Artificial Intelligence Podcast: ChatGPT, Claude, Midjourney and all other AI Tools

 Is AI Affecting The Customer Experience With Mary Poppen

• Jonathan Green : Artificial Intelligence Expert and Author of ChatGPT Profits • Episode 381

Share episode

Welcome to the Artificial Intelligence Podcast with Jonathan Green!

In this episode, we delve into the profound impact of AI on customer experience with our esteemed guest, Mary Poppen. With over two decades of experience in the B2B SaaS technology space, Mary shares her insights on how businesses can strategically use AI to enhance customer interactions while avoiding common pitfalls.

Mary emphasizes the importance of balancing AI's scalability with human oversight to maintain a high-quality customer experience. She discusses how data and AI can be leveraged to predict customer needs and personalize interactions, ultimately strengthening customer loyalty and improving satisfaction.

Notable Quotes: 

  • "The beauty of AI is not just in automating tasks, but in uncovering insights from data that we could never discern as humans." - [Mary Poppen]
  • "AI is most effective when it enhances, rather than replaces, the human touch in customer interactions." - [Mary Poppen]
  • "If AI's learning from bad information or no one is checking the validity, it's just perpetuating bad advice." - [Mary Poppen]
  • "A customer's journey should be defined by their needs and experiences, not just by the available technology." - [Mary Poppen]

During the episode, Jonathan and Mary explore the idea that AI should not substitute direct communication between businesses and customers but rather complement it. They discuss the role of AI in interpreting vast amounts of customer data to predict behaviors and improve service delivery. 

Connect with Mary Poppen:

Mary introduces her organization’s approach to integrating employee and customer experience, using AI to create efficiencies and enhance the overall experience for both. For companies looking to refine their customer experience strategy with expert guidance, this episode is full of valuable insights and practical advice.

Tune in to discover how AI can revolutionize customer experience while maintaining the crucial human element that defines excellent service.

