Choosing the Right Approach to AI

This session will provide a framework for thinking through different kinds of AI that are becoming available for CS leaders, and choosing when to apply them to various projects. We'll explore both internal and external uses of AI in CS as a part of this discussion. You'll come away with a practical structure that will help you decide on the AI approaches that fit your CS program.

Speaker

Kristen Hayer - Founder & CEO at The Success League

LinkedIn

Kristen Hayer founded The Success League in 2015 and currently serves as the company's CEO. Over the past 25 years Kristen has been a success, sales, and marketing executive, primarily working with growth-stage tech companies, and leading several award-winning customer success teams. She has written over 100 articles on customer success, and is the host of the Innovations in Leadership podcast on Success League Radio. Kristen serves on the boards of the Customer Success Leadership Network, the Women in Leadership Program at UC Santa Barbara, and several early-stage companies. She received her MBA from the University of Washington in Seattle, and now lives in San Francisco.

Watch Choosing the Right Approach to AI

‍Transcript

This transcript was created by AI. If you see any mistakes, please let us know at marketing@matik.io.

Matik MC: Hi, everyone. Welcome to Matik, AI and customer success summit. You are currently watching, choosing the right approach to AI. We're going to give everyone just a couple more seconds to just join the webinar before we go ahead and get started, but very excited for the session.

Matik MC: Again, this is the choosing the right approach to AI session.

Matik MC: Okay. So, I am very pleased to introduce our speaker. And as I introduce her, I'm gonna have her answer our ice breaker question, which is what would you use AI to automate in your life outside of work? So, this is Kristen Hayer, founder and CEO of the success league.

Kristen Hayer: Hi, everybody. My goodness. What wouldn't I automate? But I think, my most hated task is grocery shopping and what I wish there was a tool where I could just put my recipes. They would figure out what I had in my cupboard somehow magically and it would go get all the things I don't have at the grocery store for me. That is what I want.

Matik MC: The only thing missing from this is just the robot that actually does the cooking tail.

Kristen Hayer: Well, I like cooking. I just don't like shopping. So I don't wanna go to the grocery store, but I want all the like, right? Ingredients delivered to me. And now, some people say, well, couldn't your husband go do that, but he buys the wrong brands of things. No, he cannot.

Matik MC: Great answer. Just a reminder for everyone before we start the session. If you have any questions for Kristen, go ahead and use the Q and a which you can find in the Zoom control bar, just click Q and a and drop your question in there. So I'll let Kristen take it away now.

