the next sprints. Is it, you know, the ROI on these initiatives is so clear and so fast. Is it self-funding? Is there gain sharing? Or do you just give business, give money back to the state and have to scramble for more? Do you get to, you know, get a lick off that cone? Unfortunately, we don't. But I tried to see if we could get some profit. It's like, nah, we don't do that, it's all cost-based. But our ROI is very attractive, I think, for doing a whole state, you know, transformation I think our ROI is three and a half to four years. And that's pretty mind-blowing. Even if you look at private sector, I think some of the key things which people are noticing even though we're in public sector, we are very nimble. This project is extremely nimble. We've had people come in, exactly. We need this or we're going to get penalized. Okay, knock it out in four days, right? So it's that nimbleness that you just don't hear of even in private sector or public sector. And we're just able to do that for all the collaboration we do across EY, across myself and across all the other organizations that I kind of drag along or what have. What do you, do you see any limits to the opportunities here? I mean, is this a decade-long opportunity? Is it, do you have that much runway? Well, that's just not my DNA. So we're going to probably do it like in four years. But what- Well, when you say, do it. I mean, will you be done at that point or do you see the opportunity to really drive out of it? So the way that we look at it is, you know, we could boil the ocean and I think this is one of the reasons why we're successful is we could boil the ocean and it would be a 10 to 20 year program. Or we looked at it, we had some EY guys look at it and say, I said, what's the 25, 80 rule? Meaning, you know, give me, so we had 500 processes. Tell me how many processes will give me 80% of the hours. And it was 125. It was a 25, 80 rule. I said, that's what we're doing. We're doing, we're going to do the 80% of the hours quantifiably. Now, when we're done with that pass, then we'll have those other ones that are bringing 20% of the hours. That's what we might be bringing citizens in. That's what we're bringing state workers in. But at that same time, we will be going back in a wave and doing advanced AI. Right. Advanced AI. In other words, so right now we do RPA, OCR, ICR. But you know, there's NL, NL, NPS, there's virtual agents and stuff. So that's like the wave we're going to do through the ones we've already gone through. Got it. Right. So it'll probably be a two or three wave or iterations. Cool. Guys, thanks so much for coming to theCUBE. Great story. Really appreciate you taking us through it. Thank you so much for having us. You're very welcome. All right, keep it right there, Dave Nicholson and Dave Vellante. We'll be back at UI Path Forward 5 from the Venetian in Las Vegas. Keep it right there.