00:00 Intro Welcome to the Player Engage podcast where we dive into the biggest challenges, technologies, trends, and best practices for creating unforgettable player experiences. Player Engage is brought to you as a collaboration between Keyword Studios and HelpShift. Here is your host, Greg Posner.
00:15 Greg Posner Hey everybody, welcome to the Player Engage podcast. Today we’re joined by the head of product at HelpShift, Eric Ashby. This is a special episode of the podcast where we’re going to really focus on the one technology we keep hearing all of our guests talk about over and over again, which is generativeAI. We find ourselves at the forefront of an extraordinary leap in technological evolution where distinguishing true innovators versus the mere imitators is paramount. The question arises, how are we going to identify and harness these emerging technologies of the early stages of the hype cycle? Together, let’s explore the intricate art of differentiating genuine game changers from fleeting trends, gain the knowledge of expertise required to identify and leverage these technologies at their inception, propelling your organization to the forefront of innovation and reaping the numerous benefits that await. Prepare to embark on an exhilarating journey and discovery to secure your position as a visionary leader in every evolving landscape of generativeAI. Eric, that was really more of an intro to the show and not so much about you, but do you want to give our listeners kind of an idea
01:20 Erik Ashby of who you are and where you’re coming from? Well, I mean, I’ve, you know, I’m Eric Ashby. I’ve, first off, I’ve been a longtime gamer. I was thinking about, you know, when I was thinking about this podcast, when I got into game was actually my father and I, when I was a kid, back in the 80s, we owned, you know, some of the, there was, what do you call them? They were the arcades that were going on. And we actually supplied arcades with gaming cabinets. And so early Donkey Kong, you know, Ms. Pac-Man, all of those, those were the things that I just started playing because we were, you know, we would supply the arcades with that. But yeah, I’ve been actually been in the industry for quite some time. I worked at Microsoft for about 20 years. And then I moved over to HelpShift about six, seven years ago. And we’ve dove into AI soon after I moved on over to HelpShift. And so this is really kind of a topic that’s near and dear to my heart. There, you know, there’s been a lot of fads that have come and gone. We’re kind of going through a new fad right now. So super happy to talk about, you know, talk about this space because I think it’s a really, really interesting space and glad
02:28 Greg Posner it’s kind of getting the visibility that it’s long deserved. So yeah, and I agree with you. You know, we, we, a few years back, we started entering the world of the metaverse and what’s the metaverse going to be and still trying to work its way out. And now a new technology comes out, which is a generativeAI. And it’s funny because each of the podcast guests that we have on, we talk about what trends are they most excited about. And this is always the top answer. And I could see why we’re all mostly using it in our day to day jobs. I think the people that are using it correctly, it becomes a seamless integration that’s part of your workflow and works well for you. But it’s also very early and things are going to change. And we’re living in a world where technology changes quickly. Companies want to be the first to launch a product and gain that first movers advantage. We’ve seen this work in historically, right? We look at things like aim, which I’m really aiding dating myself talking about that one, Netflix, Uber, all these first movers gained advantage in the marketplace and really held onto that for years. But we’ve also seen this backfire, right? People might say maybe something like my space, although my space did stay around for a few years, they lost out to Facebook, Blockbuster and Kodak. So let’s start simple, right? We’re thinking about these companies that are making these first mover advantages here. But all these companies are releasing products. It’s like a whole bunch of things hitting the market at first. How do you keep up with this technology? What are your sources
03:45 Erik Ashby of news? How do you follow what’s exciting? Well, I mean, for me, there’s first a couple ways is just kind of kind of who I am. I am I’m an explorer have always been an explorer. So whenever I hear about some new technology or something like that, I tend to ignore kind of the news and stuff like that. And I and I just dive into it. You know, when you know, good example, when OpenEye came out, there’s a ton of hype about like, Oh, my gosh, what is this? What’s it going to do? And it was really just diving into it and jumping into it. And, you know, you talked about some of these old older technologies that, you know, you know, come and gone. You know, for me personally, it’s really about just getting into it and understanding kind of what they’ve done, what their approach was. And I’ve found what I’ve typically found is, is especially when this new emerging technology is that there’s always hype, like, oh, like, look at this thing. That’s new. There’s always hype. But the ones that tend to stay around are the ones that can bypass that hype and actually really solve a problem. You know, they can really do something. And I’ll give you I really give you a good example. You know, you brought up Uber, and Uber, of course, has AI that’s that’s all built into it. And there’d been a ton of technology that had had come out over the years in the transportation space, there’s a ton that would come out and what Uber then said, Look, it’s not about the technology. It’s about solving a problem. You know, how do we actually use this technology to get somebody on the corner of a street to their house in as short a time as possible. And when the brand focuses less on the technology and more on the solution of like, what is it that you’re trying to solve? That’s when it becomes a game changer. And those are those are the brands that I’ve seen that that have kind of gone past the hype, you know, because they’re not it’s not about the technology. It’s about it’s about the solution. And that’s kind of what I’m seeing with AI, AI today is I’m seeing, you know, a lot of brands catch up again, and be like, Whoa, like this, we’re using chat GPT or whatever. And I’m seeing other brands who’s actually picking up and saying, Well, actually, you know, this is how we’re gonna use it to solve a problem that that’s actually
05:47 Greg Posner meaningful. And maybe they maybe don’t even recognize it. It’s an interesting point. So I think the first thing you’re really talking about is understanding the use case. And we talk about use cases a lot. And it’s part of the day to day for for most of us that are on these calls and doing this thing. But right, you can’t just create a new technology and force a problem upon someone and say, Hey, use our technology to solve this newfound problem that we helped you create, right? You need to really talk to people understand their workflows, right? I think a lot of people don’t know what they’re going to use generativeAI for. But the more they talk it out with people, the more they see how others are using it, you start to formulate this idea rather than again, just saying, here’s this new problem we’ve created, and you can solve it with our own
