AI Versus the Customer Experience

AI is on the rise and Mark and Lorraine are here to discuss it all! Learn how AI is affecting the customer experience, how advertisers should be thinking about the growth of AI in businesses, and what the future with AI could look like. Join Mark and Lorraine this week to see how Meta, Amazon, and Netflix have incorporated AI into their customer experiences. 

Join Mark and Lorraine for 30-ish as they discuss all things marketing, advertising, and of course … positioning!

Transcript

Mark Vandegrift 
Welcome back to another episode of the Brand Shorthand Podcast. I'm your host, Mark Vandegrift, and with me today is our marketing magician, Lorraine Kessler. Lorraine, show us your magic.

Lorraine Kessler 
I wish I had magic.

Mark Vandegrift 
Well Lorraine, we're going to jump right into the world of AI because that is hot, hot, Hot. Yep. And directly how it relates to the customer experience. There's quite a few ways that I think AI has found itself entering into the customer experience in the business world. Sometimes customers may not even realize that they're engaging with some of these AI features. One of the common ones, which a lot of us see on a daily basis are like personalized recommendation systems, and they collect and analyze a user's data. And then those past actions provide tailored recommendations, right? You're used to it on Amazon. Things that you bought, then you see you might also like these things. So another one is Netflix, right? Netflix provides its users with a list of recommended shows or movies based on its watch history, which personally I find never to be really accurate for me, but it still is trying to do it, right? And they'll do it based on your movie rating or other things that are of similar interest, right? And then, you know, something else I've seen more recently, when shopping for like clothes or shoes or something, you'll even get size recommendations based on previous purchases. So it doesn't just end anymore with, well, you bought a toothbrush, you might like toothpaste. Now it's, you bought a certain kind of toothbrush, you might like, you know, a certain type of toothpaste and even more things like that. So I don't know what your experience has been with some of these personalized recommendations. But why don't you share your opinions and how you think it might be able to enhance the customer experience.

Lorraine Kessler 
Well, for me, it rarely does. So, and that's personal, but perhaps for someone it does. I find just like you, I'm usually after a specific thing, right? Like if it's an Amazon suggest or a purchase suggest. And I just find that the clutter that this creates is sometimes very confusing. So it's been helpful about 25 % of the time and it's been annoying about 75%. But, you know, if we know in direct, if you have a 2 % return, that's huge. So I guess if you get a 1 % response, 2 % response on an item that's served up to an audience that might be in the ball field, maybe not necessarily on the field, but in the stands, I think that's good. I think that's good marketing, you know, but they'll continue, and I think they'll just continue to improve it. And like you, the other day on Netflix, I tuned in and said, well, what new movies are there? Everything was so dark and so scary. So I'm like, I know I'm not a Hallmark channel person, but this was scary. This scared me because I thought this is what some bot thinks about my taste. It's darker than I'd ever want to be.

Mark Vandegrift
Well, I don't usually find it to be accurate because one, our family shares, right? And even though you have those profiles, sometimes we're all going into the same profile, which that's to our own detriment. But if you just randomly decide to watch a show on someone else's profile, because they said, you gotta go watch this, then it kind of messes up the algorithm. So as much as they're trying to track our usage and how that mirrors what other people are doing. It just shows, think at the end of the day, you can't put anyone in any particular bucket because we are all individuals.