Kristen Hayer: Okay. Well, I'm gonna share my screen and so give me one second here and I will get the slides rolling. All right. So today, what we're going to cover is the different types of AI that are relevant to CS. And I know there have been a lot of sessions that have covered this in more depth. So I'm not planning to cover this in a lot of depth, just gonna kind of set the stage with that slide. I'm gonna talk about what AI can do for CS. So kind of paint some pictures for you and give you some ideas of how this can be used, but really what we want to get to is when should you leverage AI and how should you think about that? I think it's really exciting to have this new technology available to us. There have been a lot of developments and changes. Obviously in the last few years around artificial intelligence. There's so much potential there and it can be really hard to not just go get the shiny cool thing. But we really should sort of stop and think about where do we want to apply? AI, and I have a framework that I'll be sharing with you so that you can start to think about how that applies to your customer success program, that's what we're kind of leading up to. And then I'm gonna leave some time at the end to answer questions that you might have on this topic. So, and I'm really excited to be presenting it. So who the heck am I? I'm Kristin here, I'm the founder and CEO of the success league. Our company is a global consulting and training firm, and we are focused on customer success. We've been around since 2015 and in that time we've had many hundreds of companies go through our consulting practice and we've had thousands of people attend our training classes on everything from becoming a certified CSM to being a CS leader to thinking about cross functional leadership and things like that. A little bit about me outside of now, you know, that I hate grocery shopping. I have a college kid. Her name is Jamie and she goes to you San Diego. I love visiting them there and they are my summer intern this summer. So it's been really fun. I also love dogs. I have two German shepherds. Our youngest one is seven months old and has a tic tac account. So if you're into to talk to, you can go check them out. He's at handsome Rob GST. We named them after handsome Rob from the movie the Italian job, and you can get, you can watch him get taken down by his bigger sister. Who's a six year old German Shepherd? So they like to fight. And then lastly, I really love CS. I, this is a field that I have focused on for the last, you know, probably 15 years of my career but very intensely since I founded the company over the past eight years. And so feel free, to connect with me on LinkedIn. My company is better at using LinkedIn than I am. So you should also connect with the success league if you want articles and info about our field. And I'd love to connect with any of you. So let's start with what you know, what types of AI are relevant to CS. We've all been hearing a lot about generative AI lately. This is AI that creates new content from a massive pool of data inputs and spits out things like writing or videos or images, you know, in our field that's really things like that. But I, which would be been getting increasingly more and more human like in terms of how they respond to the people that are engaging with them. So that's kind of an example in our field of generative AI. Then we have natural language processing. So this is enabling machines to understand a, respond to text or voice data. Amazon uses this to improve their customer experience by analyzing the reviews that they get on their site. So that's an example of how you could use natural language processing in a customer success environment. There are expert systems. These are tools where the AI learns and reciprocates the decision making abilities of a human. We've been using this in CS for a while with knowledge base routing and the development of knowledge base articles. So that's kind of an example of an expert system. And then the other one that I think is really relevant to CS is the fuzzy logic type of system. So these are mathematical models that are not just binary a, but they take into account gray areas in human decision making. So an example of a fuzzy logic system in our field would be like an eCommerce company using that to route customers through websites in a way that maximizes their spend. So all of these kinds of air are relevant to CS. I think what you choose is going to depend on what becomes available as you all know, lots and lots of different artificial intelligence tools are becoming available on the market. And so while right now the bulk of them are focused on generative, AI think you're gonna start to see systems that are focused in some of these other areas as well in the future. So it's good for you to keep an eye on all of those there's lots. Of articles online. I'm gonna mention a few at the end of the talk today, I, that you can go and do a little more research on these different types of artificial intelligence. So if you wanna go crazy and learn more, there's a lot of tools out there for you to be able to do that. But I'm gonna leave it at that for now. I think there's sort of two things to consider when you're thinking about AI for CS. One is that we can use this for internal tools. So this is things like, you know, behavior based health scoring or strategic program adjustments, journey pathing, but this can also be for external tools. So e-mail writing, slide deck design. And I just wanna kinda paint a picture for you of a couple of ways that this can really play into a typical customer success program. So imagine if you had a health score that utilize data not just from a small set of factors that you plugged into the tool that you're using but across a wide range of factors including customer behavior factors. And imagine that health score then could be communicated to your CSM in a way that makes it very clear why the customer is healthy or risky instead of just sort of leaving it to chance or kind of basing it on one factor. But in natural language, explained what was going on with that customer to the CSM. Then instead of your CSM needing to research what's going on or try to decipher what the score really is and why they spend their time taking the best actions to help expand or save that customer relationship. That's a much better use of their time. Another example is imagine that you didn't need to spend half your day pulling together a deck for a business review. Instead, you have a tool that pulls the deck together for you and includes only the data that's really relevant and most interesting to that particular customer. And imagine how much more