06:28 Erik Ashby tool. Is that kind of sound accurate? Yeah, and that’s what and that’s, you know, often I see, especially when kind of this new hype comes out, you know, a lot of brands will jump in just because they’re trying to get the you know, the eyes and the views of like, wow, we’re using this new technology, and they’re kind of missing the point, you know, the point is really about, about about solving solving the problem that you’re you’re set forth to solve. And it left the technology take take the back the background. And in the end, like, you know, imagine two, three, four years down the road, you know, everyone’s talking about like, generativeAI right now. But it generativeAI will be a part of our everyday work. It’ll just be there, you know, when I when I go to generate a document, or I go to write something, or I go to write to code or whatever, generativeAI is just gonna sit beside me. And it’s going to be like, okay, like, here’s what we recommend, or here, I see you make a mistake, because it’s not what I would have done, or things like that, this is going to be there. And it’s not going to be like, blatantly in front of our face, like, oh, we’re using generativeAI is just, it’s just part of what we do as part of the solution. And again, that’s what where I see that brands, brands, you know, have really been successful. I’ll use another example, another one that I think is is really good. I think, you know, everybody knows Google Maps, we all just use Google Maps. And of course, Google Maps is backed by AI, it has a lot of AI that’s in it, you know, when you open it up, it knows that it’s dinner time, and it knows that most people open up Google Maps at dinner time, and so it’s going to show you the restaurants that are right there. But it doesn’t, it’s not blatant. It’s like, oh, look, I’m using AI right in front of it. It’s just, it’s just part of the solution. It’s just embedded into into that, that toolset. And so you don’t even recognize it. It’s just like, oh, yeah, I want to go to dinner. Boom, done. And you go. That’s where to me, the magic is, is when you don’t actually recognize that it’s there. And that and that’s really where it kind of makes me excited about this, this new technology as it’s
08:21 Greg Posner coming out is, is when it gets used in those in those cases. It’s funny you say that because I’m going back to one of the podcasts we did with Boba and customer of help shows at Atlas reality. And we were talking about the metaverse. And he said the best implementations of the metaverse are those that are running on games and people have no idea that they’re even using the metaverse. The idea is let this technology run in the background and let them be using this new technology. But you don’t need to announce it. Let’s let it happen. And they’ll see and it seems like this is a more of a common theme is, yeah, it’s exciting to say you’re going to be the first one using chat GPT. All of a sudden our tools are using chat GPT. But that’s really just make up on a pig at the end of the day. You really want to be able to implement something in the back end of the tool that’s working. And it’s almost seamless to the user’s day to day job,
09:05 Erik Ashby except maybe the benefits that they’re getting from it. Yeah. Yeah. Yeah. And that’s, and that’s, that’s the whole point. I mean, there is a good reason for the hype. I mean, you know, AI has, has moved very quickly recently. And you know, like people were asking me last year about the Metaverse since you brought up there, like, well, is the Metaverse going to take over and things like that? And the Metaverse is definitely great, interesting. There’s a lot of technology that’s coming out. But, but when I think about the kind of the next generation of generativeAI, you know, where I’m talking about, you know, like the mid journey and chat GPT and things like that, it has a couple of things that are really going for it right now that is kind of driving the hope. The one thing is, is it’s introducing some new scenarios. You know, people are like, oh, wow, this thing can actually sit beside me and, and help me out. It can actually generate some content. It can help me write something. It can, you know, save me time. It makes my life better. And so it actually has a use case. That’s, that’s interesting. And the second thing is the barrier to entry is super low. You know, like, like Metaverse still has a pretty high barrier to entry, especially if you’re talking about Metaverse experience, where you have to put the thing on and things like that. But for, for some of these generativeAI, it’s as simple as human language. You know, I type and I ask you the question and it’s going to respond and it’s there. And so, so when you have, when you have a good value proposition, you have low barrier to entry, then you’re going to see, you know, kind of the explosion that comes out. It’s kind of that, that inflection point where the technology has really kind of come along where you’ve seen that. But with that hype, then comes everybody jumping in, trying to use it for everything. And that’s, of course, then why you have the question at the beginning is like, what’s hype,
10:41 Greg Posner what’s not. You’ve been in the industry for quite a bit of time. You’ve been at HelpShift for almost 10 years. You were at Microsoft for quite a bit of time. Historically, have you seen a trend like this before? Do you remember kind of a previous experience where you’re seeing this,
10:53 Erik Ashby whether it be web two or something else and have like similarities to that? Well, I mean, I, I don’t know. I, you know, if you take a look at usefulness and barrier to entry, take, take a look at those two, I would definitely go back to the, you know, the mobile phone revolution or, or whatever. I mean, mobile phones, nothing new. You know, we had these bricks that were that started, I remember when my dad brought the first one home in the, in the eighties and he’s like, ah, look, I’ve just spent a thousand dollars for this brick, you know, we’re like, what is it? You know, well, like I can call it by any time. It was, it was funny. And then, you know, in the, in the nineties, you know, Microsoft and, and I came out and of course there was the, the crack berries, you know, the black berries that came out. Blackberries were interesting because they started to get it. You know, they, they’re like, it, it solved the problem. The barrier to entry was still a little bit too high. Then came the iPhone where they just said, look, we’re just going to reduce the barrier to entry, make it super simple for everybody to use. And we’re going to make it super powerful. And then all of a sudden it just exploded at that point. You know, you take those two things, you make it useful, remove the barrier to entry. You know, it was all about solving problems. Like they took all of the buttons off the screen. They’re just like, let’s solve the core problems. And when we do that, then it makes it really, it makes it really interesting. And that, and that’s what, of course, that’s what they’re attempting to do. And so I do, I do see AI honestly is kind of going to be embedded in our lives, you know, just to the extent like mobile is, and it’s just, it’s just there.
