Branching Out to AI (Guest: Samuel Best)

Andreas Welsch: 0:21

Today we'll talk about branching out to AI and who better to talk to about it than someone who's doing just that. Sam, thanks for joining.

Samuel Best: 0:32

Yeah, my pleasure. Happy to be here.

Andreas Welsch: 0:35

Awesome. Hey, why don't you tell us a little bit about yourself, who you are and what you do.

Samuel Best: 0:39

Yeah, absolutely. So first, excited to be here to share a little bit more today. So I lead the business automation team at General Motors Financial. So really what my team is responsible for is using various technology tools to automate manual processes. So we've been at this journey for a few years now. So just to give you some size of what we've been working on, we have about 80 unique automated solutions that we've built for our business partners, resulting in over 9 million automated transactions. Now, what's cool about that is that was 9 million transactions that humans had to do, but now they don't have to do anymore. They're freed up to do more value add activity. And also you have to think about the customer touchpoint, right? Because whether those were being done specifically for customers through our operations and ultimately the customers at the other end of that. So we're really proud of the work that we've been able to free up for our partners.

Andreas Welsch: 1:37

That's awesome. Hey, Sam. It really sounds like you've seen quite a few things. And I think in your previous role you've been a consultant working on these kind of projects as well.

Samuel Best: 1:48

Yeah. Yeah I've been in the automation space for about nine years now. So here at General Motors Financial, and then previously, as you mentioned, I was in technology consulting doing the same thing for companies all over the globe.

Andreas Welsch: 2:02

Awesome. So again, like I said, really excited to have you with us. And for those of you that are just joining the stream. Drop a comment in the chat what you've worked on so far. Is it RPA, RPA and AI, just AI? Maybe are just learning about this as well, and you're hopefully using this as a good resource to get started. I'm really curious where you are on that journey. So Sam what do you say? Should we play a little game to kick things?

Samuel Best: 2:30

Yeah, let's do it.

Andreas Welsch: 2:31

Okay, perfect. Alright. This game is called Fill In The Blank. It's a new one. So when I hit the buzzer, the wheels will start spinning and when they stop, you see a sentence. And I'd like you to complete that sentence with the first thing that comes to mind and why. So fill in the blank and to make it a little more interesting. You'll only have 60 seconds for your answer. And for those of you, again, watching us live, drop your answer in the chat and you know why you think it is what you say. Sam, are you ready for, What's the BUZZ?

Samuel Best: 3:16

Yes, spin it.

Andreas Welsch: 3:18

Okay, perfect. Then let's get started. AI projects are like? Go. One minute.

Samuel Best: 3:33

Great question. AI projects. I think are like growing up. So what I mean by that is there's a lot of investment, a lot of time that goes into AI. There are some quick wins, but there's a lot of failures, a lot of learnings along the way. Think about a child learning to walk, right? We fall down, we get back up, and we learn how to dust ourselves off. I think where AI's getting, especially how it's getting more pervasive throughout enterprises, how we've seen some of those early lessons learned on what to do and what not to. I think it's gonna be more like growing up. There's gonna be more maturity, there's gonna be more use cases. Things that might have been impossible before are now possible because of the lessons learned that we've had previously.

Andreas Welsch: 4:18

Fantastic. Growing up, that's a great example in the learnings. So thank you and within time. So now I know you've done obviously a lot of work in RPA and at GM Financial as well, and you've seen a lot of great success. And you had mentioned as we're preparing for the session that you're now headed towards AI. The thing I'm curious about is, because I hear a lot about it, start with AI, then go to start with RPA then go to AI. What would you say how can an RPA program prepare you for AI and prepare your organization for?

Samuel Best: 4:59

Yeah that's a great question. And maybe before I jump right in, just to set up how we're structured here at GM Financial, is that my group, as I mentioned, we're doing process automation stuff primarily via RPA. However, we have other groups that have been working on AI machine learning solutions for quite a while. We. A phenomenal chat bot that helps with our customers, called Nancy actually won a AI use case a couple years ago for the way we price off-lease remarketed vehicles. So that team has done a fantastic job as far as building these capabilities internally. So when I look at, if we have these kind of two unique capabilities, we have AI over here, we have the RPA team over here. How can we combine these two mature capabilities to even unlock more value for our internal operations and for our customers? So that's how we're set up here at General Motors Financial. I would say that I think these have a lot more in common than people think, because a lot of times we have the propensity to think about the technology initiatives that we're working on. While this is machine learning, or this is maybe specific data science, this is rpa. When really there's a lot of rinse and repeat on how you get things done. So what I mean by that is having been in automation for the last nine years, The, one of the larger things that I see where people fail is that they try to do things by themselves. Maybe they're working in a silo within the business and they may see some limited success, but it's really hard to serve your global customer base and your global internal customers as well too. So really what I have found, that if you have things. How do you deliver an automation? How do you assess opportunities? How do you realize value? How do you put together a product roadmap? How do you work with different stakeholders like IT and your business partners in HR? Regardless of the technology, those things are applicable for both. Process automation and AI. So I see a lot of people fall into the trap of RPA wasn't successful, so I'm going to go now do AI. Which I think is a mistake waiting to happen because if you weren't good at the basics on how you deliver a product to your customer with RPA, it's gonna be challenging with ai. So I just think good product management and then shuffle the technology in there, I think is really a good recipe for success.