Lorraine Kessler 
Right. Well, and here's the problem too that I think is it's called narrow casting, right? And if you think about the radio market, when I was a kid and young, DJs ruled every station in individual markets, Pittsburgh, Philly, New York, Dallas, whatever. And artists would send their work, their new work, and those DJs were kind of the curators they would give a thumbs up or not. And sometimes a song that would play in one market that no one heard anywhere else would all of a sudden catch on, right? Because that DJ was more than a curator and they would repeat that song over and over again. And we know from research that repetition more than anything creates acceptance particularly a new idea because we're jointly neophilic and neophobic, meaning we like new things, but we're afraid, but we are really love old and we're afraid of the new. So what a book called, I think it was by Derek Miller, I think that's who it is. I'll have to come back and check the name, but it's called Hitmakers. And after studying all this in the music industry, what it came to is that we love familiar surprises go, God bless you. So we love familiar surprises. So it takes a while for us to like something new. What I mean by narrow casting is if you're only being served what you've already established a preference for, you're missing out on a whole world of things that if exposed repeatedly, you would start liking. And my best example of that is when I was a kid, know, my ninth birthday on a Sunday, 1964, the Beatles appeared on Ed Sullivan. I it was my special birthday gift. And we loved their pop sound and their mop hair and all that. In the round, I think it was 67 was when Sergeant Peppers came out. That was so new and different. Nobody, none of my friends liked that album. They were like, what happened to the Beatles we love? But through repetition and play and constant hearing of this, Sergeant Peppers became, that is the Beatles, right? No one goes back to, want to hold your hand, right? So it was so new that it was going to be, it wasn't familiar. And then as it became more familiar, it became very popular and it became something that we now associate. the author's name just turned to me is Derek Thompson, Hitmakers, great book in helping us understand this idea of how we're kind of of two minds, neo-philic and neo-phobic, and then you need to have a familiar surprise. So my problem with narrow casting is you don't have a DJ forcing you, if you will, to taste new things, right? And it's like kids who grow up in a household where the parents never made them eat their vegetables or other things. And they grow up, I remember my husband, when he worked, he had a guy who was 47 years old who still cut the crust off his bread. guess parents did him a terrible disservice, right? I mean, that's the downside of being so narrow-casted is you're only served up what you think someone already likes. People I think are always gonna be searching for something new, but something a little familiar and you have to help.

Mark Vandegrift 
Well, what's interesting, you were talking about the DJs. Did you know Michael Stanley Band was only a regional favorite for like most of their career? didn't become popular nationally or internationally until much later after the band was formed because they only got play regionally. And it makes you wonder what other regional bands were out there that no one ever heard of because it was just, they were from that area. So the DJ wanted to play them.


Lorraine Kessler 
That's right. And you know who would know a lot about this is Scott Edwards, our creative director. I mean, he's all over this. has such a great sense for the history of music and how it DJs and the role that they play. And what we heard in Philadelphia, I remember CKLW when I went out to Ohio, to the Northwest Ohio, we got CKLW. We heard different things from CKLW. We had a lot, of course, in Philly, the Philly set. Right? I mean, that was a huge sound and we got a lot of that. And there were groups I heard on CKLW, I never heard of it. But the point is these DJs were experimental and they were all trying to make their name. If they could find a gem and expose it and get the credit for that, that just built their cred. And so it was a win-win.


Mark Vandegrift 
Yep. Yeah. Well, another way AI is looking to enhance the customer experience. And I think most of us have had extensive experience with chatbots. And I even was on a phone bot yesterday and it drove me nuts. I was, I was, troubleshooting for, a local organization that I work with. Their internet was out. And of course I call up AT &T and I get the phone bot, I can't get a live person. And they kept going through the same decision tree. Is this working? Yes. Is that working? Yes. Is this, and you never got to the end of it. And finally, when it still wasn't working, please go to att.net and submit a ticket. And it was like, you know, a half hour later. So, you know, at some level chat bots are useful or phone bots in this case, but they also have their limitations. Where they've expanded is the ability to go past just a yes or no decision tree. Now, depending on context, they can read the point of the question and say, are you really asking this? And then they can take you down a path, but we're still limited by that. According to a study done by the IBM Institute of Business Value, did you know IBM had that?

Lorraine Kessler 
No, it's like the Disney Institute, which no one talks about anymore.

Mark Vandegrift 
Yes, they have the Institute of Business Value. Yeah. Yeah. Well, according to them, now this is coming from IBM. So consider the source. 85 % of executives say generative AI will be interacting directly with customers within the next two years. So 85 % say that. So do you think these chat bots or these virtual assistants are effective enhancing the customer experience? And what are you thinking about the growth of AI as it goes to interacting with customers directly.