Connect with Jonathan Green

AI is affecting the customer experience with today's amazing special guest, Mary Poppen. Welcome to the Artificial Intelligence Podcast, where we make AI simple, practical, and accessible for small business owners and leaders. Forget the complicated T talk or expensive consultants. This is where you'll learn how to implement AI strategies that are easy to understand and can make a big impact for your business. The Artificial Intelligence Podcast is brought to you by fraction aio. The trusted partner for AI Digital Transformation at fraction A IO, we help small and medium sized businesses boost revenue by eliminating time wasting non-revenue generating tasks that frustrate your team. With our custom AI bots, tools and automations, we make it easy to shift your team's focus to the tasks that matter most. Driving growth and results, we guide you through a smooth. Seamless transition to ai ensuring you avoid costly mistakes and invest in the tools that truly deliver value. Don't get left behind. Let fraction aio o help you stay ahead in today's AI driven world. Learn more. Get started. Fraction aio.com. Now, Mary, I'm really excited to have you here because over the next 20 minutes we're really gonna dive into how AI can make the customer experience a lot better or a lot worse. I think we've all experienced both ends of that, and we wanna give guidance to people on how they can ensure that they use AI in a positive way. As I was telling you right before we started, I just. Was dealing with customer support earlier today, and they always force you to talk to a chat bot before you can submit the support ticket. And I have to answer all the same questions. And then they go, oh yeah, this is a known bug. And then you're dealing with the person, they're starting to fix it. But AI doesn't really, at least with chatbots, doesn't have the ability to fix it. And my other worry is that there's a tendency to get lazy. Which is that we don't quality control what the AI is doing. So sometimes we have the AI check in and every project I work on, AI makes mistakes all the time. I was dealing with an AI making math mistakes earlier today, coding mistakes this afternoon, and if you don't double check its work. You start like letting it slide, then things can go off track very quickly. And I worry about how do we train our employees or how do we create a culture that. Still remembers the value of the customer experience and doesn't fall into the temptation to like, just click the button 'cause it's probably fine. Yeah, absolutely. Thanks for having me, Jonathan. It's great to be here with you, and I'm looking forward to talking about this. A hundred percent. Everyone's looking for AI to solve kinda the scalability challenge and also removing some of the tactical work, the work people don't want to necessarily do so that they can do more impactful and strategic work with customers. While I think that's true and we're heading in that direction, you're a hundred percent right that if AI's learning from bad information or no one is checking to make sure the validity of what customers are being given or following up to make sure that their issues have been resolved appropriately, then the AI's just learning. Right to continue to give bad advice or bad information to customers. So there's absolutely that governance piece and oversight piece that needs to be, at least in the interim. Put in place, right? We can't just turn on AI and expect it's going to immediately take customers through the most powerful digital experience and solve all of their challenges. We need to strategically and effectively put the human still into this process in order to make sure customers get what they need, and they do have a great experience. let's start from that point of what does divine a good customer experience? So a few years ago, my wife and I bought a hostel. We ran it for a few years until it was destroyed in a natural disaster. But the one thing we learned was that. The beauty of a franchise is there's an instruction manual with many businesses. There's no instruction manual, and I've experienced the entire gamut, and I've delivered the entire gamut of customer experiences for my customers, where often you don't think of it, but they'll find a way to break the one thing you didn't test. So how do you start at the top with what defines a good customer experience before you try and have AI implemented or try to improve that process? Yeah, it's a great, it's a great point because again, it's easy to take just pieces of a customer experience and put it into AI's hands. And then you just are left with a choppy experience for the customer. One of the most important things still today, which has been talked about for now, two decades or more. Is defining that customer journey, defining what does good look like for your company, your product, your service, right? What is the outcome you want them to have? And ultimately, what does that journey look based on the customer's needs? Once, once you nail that and get it just right now, you can start to infuse AI and start to take over some of those common processes. Or the more tactical pieces, and then you'll be able to strategically decide where does the customer need more personalization? Where do we need to put more human emphasis? But really, until you have that foundation identified, all you're really doing is taking, playing darts in the dark and hoping to hit the target. And all it does ultimately is provide the customer a choppy experience. Yeah, I think that one of the challenges with AI is that the hype and the promise makes it sound like you can just say to an ai, define a good customer experience and it really can't. What it's good at is accelerating a working process or a working system or something that already works. But when you try to have an incomplete or broken process and then accelerate it, things only get worse. And what I often see is that people want to put AI in between them and customers or in between them in communication, whether this is customer support, chatbot, or having to do social media. Having to do social media responses. And while those things can be useful for scaling, they are. Limit you or keep you from catching problems until it's too late often. The strangest messaged me the most important. I'll give you a specific example today on LinkedIn. Someone messaged me and said, Hey, your latest article had a broken link. And I looked and I go, that's weird. I clicked the link, it worked for me, and I realized my browser had it cashed and my website had expired 10 days ago. So I had to pay the renewal fee. So if I had missed that message. That would've been catastrophic, right? And it was like, I thought it was on auto renew. I usually get a notification. I don't know why I didn't renew.