An Immersive Workshop to Calculate the ROI of AI Strategy - AUTONOMOUS session

All righty. This duo. How excited am I for this one?

Super excited, I hope. There we go.

The numbers are starting to go up, so I'm going to kick us off. So AI investments can be game changers, but how do we measure success beyond hype? So now we have a hands on session with Christian Club, an associate director here at BOI and one of my dear colleagues. So today is an immersive talk where Christian will guide us through the practical ways to calculate the ROI of AI strategy, helping us think through this as we make more informed decisions about AI investments. So get ready to roll up your sleeves and let's welcome Christian.

Wow, what an intro. Akini is putting me to shame on my introductions. All right, everyone, let me know in the chat where you guys are calling in from today. We're going to be having an active subject, active session and I am watching the chat here. My name is Christian Cobb and why I'm talking about to you about the ROI of AI strategy is in part because I started my career as a cpa. I don't even know if all my colleagues know that. And so and then went into tech startups and then into. And then now I'm a consultant here at boi, focusing on AI products. And wow, we've got a bunch of people from all over. This is incredible. So what does immersive mean? Oh, no. Well, let's see. First we're going to talk. So for today we'll be talking about first a frame of ROI of a. For ROI of AI. My intent is to set you up with a simple way to reach success in your thinking about what ROI means. We're going to walk through some example contexts that impact roi and then we're going to talk a little bit about where to find some higher roi. I'll be answering questions. I'm looking in the chat today. Immersive for us means use the chat. Let's learn from each other. I'm going to ask some of those simple, difficult questions and we'll see what people's experiences have been like in the couple years of AI. So let's start with one. I'm curious to hear from you to help tailor a little bit what I'll talk about. What are you guys working on that brought you to this session? Figuring out life is a big one. I'm, I'm gonna, we're gonna solve that today. All right, let's see. We've got digital twins, innovation, roadmaps, implementing within a team, interior design, staying updated, education. Okay. I think we've got a good feel for where we're going for what we've got. We've got some reorganization. Okay. Some more specific use cases like contract management and co creation, applying to sales. Okay, got it. So we've got a feel and I hope you guys can see we're all together in this, figuring out exactly how to get the most out of AI. So let's talk a little bit about, you know, we've had all these changes. We have this tool in AI that really, you know, is so flexible in its application. But there are some things that didn't change in the last three years. One, the definition of roi. We got to get out more than we put in. Two, need for business cases. And there's a spin here. I can tell from the chat that we have a little bit of variation in terms of where we're coming from. But for us at big businesses, we need a business case that is simple because decisions are made by humans. Where does return come from? We're going to break it down into three buckets. Those three buckets are cost efficiency, revenue optimization. Think of this maybe as LT as lifetime value of your customers. Think of this as how we get the most out of our customers today. And then new revenue streams. Where can we go that is totally different? And I'm going to give some examples here on what this looks like. Cost efficiencies. We're often looking at cost reductions. Maybe you can think of hours saved, maybe you can think of minimizing getting the most out of the assets you have. Whereas revenue optimization, we're thinking about decreasing churn, we're thinking about upsell cross selling, thinking about repeat, repeat purchase rates and new revenue growth is those new product services, market expansion, and even completely new business models that we're so excited to see what, what happens here. So I'm curious, we saw a little bit of it. We saw some general answers. My question to the group here is, do you have a specific use case in mind? I'd love to hear them. Yes. Yes, we have a use case. What is. What is it? Emily, AI Agent, builder. Many. Yes, Joe, this is. We'll get into this in a little bit. We've got some industries. All right, so as people are answering that, I'm going to. Joe said. Yes, Many. We've got two people, two types of people that at boi we really end up talking to. People are targeting one use case. Sometimes it'll be a fuzzy use case. Maybe they have a user, a domain. But we're getting to one use case or they're drowning in use cases. Now I see people are still looking to identify use cases. That's a little bit out of this conversation. But there are ways we can talk about that will help us tie use cases in. Yes, Robby's drowning in use cases. You would think that would be a good problem to have, but often it can really lead to some paralysis. First, let's talk about if you have one use case like Farley here, who's looking for qualitative feedback analysis. If we're evaluating one use case, we really want to try to get it down to 1 KPI if possible. This will help clarify our thinking and really helps us understand, hey, this is what we are trying to solve. There are