12:22 Greg Posner It’s just, it’s just there. Yeah. I personally am enjoying for the most part, right? There’s these weird nuances where they’re implementing these tools to Gmail or to other email programs, to auto-generate an email. And we had someone I was talking to the other day who said they realized that they got an AI generated email back from their friend. And they were kind of pissed off about that because if you’re not going to take the time to respond to your friend, like there becomes this fine line. It’s like this black mirror episode that’s coming to be as what’s real these days, what’s not who you’re going to give that full-time effort to. And I think you start to learn how did they use these tools properly, right? You don’t let ChatGPT or any AI tool write the complete email, but maybe you use it as a starting point to generate some ideas and then
13:06 Erik Ashby you pick and pull what you want. Well, and, and, you know, there’s, there’s an evolution that’s going on with AI. I’ll just kind of quickly walk you through how, how I perceive, you know, there, there was a world where AI didn’t exist. You know, it just, it just didn’t exist. And, you know, we all did everything by ourselves. And AI has been creeping into our world as kind of these background assistants for quite some time. We just, you know, we never really recognized, I mean, it’s very simple AI. Spell checker is AI. It recognizes patterns and, and, and makes a suggestion to you. But most of the time AI was in the background, kind of recommending to a human. Okay. And it’s kind of moved more and more foreground. And now it’s kind of to the point where AI is almost like a coworker to us. It’s like somebody who is kind of with us all the time and we can ask it questions. It can respond back or it’s like, Hey, could you, could I write that letter for you or things like that? And so we’re in a different world now where we need to understand like, just like if I was, had you as a coworker who was with me all the time, there’d be some things that I’d be like, okay, you know what? I’m going to take this. This is, this is probably better for me to, to respond to or to deal with and, and things like that. But that’s, that’s really what’s happening. And what I perceive is that as AI gets smarter and better, there’ll be places where we then step back a little bit and we’ll be like, okay, look, you take this and I’m going to monitor you as you, as you do whatever this, this workload is. But that’s kind of how it’s been evolving. And of course us as human beings, we just need to, you know, we need to understand and, and, you know, and how does that, how does that work for us? And so it gets into a really interesting question then is like, you know, what is, what is AI really good at? What can it do? And that’s one of the things that I’ve been, you know, been looking at. It’s like, okay, what, what makes AI really powerful? Cause you have a lot of AI that’s out there today that can do a lot of things, you know, and, and it’s like, okay, AI can answer questions, but how do you make, actually make it so it’s like capable of doing something? And, and that’s where you have to actually add two other things to AI. And if you, and when these three things come together, you get extremely powerful. So you, you have to have an AI, which has the ability to understand the world, answer questions, you know, be able to talk, you know, things like that. And what we’re saying is we’re seeing AI to get to that point. You then have to add to that AI specialist knowledge. You have to say, okay, AI, you know, I know that, you know, you know, about the world, you know, about things, but let me teach you about this one specific area. Let me make you really, really good at this one area. Just like if you were to hire somebody from Harvard, super smart, you bring them into your company and it’d be like, Hey, let me just, let me spend some time and really get you to understand this, you know, whether it’s about the game or about the system or about my company or whatever. And then the third thing you add to it is you add to it capabilities. Like let me give you the ability to actually do things. You know, maybe I have you connected to a system so you can not do refunds or things like that. But you add those three things together. You give a strong AI plus training plus capabilities. Now you actually have a really interesting system because you’re able to actually not just answer questions, but do something with it. And that’s kind of where I see AI is kind of going next is you have these specialists, these AIs that actually have this capability. You know, if you talk about the gaming, you know, the gaming arena that we’re in is, I’m super excited about kind of some of this because I’m going to see these AIs that will pop up and brands are going to do that. They’re going to give that, you know, they’ll take kind of the general GPT model or whatever model they start with. They’re going to then train it with its in-game or in-app specialities. And then they’ll give it some power. They’ll say, these are the five tasks that you can go do and whatever. And now you’ve got, you essentially have a specialist coworker who can actually go do work for you much more than just, oh, I can respond to some. So that’s kind of where I see it going. Again, we need to manage it as we would any employee.
17:09 Greg Posner We need to manage them. So it’s very interesting. Does it become a problem, especially with some of these three pillars, right? Your AI, your specialist knowledge and your capabilities when you have companies like Google, or you have companies that have all this information about everyone, right? Clearly they’re going to have the leg up, right? Like I’m going to think big picture here and I apologize because my thoughts are going to be random and all over the place. OpenAI was the first one to really publicize, maybe not fully, but right? But like they’re a smaller company. They beat Google to the punch. They beat all these other companies to the punch with a publicly facing working tool that was nice to use. But how do small companies with these, with that lack, maybe specialist knowledge, like are they at a disadvantage? Is this a bad path for companies to start going down?