$7 bill that some reason didn't go through would've been a catastrophic problem. In catching those little things, when someone says this is misspelled, or this messaging is where these little messages are actually really critical, sometimes a comment on a post has led to a really large client. For me, my biggest paying client of the entire year left a really weird comment last year and I just responded to it 'cause I didn't understand what they meant. And then once they explained it, we had a conversation. It led to all these projects and if we kind of block that signal early. That's when we start to miss exactly when the customers tell us this is a bad experience, or this isn't working, or this is a dead end. And that's we all have it. We first build a product or software, people do something we don't expect. They click this button and this button and they end up in a cul-de-sac, right? And they end up trapped and there's no way out. And then very frustrating go, oh, you have to reload the website. It's the only way to solve it. And if we don't let them tell us they're having that problem, if we create this culture too early of blocking communication or putting AI in there because. Maybe you have someone who checks the AI logs once a week. But that means you don't notice a problem for a week. That could mean it could really scale.'cause my experience is that for every person who tells you 10 to a hundred had the same problem and didn't bother complaining. Yeah, absolutely. I think, that's a great example and actually ties to what you were saying at the very beginning of our conversation, which is at what point do you infuse. AI chat bots as an example, into the support process. And what if, you just wanna talk to an agent? You want that human to answer the question without making those options available for customers. One, it can be frustrating, even more frustrating for the customer, even if they get their answer at the end of, the process. It's not the experience that they wanted to have. And then second. To your point about missing things, if all you do is give them one channel or way to share feedback, and the cadence or the channel itself isn't effective, then you're missing that opportunity to not only serve that customer, but also to learn from it and then incorporate it into your, features or into your process. So a hundred percent the timing and the channel really matter. There's this sometimes there's a breakdown. Once a customer hits a certain waypoint, and it can be after they sign up for a free trial or after they make the first payment, or after they cross the third payment, there's always some point where your automations or your plan drops the ball. And if you don't have someone actively monitoring, it's that drop in engagement can really. Cause them like a real downturn in customer sentiment. And I guess this fancy way of saying if you don't talk to customers enough, they start to get annoyed. And it's that really weird thing where you go into a store and if the, if you're just looking around and they, Hey, can I help you? You're like, oh, that's so annoying. But if you wanna buy something and no one says, Hey, can I help you? Then you're annoyed like there's no winning. It's that unwinnable thing and it's. This, there's this worry that I experience. I know a lot of people say, oh, I don't want to email too much because it's so invasive. And if I email every day. Then the customer will get annoyed, but if they don't email often enough, they'll also get annoyed. And it's this, what's the right balance of engagement and is there a way AI can tell you which people want to be engaged with and which ones wanna be left alone? That's so fascinating because, so you just really described, there's there's more than two, but I'll just. Kind of boil it down to two, there's a tactical experience and then there's a personalized experience. And most of the time the customer experience is gonna fall between the two. But if all you're really wanting to do is offer a technology, or maybe it's, I don't know, just a basic feature that you're offering for free, having a tactical AI based digital approach generally will work because all you want for an outcome is for them. Customer to get an experience with that technology. But if you're really trying to build a customer base, customer loyalty upsell opportunity, then you need to start thinking about that personalized experience so that digital tactical focus with AI just. Leading the customer through a basic journey, it's not gonna be effective at building that relationship and keeping a long-term customer. So you need to start thinking about that personalized experience. How can we identify when Jonathan wants help or not? And that's where the power of the data and behavior of the customer and the sentiment of the customer all together tell a story that you can then start using. To predict what is the customer gonna need? At what point? What's so beautiful about ai, not only, generative ai but something that's been in place for quite a while now is that behind the scenes AI and the trending and the power of being able to look at millions of pieces of data right, and start to