a lot of reasons for this. Among them is just communicating organizationally or to a broader team, developers to whomever needs to know about this use case. But simplifying really helps in execution and it helps your messaging. For example, if we have AI trend prediction, we might just reduce that to cost reduction. Are there other things that this is good for? Does this lead to maybe new and better products, anticipating the market, understanding new segments? Yes, but the point is we want to keep it, keep it simple here. So the another as just to get you some other examples for AI customer support. Maybe we're keeping our customers better decreasing churn with an AI product designer. We're banking on net new, new sales. We're banking on hey, we're going to be designing products that people want. The point is to keep it simple. For those of you that are drowning, we can use these, the three buckets of cost efficiency, revenue optimization and net revenue growth and try to get an idea of AIs fit for that problem. And we can start to get a little more granular. But the point is we can try to measure apples to apples across these buckets. Hopefully that frame is helpful. And if you're really drowning, like I remember, a couple of us are, I'm all at Board of Innovation. You know, when, when you're looking at early ideas, a lot of times you got to go back in the napkin. And so there's ways you can just sort of score these in order to find out, okay, hey, I have so many use cases that this is helping me find focus on a subset. So another question for the chat. If you have an AI project under your belt, what KPIs were you going after? And this is a cheeky, can be a cheeky difficult question. Time saving revenue growth. Okay, yeah, saving time. So next we're going to kind of we're going to start getting into some different applications. Really cool idea. Sam Altman has said in our lifetime we're going to see a billion dollar company run by one person. Really cool. Really cool. Interesting. I'm getting a feel for the use cases that we're getting now. You can see in all of these different use cases in the chat and all these different KPIs we're going after specifically we can really, we can really see, we can really see that AI is incredibly flexible. I want to get to that point first. The thing that's so interesting about AI is that we're all looking for these use cases, we're looking for these things to solve and it's flexible enough to do that. But what's really important to understand when we think about a billion dollar company run by one person, the, you know, like a statement in pure roi, right, Is that ROI benchmarks are going to change exponentially. And that's something we need to be ready for. And that by think, by thinking about this in advance, we can really help ourselves do that. So let's talk about where optimal investments are because we have this spectrum of options and there's you know, a lot of different business models, a lot of different, a lot of different sort of contexts where you might get that optimal, optimal return. So let's start with an example. If you have a highly competitive or commoditized business, your job, your really business need at a fundamental business model level is that you need to outcompete competitors. So you might, you might start to look towards margin optimization, minimize downtime. As for, if you're facing a brand relevancy issue, it's more about refreshing the image. You want to get more out of those existing revenue streams which might lead to marketing, analytics and content generation. And if you're falling behind a shifting landscape with your products and the products aren't doing the job anymore so you need new ones, you might try to have better insight, synthesis and product and design and testing in order to connect with those customers. Now there's one other, there's another way to think about the, or to think about where to invest. And I'm really interested to hear, I think I've got a sense for it but I'm interested to hear in the chat where you guys are in this, in this, in this process. But by has recognized that AI is going to come in a couple waves and by the way, there's no wrong place to be in this journey. But the first wave is time, cost and efficiency. How can we do today's work better. The second wave is quality and better output. How can we connect with our customers, maybe in new ways? How can we deliver new value to those customers? And then finally is new systems and transformation. So wave three, now we're looking at the, the interoperability, the way different things fit together in order to, you know, in order to make a new, totally new and different way of delivering and capturing value. So investments typically map to one of these, but we can cut across them strategically. And I'll get to that in a second. So if you, you know, we have a bunch of people here who have their use case, they have their KPI. What wave are you guys shooting for? I'm curious to see all three. Look at us go. Okay, interesting. Beautiful. Okay, so what this says to me is that we, we have done a good job as a group here, probably looking down field, looking into the future of where we're going. Because my expectation, to be honest, was that everyone was going to come in and say wave one. And that is. There's no wrong answer here. Like we saw, there are contexts where what you need to be doing is optimizing in terms of cost and efficiencies. But let's talk a little bit about the