17:55 Erik Ashby Yeah, I don’t think so. You know, when I look at, when I look at these big companies like OpenAI and Google and things like that, they will always only be able to do one of the three pillars. Okay. Which is, you know, they’ll be able to be that generalist. Okay. They’ll be able to take the AI and like train it about the world, which is great. Yeah. I mean, I mean, let’s compare IBM. You know, IBM came out with Watson a while ago, which was, you know, was supposed to be this, this great AI. Really what Watson was, Watson was the neural network, but it was like a kindergartner. Like it would come in and just knew nothing of the world. And so you had to spend all your time just teaching it about the world before you could teach it about your company. You know, and so it would cost you several million dollars just to get, you know, a Watson up because it’s like, it’s like a kid. You had to like train them through whatever. So, so now when I look at Google and I look at OpenAI, it’s like they’re giving us a, you know, a well-seasoned human, you know, I won’t say human being, but artificial intelligence, a well-seasoned artificial intelligence that has been taught in the ways of, of the world and language and understands kind of general things that are going on. Knows how to code, knows how to put pictures together, blah, blah, blah, blah, things like that. But it doesn’t know about you or your company or your game or whatever. And so I wouldn’t, and I personally wouldn’t be sharing those things out. I would, I would take, I would take their, I would take the general AI and I would add it to my knowledge, but I would keep it to me. And that’s where I’d take the difference, make the difference is that it would become a specialist based on, you know, the generalists and Google and all of them. They’re trying to solve the generalists, which is by the way, the hard, a very hard problem. So they should be solving that, you know, like a small company should not be trying to solve, like let OpenAI and Google solve the generalist problem. And then we build on top of it, which is to me, is a very practical approach to it. You know, you start with someone who is smart and you add on to them specialist knowledge, give them core capabilities. Now you’ve got something very
19:54 Greg Posner powerful. I don’t know if you saw, I think it was yesterday, two days ago, a company in India got rid of 80% of their staff, replaced them with chat bots. They were a customer service based company. So do you plan to, obviously they do that and they get their name in the news. So that’s good, whether it’s good or bad for everyone. Right. But do you foresee that coming to more
20:13 Erik Ashby and more companies willing to take a chance like that? Yeah. I mean, there’s, you know, like, there’s no question about it. AI is going to make us more effective. It’s going to make us more efficient. Companies that flat out go and do that. To me, what I feel like is that they were having their humans do robotic tasks to start with. I mean, that’s what they’re doing is they’re just saying, look, great. We have this, you know, we have these people that are just out there answering questions. I’m just going to go get an automation that can answer questions. Humans are empathetic. Humans are creative. AI can pretend to be creative, but humans are really the creative ones. So humans are empathetic. Humans are creative. Humans understand priority. Humans understand how to problem solve all of these things. That’s where humans should be. That’s where we work best. And so the way that I see, you know, brands really working forward, you know, working together is that, is when it’s AI and humans together, that’s the most powerful one. And so if, you know, if a company came in and said, oh, we just got rid of half our guys, I’d be like, sounds like you were treating them as, you know, as robots to start with. And, and I, you know, it makes me question what they were doing. But, but yeah, it’s, you know, we’ll see definitely things change, but brands that understand how to combine the AI with the human power, the brands that are going to succeed. And I, and I firmly believe that it’s not, it’s like, your job’s not going to be replaced by AI. Your job’s going to be replaced by someone else who’s
21:39 Greg Posner using AI. And that’s, you know, and so, or your app or your whatever, it’s, it’s going to be replaced by someone else who’s figured out how to, how to implement AI in it. Your role as the head of product, forget the company, right? It doesn’t matter, right? When a new technology like this comes up on your radar, at what point does, does something go off in your mind saying, I need to start looking at this technology and start thinking about how to implement
22:02 Erik Ashby it? Yeah, that’s interesting because it’s kind of hard right now for me on this question, because we thought about this about, you know, six, seven years ago when it was initially, when it was initially coming. Again, the thing that the way that we were looking at it back then, and it’s still the way that we’re looking at it now is, is we weren’t looking at the technology. We’re looking at, you know, at the problem. And I’ll just give you an example using, using ourselves, like, like what is it that we do? We’re in, we’re in customer service. Customer service very, it’s very simple when you break it down to, to what it’s trying to do. It’s trying to get a person who raises their hand and they say, I have a problem to a state where that person goes, thank you. You solved my problem. Okay. And in between those two states, there’s a bunch of things that go on. And so our job is, you know, with, with HelpShift is we’re trying to build the technology to make it so that someone can get from, I have a problem, I have a solution as fast as possible. And when we focus on that, we’re constantly asking ourselves the question, what can help us do it faster? You know, and so several years ago, we’re like, AI could help this here. So let’s start exploring that. And so, you know, I, you know, I would hope that a brand, when they’re looking at their problems that they’re trying to solve, they’re like, what are you trying to do? Well, we’re trying to create a game that is highly immersive. Okay. Great. What does that mean? How are you trying to solve that problem? And then you should be looking at the, the new emerging technologies long before now, and then be like, Oh, how would AI be able to do this? And I’ve seen some very innovative things over the years. I think what, again, what OpenAI has done is it’s lowered the barrier. It’s made it a lot easier for brands to then implement it because they’re like, Oh, I don’t have to take all the time growing this child up to, you know, to be a college student. I mean, a college graduate, I now have a college graduate to work with. So it’s, it’s lowered it, but understanding how I would have used it, you could have like, you could have easily done that a while ago. So that’s always been my approach is, is, you know, understand your problem and look at it like, like that. You know, hopefully people aren’t surprised. Like I said, AI has been around for a long time. What may be surprising is how fast that it has evolved in the past two or three years. And that’s really what people need to get ahead of is like, it is evolving. It is moving fast. And if you thought you had several years before you could implement that, implement it, because it’s just not there yet. That’s kind of, I think what’s caught people