make sense of it. Something we could never do really as humans unless you knew exactly what questions you wanted answered. So data scientists are great because if you tell them the questions that you want answered, when does Jonathan want help versus not? They can start to identify what data to look at, but think of AI looking at these millions of data points, right? And customer behavior, and it can start to surface. Jonathan's looking for this at a particular point in time comes to the store. This is the time that we would wanna say, can I help you with something? And so allowing AI to surface that insight, you are going to uncover so much powerful and predictive capabilities than we ever could before. But companies have to embrace this. They have to start centralizing their data and really start thinking about their data strategy. Yeah. I think the important lesson, and this is my thought, is that you have to know if you don't have a lot of data, then you have to have a lot of communication. And like often customers will say, oh, why isn't this button here? Why isn't this button there? And I go, oh, okay. We'll add to the header two. And that feedback's really valuable when you don't have a large number of customers at a certain point. You have a lot of data or you need someone else's large data set of customer behavior. Because what I have seen sometimes people think, oh, we have 50 data points. I'm like, no, you need like hundreds of millions of data points to do predictive analytics.'cause I work with people who do really large financial modeling. And they're like, oh, we have 400 million data points. So you can predict six months of behavior, like four million's a lot. And that's what's. Very critical is that figuring out where you are on that spectrum. So at first, the customer data is super important. In the middle, I think you can leverage other people's data because people act the same whether they're in one store in the mall or another, right? They make the same face or they walk the same way if they wanna be left alone. So you can use your universal data until you get to the point where you have enough of your own data, and that's where I think AI can become really powerful. And there are some things to be said for the objectivity of ai, like sometimes. The length of a message is more important than the content of a message. Like the people that write the longest messages are the most invested, and that's something we totally miss. If the person's angry, you go this person is angry, but they wrote four paragraphs, which means they care. If they don't care, they'll write two words. So there are still ways to use AI to create that experience, and I think that we're seeing. Definitions of experience change a lot. Like when I watch how my parents shop versus how I shop versus my wife versus my kids. My kids cannot understand why non-touch screens exist they're, until they start having to take computer class, they're like, what is the point of a keyboard? And these different ways of communicating, right? And we're seeing now with shifts in customer experience that like millennials versus generation X versus Zoomers versus boomers. I always mix it all up, but everyone interacts in different ways and wants a different customer experience is one of the first places to start. Looking at the kind of age brackets of your customer, even if you just have like gross data, you go, I don't know. And did, before you do individualized experience, you go, let me just do an experience based on this person's age, or a few psychographics and start small. Is that still doable? That's really it's very fascinating and very powerful, but I'm gonna come back to that baseline customer journey. And the outcomes that you want the customers to have, and then starting to segment the journey by what you're describing would be amazing to get at. Not only building a better customer experience, but getting better data for your organization. Because the people, the younger generation that does not wanna sit and type on a keyboard will pull up their phone and just talk to it. Hopefully the AI is picking up now their speak and it makes sense. So you don't get a lot of qualitative data that's nonsensical. But the AI's getting really good at picking up on that too, right? But if you going back to only offering sort of one channel or one cadence to get feedback, you're gonna be missing a whole lot of information. So I think you're, you're spot on. Start collecting the feedback. I don't know if it's necessarily by age, maybe it is by preference of channel. You need to provide feedback. But looking at response rates and looking at the type of data and information that's returned through those options allow you to make better decisions in terms of how to collect that going forward. So I think. The strategy at the beginning, you need to have a good baseline and a good foundation, but you can't just be one and done. This isn't our listening strategy in a, in a box, right? It has to be ongoing. Someone has to be overseeing the evolution of it, right? Because otherwise you're just gonna get stuck and you're gonna be missing a lot of pieces. So I agree with you. I think that would be a really good next step. It's probably a little bit more advanced than where some organizations are, but it is absolutely, I think, the right direction to go. What are the types of tools that, the types of AI tools that can help employees to give more personalized or more. Informed or just a better customer experience? What is the starting point or what like features should a company look for in a tool that they're thinking of bringing into their toolkit? Yeah. I would tell you one of the things that I've seen and deployed is the ability to leverage data holistically and then start to surface these insights about customers. So I'm just gonna go with product usage as one example, right? How your product is consumed and adopted might vary by industry. It might vary by maturity of organizations