What's Up with Tech?
Tech Transformation with Evan Kirstel: A podcast exploring the latest trends and innovations in the tech industry, and how businesses can leverage them for growth, diving into the world of B2B, discussing strategies, trends, and sharing insights from industry leaders!
With over three decades in telecom and IT, I've mastered the art of transforming social media into a dynamic platform for audience engagement, community building, and establishing thought leadership. My approach isn't about personal brand promotion but about delivering educational and informative content to cultivate a sustainable, long-term business presence. I am the leading content creator in areas like Enterprise AI, UCaaS, CPaaS, CCaaS, Cloud, Telecom, 5G and more!
What's Up with Tech?
Future of AI in Business: Deluxe's Journey with Gen AI Bots and Automation
Interested in being a guest? Email us at admin@evankirstel.com
Ever wondered how AI and automation can revolutionize your business operations? Join us for an enlightening conversation with Deluxe, who share their experiences in transforming Deluxe into a modern payments and data powerhouse. Learn how they navigated the complexities of managing a vast client portfolio and implemented groundbreaking AI solutions to streamline their processes. Kari recounts the challenge of extracting data from 1,200 legal documents, and Srini's innovative use of a Gen AI bot from UiPath that not only enhanced efficiency but also improved accuracy.
Discover how Srini and Kari are unlocking broader opportunities for enterprise AI and the importance of having change champions like Sharini to drive awareness and refinement of automation tools. Hear their excitement about the potential to handle multiple projects simultaneously and transform tedious tasks with cutting-edge solutions. We'll also preview the upcoming UiPath event in Las Vegas, where both Srini and Kari look forward to networking with professionals and scaling their automation initiatives. This episode promises valuable insights for anyone looking to leverage AI for business growth and operational excellence.
More at https://linktr.ee/EvanKirstel
Hey everybody, fascinating chat today as we talk about revolutionizing business operations through AI and automation with Deluxe. A UiPath customer, srini Kari. How are you Good? Good to see you both here. Let's get started with some personal professional introductions here. Srini, I'll start with you. Tell us about your role at Deluxe and who is Deluxe.
Speaker 3:For those who may not be familiar, Sure, I'm Srini, For those who don't know Deluxe. Deluxe is a longstanding original payments company. We're actually transforming ourselves to be a modern payments and data company. We really help small businesses and large institutions and organizations really scale their business and grow their business through our payments and data products and assets. I, within the operations organization, am responsible for product management function and also driving process improvement and automation across the enterprise for Deluxe, leveraging the automation and the technologies.
Speaker 1:Fantastic and Kari, if you could introduce yourself and your role as a colleague of Srini's?
Speaker 2:So my name is Kari and I'm a business lead in treasury services. So my name is Kari and I'm a business lead in treasury services. I'm actually in product management, so Srini and I partner quite a bit on various opportunities.
Speaker 1:Well, it's an exciting time to talk AI and automation, so let's dive right into it. Maybe you'll start with Srini. What is the need for automation? What drove the decision to focus on automation at Deluxe and what were the specific challenges you were originally looking to solve?
Speaker 3:That's a great question, amal. Yes, so, as I said, as Deluxe is more transforming into a modern payments and data company, as we continue to grow, our customers also grow along with this, and their needs actually change along with it as well, and so, as we go in through this journey, our needs and challenges also come up, and one of the challenges that we faced was really around how do we make some pricing adjustments as easy as possible for our customers, as we have a very large portfolio of clients. So that is one particular challenge that we face where we want to do it not only quickly, efficiently but, more importantly, accurately. That really delivers a really positive customer experience for our large clients, and so that's where automation came in. We fell in love with the problem first, and that's why I partnered closely with Kari and the rest of the business teams in this case, contract team, legal teams, product and technology and we fell in love with the problem and we found eventually a solution through automation and technology.
Speaker 1:Wow, what a great setup for the discussion. And, kari, maybe talk about some of the challenges you were aiming to solve in your specific role. Function in the company. Yeah, you have a lot going on. Maybe describe what you do a little bit as well.
Speaker 2:Right. So in this situation we had a challenge where we needed data elements, data attributes, that we didn't currently have housed in a contract database. So I needed to read or reread 1200 legal documents to pull out these attributes. So, like Srini said, we could fulfill the project accurately and hopefully quickly. So I went to Srini and said, hey, can I have 10 people for two to three weeks to reread all these contracts and type all this data into an Excel spreadsheet? And Srini's like, ah, counterproposal right. And he said, how about a Gen AI bot? And I'm not IT, srini's not IT, but Srini had this partnership with Kibatica and Kibatica had this tool, uipath, and it kind of just all came together from there.