24:24 Greg Posner off guard that it’s moving very quickly. It is moving quickly. And with that, right, we have HelpShift, right? We basically have our own LLM, our large language model that learns about players intent, right? But when, when chat GPT offers, I’m going to pick on them because they’re the most well known, right? Not the similar offering, but sort of the same technology. Do you, do you look back at it and think about it? We need to implement some chat GPT features just for the appeal, just to get the broader audience knowing that we’re using that, even though we may
24:57 Erik Ashby have a more sophisticated technology already in use. Yeah. So, so there’s, when I look at AI, I actually break down AI into, into kind of three, three major components. So the first, the first part of AI is, is pattern matching. You know, having the neural net, the neural kind of whatever to be able to say, oh, I recognize, I recognize this pattern before pattern matching has been around for quite some time. It’s, it’s used quite well. Then intent detection, using that pattern to then say, oh, I understand what you want. I understand what you need. And that is something that has been coming along, you know, and that’s, in fact, if you take a look at what we’ve been doing with our, our intent engine, which has been extremely powerful, we get easily 90 plus percent of all conversations that come in, we understand the intent and can route it and, and, and handle it just fine. And so the, the intent detection is there and, and you know, Alexa and things like this, they’re all built on that. You, it’s like, oh, I have a question. They understand the question and then they just go Google the answer and bring it back. What is new, what we’ve seen in the past couple of years is the third part, which is the generator of AI, which is instead of just responding, I can now actually generate a response based on a combination of other responses that have happened out in the world. So I can, so it used to be like, oh, you know, it would detect that I wanted to write code. It could point me to code. Now it can say, oh, I intend that you want to write some code. I can now generate based on a thousand lines, you know, a thousand samples of code that I read before. This is what I think the code should look like. And that’s what, when I look at the generator of AI is doing is it’s, it’s adding kind of that third pillar, which is interesting, especially when you’re in a communication experience, whether it’s in a game and you’re trying to generate some, some chit chat, which is really interesting to be able to use chat, shape, EIs first kind of some chit chats and to make, you know, the people there feel real, but, but also in customer service so that you can then respond and have it feel like, you know, like a human being or things like that. So, so I definitely do see chat GPT and the generativeAI impacting and being an addition to the tools and services like that have been, you know, that have been there in the, you’ve seen several of these things come out where like issue summaries, you know, where it summarizes the issue or, you know,
27:21 Greg Posner suggested responses or, or things like that. It’s interesting. I know this is a completely different vertical than ours, right? But right now at the time of filming this, right? The, the actors and the screenwriters and all of them all went on strike, right? And I understood the strike, but then I heard today that they’re trying to replace all, all extras as just being AI generated extras. And that’s just insane to think that these jobs are like legitimate jobs and right, those are movies, but think about voice artists and games and regular artists and games. And like, I dream of the day when I could just talk to an NPC and the NPC knows all about me, but I understand at the same time that’s cannibalizing jobs in the industry. It’s kind of this, what’s good, what’s bad. It’s crazy to, again, with the speed of happening, it’s not.
28:08 Erik Ashby Yeah, let’s talk about generativeAI, you know, for a while. And I’ve learned this in, you know, over the, you know, kind of my experience over the years, kind of being in the industry where, you know, we create stuff, you know, software is the creation of stuff. We are always looking to create something better, more interesting and things like that. And the way I see generativeAI is it allows us to be able to do that, to create more and create faster, which just means that what is produced will be even further and further better. Getting into the industry of, you know, like entertainment, you know, it used to be that you had to draw the pictures, right? You know, it just like, and so to create a, you know, a two hour movie would take like millions and millions of man hours. And they’re like, oh, now computers came in, I don’t need all these animators. Okay, so then what happened? The stories got better. You know, the medium got better. It became improved. And so as, as these technologies come in, that makes it more efficient, you know, humans have to up the game. We’ve got to just say, okay, great. So now we just have to leverage this technology to make it better. And I’ve seen that I talked to a good friend of mine who’s an artist. And I said, you know, there’s this mid journey, what do you do? And she’s like, you know, to be honest, like, there’s a lot of, you know, hype about mid journey. And is it, you know, ethical and all that stuff without getting that she’s like, what’s done is it’s helped inspire me to make better art. You know, that’s what it’s done is inspired me to make better art. And, and that’s how I make my business is I create unique, better art. So, so yeah, it there’ll be things that we have to do as human beings as we as we go through this. But the thing that I always come back to, in all of my time that I’ve worked with AI, is it comes back to taking a practical approach, you know, taking a look at the AI and saying, okay, it is a tool, how do we use that tool in a practical way to have the outcome that we want. And, and to me, that’s just that that’s the best way to use it. It’s like, it is a tool, let’s use that tool to generate the outcome. And if, and if we can now be faster and better, then let’s be faster and better and build up the world with