in terms of a process, those kinds of things. And once you uncover the expectation of adoption, now you can start to monitor product usage and adoption and start to raise flags that. Customers are using your solution in a way that's either effective or not. And it's gonna vary by organization based on, again, all of those factors. But if you effectively use the data and the AI to uncover those, now you can start to make really smart decisions as an employee in terms of when my customer's ready for new features. Are they ready for upsell? They've already consumed 50 of the 60 licenses that they purchased two months ago. I need to reach out proactively and help them purchase more licenses before. They run out. And those are the kinds of insights that employees can use to do a better job with customers and be able to provide more value. And that's just one example. It could be number volume of support tickets, right? When you start to see a spike, is that a good thing or a bad thing? Is it expected? Because in and of itself, volume of support tickets doesn't really tell you much. If you look at it in the context of how the customer's getting value from your solution, you can start to piece together. When do I need to reach out to the customer? When are they struggling and when are they actually doing what's expected? I think this is really important because we look sometimes at the wrong. Number and one of the most important things I think is what are the different issues? So my very first product, this is 15 years ago, you could, when you purchased, you could log in once, but you couldn't log in a second time. So when I did all the testing, debugging beforehand, you test it, make sure the thing works, you do a login. And so 1,760 customers generated 1,760 support tickets. And or, but the thing is, once you do one fix, you can then email everyone and say, it's fixed. It's fixed for everyone, right? But it's still a massive number. And so you can look at it and go, wow, there's a hundred percent complaint rate. But once I figured out what was happening, I learned my lesson.'cause I never would've thought of that. I never thought, log in, log out and see if you can log in again. Now I do when I create something, I test for that too, but I learn my lesson and we can sometimes get distracted by the wrong variable or the wrong element of, oh, there's a bunch of support tickets and it is important to look at. Are they all about the same thing? Is this universal fix? What's causing it? And certainly, I've been building a lot of software lately and getting a lot. Of customer support tickets. Every time there's the tiniest issue, I have to go and fix it, which is fine and it's certainly helped me modify it. But I do think that a lot of. Times we start to think of the customer journey as our own journey. This is something that I've had to learn over the years, which is that my customers don't act the same way I do. This is a talk I have with a lot of people. They go, I don't do that. I'm like, yeah, but you're not your customer. A large number of my customers, for example, will buy the digital and audio book version of the same book and they'll read some at home and they'll listen some in the car. Then they'll read some more later, and maybe you do this. Like I don't, I've never done this. But 'cause it's very hard for me'cause the audio book's so much slower than the written book. Yes. But tons, like 40% of the readers of my books do this. It's a huge percentage. And I had to learn, 'cause I was like, I'm always like telling people, get out the audio book as fast as you can. You'll be amazed how much bigger audio books are than the other format these days. And the way people engage and people will surprise you. And I think that there's this maybe it's POV bias where we start to think I don't do that to, my customers won't do that. How can companies maybe break outta that and say our customers. Act differently than us. And one of the things I, he is you don't, your customers don't have the same expertise you do. They don't understand the logic. And you can certainly see this you look at the experience of chat GPT, it's terrible. It's a blank page. It's the worst customer experience you could possibly have. And that's the template for ai. So how can we maybe use AI to get out of our own heads and start to think about how our customers are different than us? And to start to see what they like and don't like and what they want and don't want. So it's really interesting that you brought that up because it leads me to this story that I have. I worked with a customer that has a financial modeling software solution, and it's really powerful and it's big. They have lot of thousands hundreds of thousands of customers and, the assumptions that are made by the organization in terms of what will make the customer more successful, to your point, the mindset of this makes sense to me because it don't make sense to our customers 'cause it makes sense to me, which is the more training we give, the more successful they'll be with our product. And so this company invested. In spinning up a full team, lots of FTEs, creating this mastery kind of library of training with four levels and decided, the more time they spent getting their customers through all of this mastery training, the more successful their customers would be at the end of the day when the data was actually looked at. And some AI was used to do some linkage work. Only the level one training had any bearing and validity to the usage and the success of the customer with that product. And so they were spending all this time getting customers through four levels of mastery training and their long tail of customers, right? A hundred thousand customers. They weren't even worried about going through level one. Level one was the most important level for any customer. So here their long tail had started to churn, right? They had an increased churn. And all they really had to do was sit down and focus on that