Speaker 1:Well, it's a great case study, great story and best practices. Maybe talk, shreen, a little bit about the proof of concept and some of the benefits of automation. You know how did you initiate and start, kick off the engagement and you know what were some of the outcomes that you saw from automating the contract extraction process. You know short-term, immediate as well as some of the longer-term benefits process. You know short-term, immediate as well as some of the longer-term benefits.
Speaker 3:Absolutely. That's another great question. Before I jump into the proof of concept itself, I just want to take a second to spend on the problem itself. The problem was pretty complex, as Korya suggested, and these were documents that had handwritten, in some cases, images of not the best quality, because we have very historical, long-lasting clients with contracts drawn out many years ago, and so the scale and the complexity of the problem is very high in terms of the data that we need to be extracted out of these contracts. And so that's when we had actually went and searched for a solution where we want to bring innovative and a creative solution that actually solves the problem quickly but also accurately. That's when we came across, or we identified UiPath as the technology provider of the platform that allows us, I think, we leverage the document understanding aspect of the platform, which really helps us to quickly read documents very quickly and extract some attributes, and then leverage our Kubotika partner to go implement this rapidly so that we can do this proof of concept.
Speaker 3:Spending a little bit more on the proof of concept itself, the first step was to really create a solution where we are able to automatically read all of these documents 900 plus documents that Corey has suggested, that are in various different formats, images and handwritten something like a lot of those information. So we had to automatically read those information. So we created a bot that goes into our Salesforce system reads these documents. Second, we leveraged, like I said, the document understanding portion of the UiPath platform to read these attributes. The beauty was that we didn't have to go build our own Gen AI algorithm or Gen AI solution.
Speaker 3:The capability existed within the UiPath platform, which was pretty easy to use. Business folks and other folks like Kari was able to really provide us the keywords that need to be extracted and the algorithm and the capability was easy enough for us to tune and configure and set up. That was step 2, is to really provide the keywords and the type of variances that we would get so that we can extract those attributes very cleanly. Third, once we do that validation, then we actually put them into an action center where the human aspect comes in. This is where just the technology alone doesn't solve the problem. The human is still very much important and the human is the last step in the sense that it ensures that the data we extracted is correct and is accurate and does the once over. Then, finally, we then is correct and is accurate and does the once over, and then, finally, we then scale it and deploy it so that we're able to take that specific information and be able to be surgical about our price adjustments to our customers.
Speaker 1:Wow, fantastic approach. Maybe we should have invited the bot to this webcast and ask a few questions. It would make it better. We only have human-seer. So, Carrie, maybe talk about the human in the loop for final validation and the role of you as a human in this technology project.
Speaker 2:Yes, as a non-IT human right. So it was pretty amazing because that action center that Srini mentioned, that is so intuitive. It was easy for me to, you know, go through line by line and just quickly see which attributes were pulled, and right to the side of it, if I wanted to, I could look at the kind of the extract it pulled it from, or I could look at the whole contract. If I was doubting it or if I needed to change it. I could do that After several refinements though.
Speaker 2:So we did the action center with like a smaller, like a sampling of the contracts and we're able to really refine the bot better. So by the time we did the 1200 and got back into the action center, it was pretty bot better. So by the time we did the 1200 and got back into Action Center, it was pretty darn accurate. So we weren't finding many issues anymore, but there's always a couple right. So. But someone who's not technical, who's over 50, but with a lot of institutional knowledge can be very impactful, and I think that's kind of what's so invigorating and it's got me so interested in upskilling on AI now.
Speaker 1:Yeah, it's exciting times and talking about impact, maybe, srini, talk about some of the outcomes that you realized from this automation project. You know whether it's saving time, data quality. You know allowing other initiatives to proceed that you now have time for. Give us a holistic picture.
Speaker 3:Absolutely yes. So when it comes to the outcome, I would like to kind of separate this into the qualitative outcomes and the quantitative outcomes, if you will, for our company. For us, what's more important than the qualitative that I'll touch upon here in just a second is number one is this really helped us collaborate across Deluxe? More importantly, we're trying to be more of an innovative company and this really allowed us an opportunity to bring an innovative solution to the mix and teach the organization, share with the organization new ways of working, for example, leveraging AI and automation very easily to help solve not only our own problems but our customer problems. So that was definitely one of the big outcomes for us. It's the new ways of working that we were able to share and kind of teach the organization in general.
Speaker 3:Second is, on the quantitative side was really around a lot of benefits for Kari and the business, around time to value. Here we had 99% time to value, meaning we were able to accelerate this process. This whole bot and solution was built in less than a sprint, which really, in other words, without this automation, which was very manual, it would have taken us, pretty sure, many weeks and nights. I think it's probably around in the realm of six to eight weeks at least in the minimum, and so it really helps us to accelerate a solution, and the beauty is it's not just a one-time use case. This is basically we can rinse and repeat this as many times as you want, and so that's where the scalability comes in. Second is, I think, overall from a timing perspective, in terms of the efficiencies, no-transcript to go, focus on high value activities, spending time more with our customers, and so on and so forth, and that's really exciting.