30:17 Greg Posner it. That’s, you know, that’s how I see it. It’s an interesting point you bring up, right? Because you’re not against that concept, right? And you’re trying to look for this win, win, win. And with AI, right, you have a win for the company, because the company can, can save money by outsourcing, you have a win for the end user, well, sometimes a win for the end user, right? Because you’re dealing with someone right away. But how do you get the actual agent, or how do you get that person at the bottom? I think it’s, I don’t want to call it Darwinism, right? That’s a terrible way to look at it, or the old GE, where you always nix the bottom people, right? But if you’re not using this technology to up your game, if you’re not using it to leverage yourself, you’re going to be eaten alive. Probably not the best way to put that either. But right, embrace the change. This is potentially the dawning of this new technical era, right? And
31:04 Erik Ashby how do you take advantage of it? Yeah, I listen, I, there’s a book that I, that I was recommended to me a while ago, I’m actually going to show it to you here. And it’s, it’s called Factfulness. It is an, it is an amazing, amazing book, because it’s 10 reasons why we’re wrong about the world, and why things are better than you think. As, you know, we’ve changed, you know, in history over a number of things, you know, everyone’s like, Oh, you know, worlds get worse because of this or that, there’s always, there’s always that hype. But the reality is, is that, you know, as society, as human beings, you know, we take these opportunities to better ourselves and end up in the better world. And that’s, and that’s why really, I, you know, I always have come from approach of when these new technologies come out is like, what’s practical? What makes sense for us to, to leverage this? Because it’s, again, and as I said, in the beginning, it’s not about the technology. It’s about the problem that you’re trying to solve. And if AI can come and help you create a better experience, or to solve that problem, then use it appropriately to do that. If you’re bringing AI in just because you want to be like, Oh, I did look, I have AI, then yeah, you’re going to actually have lots of problems, you’re going to make mistakes with it guaranteed. And we’ve seen that before
32:19 Greg Posner with a number of technologies. So am I allowed to go in to start asking questions about what has helped shift on taking a look at these technologies and putting thought into how they’re going to
32:30 Erik Ashby implement them? Yeah. Are you asking me if you can ask that? Or are you asking that? If you said no, I would edit it out. But since you said yes, let’s roll with it. Let’s go with it. Yeah. No, absolutely. I mean, we, you know, like I said, I talked about the, you know, the pattern recognition technology. You know, we actually have been using that behind the scenes for quite some time so that we can gain insights on what is the, you know, kind of what is coming in the system? What are the problems that you’re seeing? What are, you know, kind of so that you can start to understand what’s going on. So the pattern, the pattern recognition has been something we’ve had for quite some time. Again, it’s kind of behind the scenes. You know, it’s this technology that we have. We introduced our intent technology. Oh, I see, I think it was 17 or 2017, 2018 is when we started to introduce our intent technology. We’ve gone through several iterations of it. And what’s really interesting about that is when we took, we first is like, Oh, let’s, let’s just go out and do this. That’s where we learned about kind of the importance of taking a practical approach. When we first implemented it, we’re like, okay, hey, let’s just have everybody type in a word. Like if someone would say, you know, I just purchased something by mistake, right? Well, what did they want? They want a refund. Okay. So we intent, there was a lot of refund. And so we would do that. And we’d have the intent engine doing that. What we learned is that if we put up a menu with the common, you know, with the common options that people would have in this area, that often people would use the menu more than typing. And then you’re like, well, wait a second, the menu is not it’s that’s not AI. It’s just a, it’s just a menu. But what we found is that when we combine those two together, that brought us up to 90 plus percent recognition. And that was where we took a practical approach because we’re like, wait, we don’t have to have them type. We can put a menu or we could do them both. That actually made a lot more sense of kind of merging AI into standard, you know, user interface technology. And it just boosted the, it boosted the outcome. And if I was someone who was like, you know, hip on AI, it’d be like, no, we just want to be pure AI. I’m not, I want to find out what the customer wants. And if I can do it with a single thumb, I’ll do it. And that was when we started to be like, yeah, if you actually, you know, if you bring in, in kind of the practicality about it, you know, we explored AI bots as compared to workflows. And we found out workflows actually perform really well. And so again, we, we mixed AI with workflows. And now that generativeAI is out there, we’re, we’re looking at how does that fit into this? And, and essentially when we’re done, if you were to look at our, what we call the modern support journey, it will have AI from start to finish, but you may not necessarily know or see that there’s AI in it. You know, from the very beginning, you’ll have AI that’s talking in your language so that you can be in whatever language and the bots still understand you, you know, or you may have AI that’s giving you option one versus option two, or you may have AI in the backend that’s suggesting something to the agent. Like we’re just going to embed AI just along the entire journey and not necessarily be like, oh, we have AI, it’s we’re trying to just get, you know, just get the problem solved and we’ll just use AI throughout it. So that’s the approach we’re taking. That’s, you know, there’s, there’s a number of technologies that we’ll use in there. And if you want, I’d be happy to do a, do a podcast on, Hey, how do we use generativeAI or how do we use, you know, language AI or intent AI or, or, you know, pattern matching.