strategic than your conversation could be if you didn't need to Wade through a bunch of slides that aren't relevant, but instead had slides, that were focused on the value that customers could get from making the most of your solution. And then this last example is when I love and it just came up yesterday. So, I wasn't really planning to include this in the presentation but it was so great. I have to share it with you. I was teaching a class on proving a return on investment to your customers. And as a part of that class, we had a workshop and the homework that I gave to the students was that I wanted them to create a set of goals. For a particular customer that they work with and I wanted them to have a conversation with that customer about the goals and their values and what they wanted to get out of the solution. And then to turn that into a goal plan with the customer. And I have a little table template that I share with people that they can use to do that. We had then three people volunteer to share their examples with the rest of the group. And the first two people, they did it manually, and, their goal plans were good. There were a few things that they could have changed. But, you know, overall really good results. The last guy that presented was just, but he's like I hate filling out tables and charts. And so what I did was I used AI to take the text of the conversation that I had with the client that was strategic and turn it into goals for me. And he fed my example and form into his AI tool and it spit out a five point goal plan for him that he had to make a couple of minor tweaks to. But overall, it was fantastic and much closer to what I was looking for from the group than anybody else's and he didn't have to spend the time filling out the form he got to spend his time having the strategic conversation with the client and it's one of the most creative uses of, I've seen recently and I thought it was fantastic. So I wanted to make sure I shared that with you today. So those are some things that you could think about as you're thinking about artificial intelligence for customer success. There's some problems with a. I, and I know some of the group discussions today have focused on this, but I wanna make sure I cover it with all of you who are sitting in on this session. You know, accuracy is a big deal and garbage in garbage out. So the quality and quantity of the data that's driving your tool has a big impact on its accuracy. And one of the things I noticed from working with a lot of CS teams is sometimes we struggle with data availability and cleanliness. And so… I think you need to make sure that you have the data that you need to provide an accurate result before you go down the road of a, and depending on your project that may be harder or more easy. The other thing that somebody mentioned, I was having a conversation earlier today with somebody about artificial intelligence and what they mentioned they do to help companies with this accuracy problem is have them think about building in kind of a traffic cop layer above the core. I just to keep the AI on the rails. So it doesn't kinda go. Off in a weird direction, another challenge with AI is by us and this is around the sources and where the sources are coming from and are they skewed toward particular groups of people? I think in general, a lot of discussion around artificial intelligence has been around the fact that tech is driven largely by a white male audience. And so it's missing a lot of group groups of people, you know, in CS. I think this would be less likely to happen around demographic groups. But I think you can think about this in terms of customer groups. So if you're using data and it skews toward one group of your customers, you're going to be missing out on potentially other segments of your customer base and your data won't be as accurate and as unbiased as you need it to be. And then finally illegalities is a big concern. And I think here there's just a lot of controversy around especially copyright infringement and artists, right? I, in customer success, I think there's also a concern around confidentiality, especially if you're working with clients who are in regulated industries like government or healthcare. You have to give thought to how you use AI and how you use customer data in a way that doesn't accidentally share confidential information with customers at large. So there's some risk around that. I think the other thing to keep in mind and this is for just all of us as we think about and build AI tools, especially generative AI tools. Artists make a lot less than all of us do in the tech industry. And it's really important for us to support artists and artists who are not making money on their art, can't afford to live and they're gonna have to make the tough choice to not be artists. And I think that lets our whole society down. And so it's something that we, you know, probably don't have to deal with as much in customer success, but I think it's something we should all have in mind as we think about AI… with that. I wanna go to the model about how to think about where to leverage AI inside of your organization. So I think you wanna think about cost. I think you wanna. And in there, you know, if you're thinking about something that you can't afford to spend a lot on, you may already have tools that have some component or level of artificial intelligence built in a lot of BI tools… you know, have some things built in. Some of the CS platforms have things built in. So just consider what you have if you don't have money to spend, if you're trying, to tackle something where you need your solution to be predictive. There's going to be probably a medium cost associated with that. So those would be things like industry specific tools like Matik, and there's a lot of other tools that are made for sales and CS team out there. But those tools are going to have kind of a more medium price point and are probably going to be sufficient for trying to predict or trying to produce content that isn't trying to solve major problems when you move to being prescriptive. So asking a system or having a project where you need the system to tell you what should I do next. Then the cost is probably gonna be pretty high because there's going to be significant customization that you need to do. A, you're going to need to make sure your input is really high quality and there's cost to that. So, you know, as you think down this chart from descriptive, so, you know, a tool that tells you what happened in the past… that's going to be lower cost. And as you move up the scale to prescriptive, the cost is going to get increasingly higher and higher. So keep that in mind and.