different impacts of these returns and why we might want to. And why we might want to be looking for wave three impact. So naturally, if you're in wave one, if you're looking at time cost efficiency, you're minimizing AI time to value. It's saying, hey, I'm doing something today, I want to do it better. And I know this will work because I already do it. This can be done within a team. I know someone before said they're looking for AI use cases within their team. You might want to lean this way. Wave 2 has better quality, better output. We're trying to increase that size of return, maybe increase, maybe adding features that help us increase our price one way or the other, trying to get more out of what we already have. This is done with certain functions, have a little more control over that interaction with the consumer. Now when we get to wave three, now we're looking at the way these things live together. And then we sort of introduce this contagion of return. So let's see. I see a question from George. Is it hard to get to wave three without learning from wave one and two? I will say yes. I'm going to say yes. I think that there are foundational transformations that you need to make. And I will, in order to get to those New systems and I'll actually do an example right now. So let's take those three use cases we had before. AI Customer Sort, AI AI Customer Support, AI Trend Prediction, AI Product Designer. Let's say that we have implemented these. Now. What does it mean to get contagious return from these different implementations? If we have AI customer support, you can imagine us more deeply understanding pain points which helps us identify more relevant trends. If these start to talk and live together and then where we start to actually drive this value in our AI product designer, we use that data in order to make better selling products. This is just a simple example from before and showing how these might live together to get us that contagious return. Now, what's really interesting, at least to me, is now when you turn the knobs on each of these, as these improve, you get a systemic return and you get a multiplicative effect, dare I say exponential, on how exactly on the return you're getting from each of these. Now I want to leave you with, and we may have time for questions more formally, I want to leave you with something for those of you that are looking at and evaluating use cases. What does a good use case look like? Ultimately we really need three things to start. Does it solve a user problem? This often? And I don't want to be overly prescriptive, but this is general. I'll say if that you start with a user problem or if you're building a new capability. This often means you're in wave one. It often means you're at least solving some efficiency issues and able to cut across the waves that way. And it's going to help adoption, meaning if you're solving a problem, people will use the thing, use what you build. Number two, if it solves an organizational problem, now we're talking about our roi. Now we're talking about getting that return and then if it supports your business model, sort of those situational things. Now we're thinking about, okay, now we're looking to out compete competitors and frankly if we're not competing, supporting the way we make money and the way we compete, then we might be falling behind. Now there's a couple of bonuses here. Number the first bonus is if you get, if you solve a leadership problem, then you can really get this sign, then you can get some but A budget sign off. But B we, you know, there are ways to build AI tools such that it helps steer the ship and that helps you be strategically agile. Not every use case has that. But it's worth bearing in mind that these Especially with custom solutions, those exist. Then there's the vision for a clear capability, meaning if you're going towards something that wave three, then you're getting faster to contagious return. The last thing I'll say, because I think it's really important and there will be variants in different use cases that you evaluate, is that it needs to activate your data. This is important because it sets the foundation for AI. Yesterday I was moderating a talk and Nigel said that AI transformation is really a data transformation. This is absolutely true. A lot of the SaaS use cases we see that are emerging, for example, AI note takers. What makes them powerful is they start with the data, they already have the data one way or the other. Google Gemini, you know, in their in workspace seems to be promising, but it's because all of your data is already there. If you don't know that your data is ready and you wouldn't be, and you wouldn't bet on it, then it's probably not. This is a big part of why at boi we like to make custom solutions where we have the ability to get our hands on the data. It looks like we won't quite have time for questions. As a host, I know that we get cut off here. I will say please connect with me. This is a topic I love. I am intending to put out a lot more content on this shoot me messages if there's. I'm going to go through the chat, go through the questions, get ideas, put stuff in there or DM me if there's anything that you guys really want to see. Because then, you know, I can, I can put some, put some opinions out there, see some people in the chat so much.

Christian, I loved it. I loved how interactive it was and how you really worked with the crowd. So thank you, thank you.

Thank you.