35:52 Greg Posner We can get it all into that sometime in the future. Yeah. I’d love to see that over time. One of the questions that first comes to mind are one of the big news things that come up a lot now is that more companies don’t want their data to be part of that model. Right. So how is what Help
36:07 Erik Ashby Shift is doing different than the open AI is at the bars of the world? Well, and that’s like, like, first off, there is some risks with AI. Okay. There’s a couple of risks. One is the one that you’re bringing up, which is about data sharing. And then there’s also risk about, you know, about errors and things like that. And I’ll talk about both of those, but data sharing is definitely something that is a, is a, is a problem, is a concern. And it’s just like, if you were to bring somebody into your company and you’re like, okay, I’m going to train you on your company, you don’t want that person then going out and being like, Hey, look, here’s all the secrets that I have. And so, so typically what we would do is that we would say, okay, great. We’ll take a generalist that comes from say open AI, and then we will, we will build our models on top of that, that is makes them a specialist, but it’s going to be an internal model that is only access, access by AI that’s built for that company. And so you don’t share your internal secrets outside of, outside of your thing. And so that, that’s where, where people, brands make mistakes. And so if they’re like, Oh, we’re just going to throw in chat, chibit and you’re like, I actually would be a little bit careful on that. What I would do is I would take chat chibit as a base. I would build a model on top of that so that anything that’s, that is secret about my company, that stays within my realm, but I still would leverage chat chibit to maybe summarize a response or to, you know, to, to find a general or language knowledge or any of that type of thing, I’ve leverage chat chibit for. And so like internally at HelpShift, we actually have both internal external AI that we, you know, that we leverage for that, just for that, that very
37:43 Greg Posner reason. With the sudden rise again of the popularity here, did you, have you looked at your product team and decided that make any changes as so maybe you need someone focused fully full
37:53 Erik Ashby time on the different types of AI that are coming out or, or how do you take a look at that? Well, I mean, we’ve always had somebody focused on AI. Do I need to have more people focused on AI? Yes. You know, you can’t have, you know, always have too few, but again, like I said before, like the technology is going to have AI infused in it. That’s, that’s kind of the way the future is, which means that every one of your product managers needs to be AI specialists. Like they need to understand AI because they’re going to be infusing AI across, across the board. And so do I, do I want to have just one person who’s on AI? Yeah, probably so. Maybe two, maybe three or whatever, but I also need to have all the product managers to have an understand how AI is going to be impacting their area and things like that. So it’s, it’s more of a AI needs to be seated. And sometimes you need to take a person to be able to do that seating. So if you were to go back a year, what would you tell Eric or, or how would you approach things differently? If at all, knowing how this has happened, this rises? Yeah, I would say Eric, it’s going to move faster than you thought it was going to be like, you know, if I was looking at my roadmap a year ago, I would say, because you know, AI is always on the roadmap. We’re already, you know, we’ve been developing things, but I would be like, get ready for generative, generativeAI is the next step. And it’s going to move faster than you think it will. So, and like I said, we’re kind of at that inflection point where the barrier to entry is now at a point where it is so low. And the value is clear that you’re going to just see it moving very quickly. And so I, that’s what I would have said a year ago. Just, I would have been, yeah, yeah, buckle up, you know, make sure you’ve got, make sure that you, you understand this and, and, and they’re ready for
39:39 Greg Posner it. So have there been any implementations or tools that you’ve used? Right. And it’s kind of a multi-part question, right? Cause you have your mid journeys, you have, I use something called swell for the podcast. I have, right. Like there’s no true all in one tool yet, but have you seen any tools or implementations of any of these tools that have really impressed you? And you kind of
39:58 Erik Ashby wrote down like, I want to check out how they’ve done this or so on. Yeah. I mean, obviously the, the two big ones, you know, it’s like mid journey, jet GPT. Like, I mean, I, I use those in my, in my everyday, everyday life. And so I can’t, I can’t deny that I don’t use those. Obviously we use them in our, you know, we’re building them into our, our own technology and things like that. You know, I, I’ve actually seen some, some interesting uses of them and, you know, in this, in customer support, I’ve seen some poor uses too. I’ve seen some where people are like, oh, we’re just going to let chat GPT answer all of our questions. And I was like, oh, okay. Let’s see how, you know, let’s see how that goes. But I’ve, I’ve seen other, other brands that really took a more practical approach and said, okay, wait, wait, wait, let’s, let’s really look at how we can kind of leverage it for generative, but maybe not, you know, maybe not for everything. So not without getting into all of, all of the different, you know, different customer service tools that are out there. I definitely think there’s been some, some interesting ones that, that have, have come up. So what worries you the most about this trend? Okay. So there’s two things. And one of the risks that, that I think is actually really real is that chat GPT and these generativeAI have, have one major, major flaw or major concern in the industry, which is they will do things very authoritatively that may be completely wrong. Okay. They just, because they don’t know it’s like, you know, let’s like when you hire somebody, right? The worst, the worst thing that somebody can do is to be somebody who doesn’t know what they don’t know, because then they’re running reckless. All right. And chat GPT, these tools, there are times where they think they’re completely right and they’re completely wrong. And that, you know, there’s this big discussion about, and you can see it like most with mid-journey, if you ever have it two fingers, right? And it does like 20 fingers there. And you’re like, Oh my gosh, what are you looking at? And it just doesn’t know. Like it just doesn’t know what the human anatomy is. And so it doesn’t know what’s