Speaker 2:So important because that the work they would have done would have been so tedious, right, rereading twelve hundred documents and typing it into a system. I mean, these people have customers to serve, right and and good, high quality projects to work on, and so no one had to work nights and weekends. It was just a kind of a win all around nights and weekends.
Speaker 1:It was just a kind of a win all around, yeah, and I bet you have other ideas for how you could expand this into other areas of the business within your personal domain, maybe share what the future might hold.
Speaker 2:So I'm working with Cubotica right now and we're further refining the prompts so that we can use this spot more broadly, because now, thanks to Sharini, other parts of the business are aware we have this capability, so we want it to be able to be used across the enterprise and then also, like you said, for my own next upcoming projects. Of course, I have data attributes that I want that, as the business evolves, I need different attributes, and so all these attributes, I'm just writing those prompts as well and we'll be able to run the bot for those.
Speaker 1:Fantastic and you know, srini, you have a lot of takeaways and lessons learned here. Just for those who might be watching or listening, who haven't embarked on an automation or AI journey, any advice, lessons learned both on the technical or business side that they could benefit from your experience.
Speaker 3:Yeah, I can share a few things that we've learned. I think the first and foremost that's important is it's less about the technology, it's more about identifying a chain champion and really advocating for the change. In this case, I was very fortunate and lucky to find Kari within the business, who was a big chain champion. That really allows us to showcase our story and the problem and the way we solved it. But I would say, in your organization I'm sure you have problems to solve, for there are definitely lots of inefficiencies and things that you do manually, repetitively, where it's not adding a ton of value. I would say identify those set of activities and go do engage in a pilot, a proof of concept, and find a person like Kari within your business who is bought into this problem and she has a problem. Or you want to go solve this problem for somebody.
Speaker 3:But that's what I would do Identify a problem and a need, fall in love with the problem and then find a champion and then probably don't be afraid to go do a proof of concept. I think the solution is easy Most of the time. It's more important to spend time on the problem and then doing a POC to validate the problem and finding a solution would be a good next step for you, so I think it'll be worth it. Give it a shot. The worst you can do is it may not work out, but I'm sure you'll learn a thing or two from that process.
Speaker 1:Fantastic, and maybe you could share a bit of your long-term vision, not just for automation, but AI at work, something that UiPath is doubling down on. What does the future look like and where do you see the most growth opportunities into the next two, three, five years?
Speaker 3:Sure, like I said, for us, automation is just not more than that. Automation is going to be key for us in the long-term. Our vision is to really make automation a part of our DNA in terms of how we work. That means we're going to scale and find new opportunities for us going forward, which means we're actually leveraging some of the other capabilities within the UiPath platform. Sorry about that.
Speaker 1:No, it's exciting times and hopefully not an emergency, but definitely there's a huge opportunity for folks, kari, who aren't necessarily technologists or IT practitioners. What are you excited about personally and from a professional standpoint, about what's next with automation and AI?
Speaker 2:Yeah, no, I'm excited. You know just when you look at the parts of your job that are tedious or repetitive, like Srini said, or the parts of a project that you think, oh gosh, how are we going to get that information, or how are we going to do this in the timeline that's you know timetable that's been given to us. This just really, just you know that whole thing about you can do anything with time and money. In this situation. It was very little time and very little money.
Speaker 2:You know it was more about who did you partner with and what tool did you use and, you know, did everyone stay focused? So it kind of creates a new paradigm and it's just really exciting to think of all the things we can pursue and kind of the number we can, the number of projects we can take on and maybe concurrently work, because it's so much more efficient.
Speaker 1:Yeah, I love that. And so we have the Blockbuster UiPath event coming up forward in Las Vegas. It's the annual get-together for UiPath customers and partners. I've learned so much at these events. What's on your radar, srini? What are you looking forward to most over the next weeks months?
Speaker 3:Super excited to be at the conference to really learn more about UiPath and UiPath's customers in terms of how customers of UiPath are actually solving some complex problems like us and definitely an opportunity for us to learn. More importantly, we have a really good backlog of problems that we want to solve really some complex problems. I'm super excited about that. So I'm really looking to double down on leveraging UiPath's capabilities to solve some additional problems for us, and so it's a great learning opportunity for us and sharing some of the things that we have done. Hopefully we can really scale this within our organization and leverage AI and automation to really help grow our business.
Speaker 1:Well, that's quite a mic drop moment and, carrie, I guess you're looking forward to that as well and networking with amazing like-minded professionals not all IT geeks which is a refreshing change from many of the tech shows. I go to Okay, thanks guys, Thanks so much for sharing your update and insights, and congratulations on all the success.
Speaker 3:Thank you.
Speaker 1:Thank you so much. Thanks everyone.