Matik MC: Then you…

Kristen Hayer: Need to decide is what I'm trying to do worth the cost? I think, you also want to consider, what are you trying to ask this tool to do? What is the role of AI in this project? Do you want something that helps you to understand what's going on or do you want something to support a decision that you're making or do you want something to tell that tells you what to do? And, and so you can kind of think about that spectrum as well. I think it's important to think of the size of the opportunity. So if what you're trying to do is you're trying to solve a one time problem, you know, maybe run a report and figure out what's going on that's not necessarily something you need to leverage AI for. Whereas if you're trying to increase revenue or save money across a broad swath of your customer base, and maybe even on a repeated basis that's something that it may be worth spending the money on because there's a large opportunity there. So you really wanna think as you're building out your business case, what are you asking this tool to do? What's the size of the opportunity? And, you know, what is the cost? I, I've got some examples, these slides are going to be available later. I wanna leave enough time for Q and a here. But AI is really exciting and it can be tempting to want to apply to it to everything because it's really cool and shiny right now. And this framework is really to help you think about when to apply it and when it might not make sense. So you wanna make sure there's always a compelling business reason for purchasing technology. And rather than by first use later, you wanna think about the use cases for technology. And this is a tool to help you think through that process. I wanna note before we got a question that there are some articles available here. I also found there's a really interesting MIT course. If you've got a little bit of budget to spend on training on AI and the implications for business strategy. So it's about 3,600 bucks. So it's not a cheap course, but it looks really good. So I'll let you guys play with these resources. And with that, I'll open it up for any questions that you have.

Matik MC: Thanks so much Kristen. So we do have a couple of questions. Okay. How do I get by in cross functionally on using AI, especially when a lot of its impact has not yet been seen or is hypothetical?

Kristen Hayer: So, I think this comes back to like thinking about your business case, right? I think you have to sort of put some estimates down and it can be, it can be scary to do that if you've never done that before, but you can use some of your past data to estimate, you know, how your use of AI will impact things in the future. I do think you wanna think about depending on what it is you wanna do. So let's say you want to try to identify some areas where you could potentially grow customers. So you wanna do some expansion work. You can, you can use AI to figure out who's ready for expansion and that would save your team time which has a cost savings benefit to it. You could also estimate how much money you think that could generate for you. And you wanna be somewhat conservative, probably if you've never used it before, but, you know, you can kind of estimate how much money could come in. So you've got money being saved and money being generated and you wanna think about both of those things. Then I think you wanna think about and, this is a great way to get leadership and your peers bought in on this. Try a test. You don't have to run this across every customer you have, which can be risky depending on what it is you're trying to do, you can test it on smaller groups. And most of the organizations that are, you know, bringing tools into the artificial intelligence space right now are fairly new themselves and are probably going to be willing to do a Pilot with you so that you have the opportunity to test it out for part of your customer base. And I think if you can pitch, that softens the cost of whatever solution it is. You're using it, softens the risk and it makes it much more likely that your leadership or your peers will see the balance between the potential benefit to the organization and the cost.

Matik MC: What part of the customer journey do you see AI having the most impact on?