wrong. It’s like, yeah, 20 fingers. That’s, that should be right. You know, and, and chat GPT will do the same thing. It’ll, it’ll, you know, it’ll say like, yeah, this is exactly how you do it. And it thinks it’s right, which is why having it run aside of a human is great, which is why putting guard rails on it is what you need to do. It’s why, you know, it’s, it’s why you want to have it, you know, you want to have it specialized. So you can be like, here, let me actually nudge you in the way that I want you to do. And so, so with that, so that’s the first risk. Then you have brands that don’t understand and manage that risk. And then it scares the world. You know, people using these AIs in a way that they’re not really meant for. And what I see is I, you know, I see brands that treat AI, not like they would treat a human being. Like if I were to hire a human being and I bring him and I train him, I give him guard rails. Here’s the, here’s the company manual. Here’s how I want you to work. And people will think, no, AI is magic. We’ll just bring it in and put it in front and say, here, you go answer all our, our most important VIP customer questions. Go for it. You know, and of course it’s, you know, that’s, that’s not, not a good use of, of the AI. So, so those are kind of the two things that I’m, I’m most worried about is that we don’t understand this risk and
43:08 Greg Posner we don’t manage it appropriately. Well, if all go, if all doesn’t go well with this, uh, writer strike, maybe with the next mission and possible, we’ll have Tom Cruise with 20 fingers on each hand. Yeah, actually it’s, it’s been getting pretty good reviews. I might be going out. I know always good reviews with that, that franchise there. How do you feel the next few months and years are going to go? Do you foresee, I guess it’s all moving very quickly, right? What used to take the span of years is all happening. What feels like in weeks now, right? And it’s just going to continue to speed up. So, so if you had a crystal ball, where, where do you think we’re
43:41 Erik Ashby going to be here in a year with all these tools? Yeah. Um, I actually think that the hype’s going to die down. Uh, I think because what’s, what’s happening is that there’s, you know, I’m reading every day. I’m, I’m, I’m out there reading different reports from different people. And I see a lot of the people are like, they’re starting to get there. Like, oh yeah, this is, this is what it can do. This is what it can’t do. And then you’re going to see the more practical solutions coming out. And from that, it’s just going to be, like I said, I’ve talked about the future. It’s just going to get embedded versus part of what we do. And so I’m just, honestly, I like, there’ll be the hype, there’ll be the mistakes, there’ll be things like that, but what’s going to happen is, um, you know, the, the, the practicality and the things that work well, those are going to bubble up to the top. Those will get implemented. The things that, that we’re going to fail in the first place, those will fail. And then we’re going to end up with actually a pretty stable, you know, environment where AI is embedded a lot in the solutions that we have. So I just, like I said, I think it’s just going to happen faster than, uh, than we were expecting it.
44:45 Greg Posner So, um, that’s just to be clear, you don’t, you’re talking about hitting the end of the hype cycle. The technology itself won’t go away. Yeah. Yeah. Yeah. Yeah. Real stuff is going to
44:54 Erik Ashby start coming. Yeah. Yeah. It’s exactly. It’s the hype cycle that’s going to, um, that I, you know, because there’s always, there’s always this hype cycle, like, oh my gosh, what is this? And whatever. And then the experts start to step, step in, they understand it and they’re like, oh, here’s how it can be used. They figure it out. It gets used appropriately and it makes an impact. And then, you know, we, we all go on. And so it’s, it’s really the hype cycle, I think will die down, but I think there’ll be a ton of, um, you know, places where AI will be there. And like, you know, it’s like the comment that you said earlier, it’s just going to be there, but it may not be front and center. Like, you know, and it’s not, and the reality is it’s already there. You know, you go to the store, you buy stuff here, you know, there’s AI that’s that’s there. You drive a car,
45:37 Greg Posner there’s AI that’s there. You’re, uh, you know, it’s, it’s always, it’s already embedded in our lives. So Eric, I think that’s most of the questions I have for you today. I have one last one. I typically ask in the beginning, but you kind of talked about, but what games are you playing today?
45:49 Erik Ashby Oh yeah. That’s a, that’s a good question. I mean, I told you some of my old games. Okay. But I don’t play those, but, uh, uh, and it kind of depends on, on the device that I’m on. Um, you know, I, uh, I have, I have my Nintendo switch right here. Um, and so that one, uh, this is like whenever travel, I take this, it’s kind of embarrassing, but I’m a Mario guy. So Mario, I see that’s, that’s what I’m going through. And I actually, yeah, one, uh, on my phone, uh, it has to be Marvel snap. Uh, Marvel snap is a great, a great thing because I can just play that. I can play that anywhere. I think if I get in VR, I mean, I try a lot of games, but to be honest, Beat Saber is king, you know, just to be able to sit down like that. And then when I, when I get on my, um, my, my PC that I have set up, it’s flight simulator, uh, it’s just a full joystick and everything you have. Uh, yeah, actually I just barely ordered it on prime, uh, on prime team. My wife hasn’t yet figured it out, but I got the, I got the whole setup. I’ve got the, the, the,
46:53 Greg Posner the pedals and everything. So, so yeah. That’s perfect. And make sure you get one use out of it before she finds out. Yes. Yes. I know. So, but Eric, this has been great. I think we learned a lot. I’m excited about the future of the technology and I think you helped spread some more information and more excitement about that. Uh, before I go in, all the information we have will be on the player engaged website as well as the help shift website. But is there anything else you’d like to
47:16 Erik Ashby share with everyone? No, I just, I honestly, Greg, I think it was a, just a pleasure to be able to sit down and talk about this. This is a very important subject and it’s one that, you know, we, you know, are taking a lot of time, a lot of energy working on. And so I really appreciate you,
47:32 Greg Posner uh, you know, offering to have this podcast. So it’s really great. Great. Well, thank you, Eric. And thanks for listening everyone. And look forward to, uh, next week. I’m going to cut that out cause that didn’t come out well, but thank you.