Andreas Welsch: 7:25

Fantastic. Thanks. Thanks for sharing. And I think what you said about getting the basics right and the fundamentals in place and then. Adapted and roll it out to something that is in inherently more complex like AI than I think that's a great suggestion. So looking at the comments just real quick so we have some here where where folks are on their journey. And from Petr RPA and AI in different industries, from Glenn Process orchestration with many large financial services institutions, Jesse, Martin and the educational materials. Perfect. So I think a good mix of where folks are. So maybe let's let's switch gears a little bit. And on one hand, like I said, start with RPA. See what you can replicate and build upon for AI. But I think that the part that I'm curious about and that I'm assuming our viewers are curious about as well, is what is one of the use cases that you guys are looking at, maybe in addition to the chatbot Nancy that you.

Samuel Best: 8:33

Yeah, that's a great question. So when I think about, again, where we're at, the RPA process automation team, and then the AI team, how can we amplify the technology strength that produce more value for our customers and our internal customers as well? And really what we're beginning to experiment with is how do we improve customer interactions? So what we like to look at is what is RPA good at? It's really good at rule-based processing, processing stuff and systems, doing requests, those type of task activities where AI is really good at, interacting with customers, trying to understand what they need and what they want. So really we're taking from both programs what the strengths are and trying to combine those two together. So where we may have a request on the front end, right? That AI is able to collect and determine what the customer needs. Could we then pass that to an RPA bot if we didn't have an API available? To then actually execute that task and solve the problem right there. So that's one of the the use cases that we're looking at. Again, trying to combine the strengths of both technologies to automate more of an end-to-end process.

Andreas Welsch: 9:43

Perfect. So really leveraging what each one technology is good at or in intended to be used. And then connecting them. Now, what does it look like organizationally, because I heard you talk about different teams on one hand, RPA and on the other AI. How do you collaborate?

Samuel Best: 10:02

Yeah, that, that's a great question. So it's not one group or one team that does everything. It's, it really goes back to the old adage that it takes a village and that's I believe with anything worth worthwhile does take a lot of very smart, collaborative people working together. So really, I know when my team has done experiments, we really try to take a product. Type of view where, how do we be more collaborative? Let's not look at this as traditional. Let's do all the paperwork up front and, document everything. It's real. We try to replicate, although we're virtually quite a bit. We try to replicate if we are in the same room and we had a whiteboard and we had to figure out a solution and document things on the whiteboard. Really that collaborative approach. So that's really how we we try to work. Do short sprint experiments to either try to prove things or disprove things, but either way, you're learning along the way and you're able to take those learnings and apply them to future initiatives.

Andreas Welsch: 11:04

Great. Now I had a conversation with somebody just recently about how do you decide how long these projects run in at what point you maybe pull the plug and say that's been going on for so long. We're not seeing the results that we want and we need. What have you seen there? If anything, how do you think about that?

Samuel Best: 11:28

Yeah, I think that's an interesting question and I think you can easily fall into this if you don't have a predefined measure of success upfront, right? Even if it's as simple as, hey, we're trying to prove this technology or whatever you need, some sort of mission statement or some sort of measure of success before you start an experiment, because then you can get into no man's land, right? Where you're in there, you're still churning through it. It doesn't feel great, but we don't know if we need to pull the plug yet. But if you all have kind of this alignment up front that, hey, this is what we're trying to prove via this experiment, and this is the time boxed activity in which we're going to do that, then it's a lot easier to have conversations on do we need to extend, do we need to modify, or do we just need to pivot to something else?

Andreas Welsch: 12:15

Yep. Perfect. Thanks for sharing. I think that's very valuable. Looking at the chat I see a few questions from Michael Novak. Hey, Michael, great to have you with us. I know you're very active in the conversational AI space and voice space. So one of the questions that Michael asked was, who collects the analytics of RPAs to measure and improve them? A business team or tech team?

Samuel Best: 12:42

Yeah, so that's something my team does as part of the product management. Not only do we have metrics and things that capture, for use case management, assessing processes for automation, but also in production. We go out there and grab exactly how many transactions the automation completed. We also work with our financial planning team to understand what the, that the monetary value is and a lot of other things, and then we're able to roll that up to a nice d. So that we can see what exactly our automations are doing. And also that drives insights, right? Because we're able to review that and see are there enhancements or are there other things that we can do to make things more efficient?

Andreas Welsch: 13:24

Thanks. Yeah. Great point on the dashboard as well because I think if you can actually see the data and show the data, then it's a lot more tangible than just a gut feeling of how this is probably working or how it's not working as, as well as we would like it to be. You can, show the data. That's awesome. Very good. So you see. Just about coming up on time. Maybe let's summarize. So what I take from our conversation are three things. If you have an RPA program in place, reuse, adapt, the governance, delivery lifecycle, business case evaluation and so on. So everything that you've learned around RPA is something that you can readily apply to AI and it's only getting more complex from that sense of how do you manage. That when you add that uncertainty or that exploration of AI into it so you don't really need to start from scratch, but really build upon something that you have. I think the second point that you mentioned was Again, leveraging the strength of the technology rpa more for the repetitive tasks executing, automating certain things, but then AI more for the reasoning side in, in where it's a little fuzzier with what happens. And then the third one what I heard you say was team up with others across the organization and compliment your skills and your skill gaps. So everybody learns and everybody gets the best of both worlds. Would you say that captures it?

Samuel Best: 14:58

Yeah, I think that's great.

Andreas Welsch: 14:59

Awesome. So folks, we're really getting close to the end of the show. Thank you so much for joining us. Thanks Sam.