Kristen Hayer: Gosh it could have, there's so many ideas I have for this. I think it could, it can benefit all of the parts of the customer journey. I think where I see it having the most benefit to people and by people, I mean, both the customers and the internal CSM folks is in higher touch programs. I think we often think about artificial intelligence as something that you wanna apply to your digital program or to your long tail customers or whatever. But I think that, the savings from focusing CS professionals on strategic work… allows those professionals to serve more customers, which is a cost savings for the company. It allows those conversations to be more fulfilling and interesting for both the customer and the CSM them, which has a qualitative benefit that's huge. I think a lot of programs, the CSM never really get around to having strategic conversations because their whole day is bogged down with trying to fix problems for customers. And if some of those problems were handled by artificial intelligence, then you're applying a CSM to positive things that are going to demonstrate value to customers. And that's the whole point of customer success. So, I think it's less about where do you apply it in the customer journey. In my mind, it's more about how do the people and the customers benefit the most from the tools that are available? And, I think if you're creative, you could come up with all kinds of interesting places to apply it across the entire customer journey. So, yeah.

Matik MC: The next question has two parts.

Kristen Hayer: Okay. First, okay.

Matik MC: Curious about how to handle security buying et cetera. Especially for small businesses who have less resources and risk they may be able to take on, for example, how to cut through the noise and find good tools especially with so many. And the ease in which given these tools, any tool can be made to look legitimate.

Kristen Hayer: Okay. So this is what tools like G2 and the other rating tools that are out there in the market and Gartner and Forrester are for like there's the Forrester wave there's the Gartner quadrant. Everybody's ranking the tools that are out there in artificial intelligence. I think that's a really important way to look for something that's going to be the right tool for you. That said the newer players in any space are not necessarily always gonna show up on G2 with thousands of reviews because they haven't been out there long enough to generate them. So just because their, you know, their reviews are not as plentiful as other companies because it's a new space, I think you wanna look at the reviews themselves and see what they say. And that to me in a space like this that's evolving so rapidly is much more important than that they have thousands and thousands of. So, yeah. What was the second part?

Matik MC: Second part? Are you ready for this? What are your thoughts? What are your thoughts on some of the advancements and things like the zero retention and anonymization layers of some models such as what Salesforce is doing with it?

Kristen Hayer: I don't know much about that. I'll admit it. So I'm probably not the best one to give an opinion on that. I think that people are being very clever with solutions all across customer success right now. I do think that the best way to be careful. And I actually was just interviewing somebody for our podcast this morning and we were talking about AI and I was bringing up all these, you know, like what about security? What about, you know, how do you make sure, your artificial intelligence tool is staying on the rails and that your accuracy is there? And it's not just looking at all of human knowledge because it got fed this huge data lake or ocean or, and it's looking at everything and just kind of pulling stuff together because one of the problems with AI is it can make stuff up and it sounds very true and real. And so, you know, I think you have to, you do have to be careful about that. And what he suggested and he's a PHD level, you know, data science guy. He said you can have a layer that sits above the AI that is sort of like the traffic cop for it, that keeps it in the rails. And I think that that's probably the easiest and best explanation I've heard on how to keep your, how to keep AI in line. So, yeah, that's my answer. Yeah, that program specifically, I don't know a whole lot about it. So I can't comment.

Matik MC: And then I think we have time for one last question, right? What do you see as the biggest risk with AI for?

Kristen Hayer: I think that we need to make sure that we don't… lose the humanity of it, of our interactions with customers. You know, you can't always afford to have, you know, a person assigned to every account that's just the reality that we're all dealing with, but you can still have systems that feel very human even if they are run by AI and that's improving and approving. I think chat bots are a great example of this over time that has improved dramatically. But there still should be a person looking at these programs holistically and thinking about what the human experience is like because yes, we have accounts, but we also have people who are customers. And sometimes we start to forget that we work with people not companies. And so if you're not looking holistically at your programs and thinking about how people engage, I think you run the risk of having a experience for your customers that feels very cold.

Matik MC: And with that, we are at the end of our session as a reminder to everyone if you missed the beginning or whatnot, we will have the recording of this session available once the hub goes live. Thank you so much, Kristen, for such a wonderful and educational session, and thank you to everyone else for attending.

Kristen Hayer: Thank you so much. I appreciate it.

---

Click here to see all sessions & speakers, and navigate to recordings.