customer experience and look at the data before making that assumption that more training is just getting, better outcomes. So it's pretty powerful when you think about the decisions that are made based on assumptions, and you spend all this money and effort. At the end of the day, only a portion of it really matters. Yeah, I think that. I've seen this happen so many times with people creating their first product, they go longer is better. And so they'll say, I want this course to be like 75 hours. And I go, great. I'm actually releasing a new 75 hour course. Do you wanna buy it? And they go, we're too busy. And I go, so are your customers? Yes. And it's this thing we do when we're author. When you're creating your first book or your first product, your first deliverable, you always go too big. Because you start to think quantity is the most valuable thing. I, as I've become more successful, my, and I have five kids, my time is more and more precious. So I, if I can learn or get it done in less time, and as you were telling that story, I was just thinking like, how many people, and this is obviously aging myself, never read the VCR instructions. It was always flashing 12. Maybe you learned to set the time on it once. I was like 20, 30 years ago. But it even back then, right? The instructions were probably perfect, but no one ever looks. And that's the problem with training. And I think about this, I actually have this new thing I do whenever I play a demo of a video game is I start a timer between when I click start and how long until I'm doing something. Because some games, it's three minutes. I was trying to play this game earlier today and they started. Like the main character's blind, so it's a blank screen. And I'm like, okay. Oh wow. And they're talking in German without sub titles. I'm like, okay, no, this is not, this is too much. Like I'm not gonna sit through. I sat through 10 seconds of it and then I was like, all right, I'm outta here. You lost a customer and it's the thing of being too clever, which you can sometimes fall into. And that's it all comes back to, it's like you can have bad ideas, it's gonna happen. I've certainly had plenty, but if you're not. Checking and adapting to the behavior. And this is one of the things, another lesson I've learned with a lot of sales webinars and training webinars. That's how I do a lot of my business. It's that you can often just delete the first 20 minutes of your hour presentation and the numbers don't change. So once you start having a baseline, be willing to test different ideas, be willing to test different experiences, and people will tell you. Something like customers will often say, oh, we want more training, we want more of this. And I'm like, yeah, but the numbers don't match that. So I've had to learn to give based on data, not just based on my opinion that, oh, if I do more, the training has to be this many minutes longer than that. Sales decrease if you give people too much.'cause what you can do is if you give too much on the free side, they go, oh, I need to go through this before I make the buying decision. And you actually lose your customers 'cause you, they punish you for giving you too much value. And so being data driven is really important. I think that. It's really important, especially now for companies to be like cautiously optimistic to use their data to provide a better customer experience, but not to, like you demonstrated there with that story go too far down a rabbit hole without double checking their assumptions. Because sometimes we correlate to data points that aren't related. And that can. Happen all the time. So I think this is really valuable and this is really helpful. Thank you so much for so giving us so much of your time, Mary, for people who are interested in kind of what you do and how you help companies to improve their customer experience and come in strategically from the outside and sometimes that's so important to just have someone else look at it because if it's your baby or your passion project, it's very hard for you to look at it as objectively as needed. So tell people a little about what you do and how kinda people can find out all the amazing things and where they can find you online. Yeah, you can find me on LinkedIn under Mary Poppen, and that's a great way to connect with me. I have spent I'm also aging myself at over two decades in the B2B SaaS technology space. Mostly focused on customer experience and HR technology. But I have, being in HR tech, I've seen the linkage between. The employee experience and the customer experience and how much overlap there is between the programs that you can roll out to really impact the organization along with your customers and your employees. There's just so much powerful data and insights and now AI technology that can be incorporated. And that's what my organization now is really focused on is looking at your employee experience, customer experience. How can we, bring these things together, create more efficiency, but create a better experience really for both. That's amazing. So for everyone out there who's thinking about how to give their customers a better experience, Mary is amazing. I'm so glad you came on the show today. Thank you for being here for an amazing episode of the Artificial Intelligence Podcast. Thank you. Thank you for listening to this week's episode of the Artificial Intelligence Podcast. Make sure to subscribe so you never miss another episode. We'll be back next Monday with more tips and strategies on how to leverage AI to grow your business and achieve better results. In the meantime, if you're curious about how AI can boost your business' revenue, head over to artificial intelligence pod.com. Slash calculator, use our AI revenue calculator to discover the potential impact AI can have on your bottom line. It's quick, easy, and might just change the way. Think about your business while you're there. Catch up on past episodes. Leave a review and check out our socials.