Lorraine Kessler 
Well, I don't think you can stop a running train. And I think this is a running train, right? I mean, it's a tool. It's going to be there. And so many of the questions that you might have about any product or service, no matter how complicated, are common, right? So I found that the chat box, you know, what does it deliver? Immediacy, right, with limitations. So it's like immediate. I don't have to wait to get someone on the phone you know, go through voicemail jail, can, you know, quickly use my computer, which I'm already on. And in a lot of cases, I can get my issue resolved. And when I can't, that's when they say, well, you need to have a live agent. You know, can I connect you? So for all that chat boxes deliver an immediacy, they have limitations, right? And that limitation is if you ask a question that probably a hundred people before you didn't ask, or it's off the skew, then they're going to have to connect you to a live agent. I think where  this needs to improve and it hasn't yet is the communication between the chat and the live agent. Because I don't know how many times this has happened to you, but I'll get a customer service rep live and they have me repeat everything I already inputted or start back at ground zero. And I don't know why they can't just review the chat and say, I see what the problem is. Let's pick it up from there. Right. So I think that's going to get better. You just can't stop a running train. Now does this present a way to differentiate? Possibly if you are a company that's very much about intimacy, know, there's like you can be you can from a master strategy standpoint. You can either be product specific. You could be organizationally specific you could be intimate customer intimate. I could be a master strategy. So if I were a company and I wanted to be customer intimate, and that was going to be my real difference, I might not rely on chatbots. I might do something differently using technology to connect to a live agent right away and make that part of the experience, right, that I can offer. But I think that's going to be limited by scale. Like if you get really big, that's going to be really hard to do. So if you're smaller and you have a customer intimate and your experience is very personal, I think this would be something to think about. It's a way we could differentiate and kind of bring tangibility to be more personal, more customer intimate.

Mark Vandegrift 
Okay. Now we're going to get into something a little more obtuse. Okay. And this is predictive analysis. Okay. Well, what's interesting about predictive analysis is that it uses historical data, historical statistics and machine learning, right? To forecast future events. Okay. Now you have kids and you had a lot of history on your kids, how often were you able to predict their next step?

Lorraine Kessler 
Let's say they can predict our next step easier than we could predict theirs.

Mark Vandegrift 
Yes, that's right. Well, predictive analysis is the goal is to predict or forecast future events or trends. And part of that is predicting customer recommendations, behaviors, buying patterns, preferences, many different things. And obviously if we could predict that, wouldn't that make our job easier?


Lorraine Kessler 
Possibly.

Mark Vandegrift 
Yeah, if it were accurate. Yeah.

Lorraine Kessler 
if it were accurate, of what, see the problem is right away, it's doing a predictive model based on what people, the past, what people have already done. It's not able to anticipate what they might do given certain circumstances.

Mark Vandegrift 
Right. So. One of the ways that it does it is called sentiment analysis. So any time a behavior happens, it can be graded as negative, neutral or positive. Makes sense, right? OK, you make a comment, I can say, that was neutral or that was positive or that was negative. So that's kind of around natural language processing and that enables the computers to understand with and interact with human text and language. It's the only way they can analyze a customer's attitude or opinion or emotion towards something. So, the goal is to give feedback on the predictive behavior of customers and then eventually get to the point of developing strategies around that. Now, you and I have always discussed the beauty of the mind, right? And how challenging and rare it is even among human minds to be able to develop the right strategy. As it stands today, what's your guesstimate, one of my favorite words, as to whether AI can challenge humans when it comes to the creative nature of strategy?

Lorraine Kessler 
My guesstimate, are you looking for like a percentage or what are you looking for? Statistically reliable.

Mark Vandegrift 
Yeah, would you say that, okay, Bill Gates a couple of weeks ago on the Tonight Show with Jimmy Fallon, I think it was, said that humans won't be necessary in 10 years. Okay.

Lorraine Kessler 
Yeah. He actually said that. Wait, he actually said that?

Mark Vandegrift 
I would say, in counter to that, yeah, yeah, Lindsey will pull up the clip for us. Okay? So Bill Gates on Jimmy Fallon said that. And I think it was 10 years. If I'm corrected, that's fine. But Lindsey, you can pull it up hopefully and let us know for sure. And just put a caption on the screen if you need to correct what my takeaway was. But that was my takeaway. I would counter with, I'm in one business, a very discreet ad agency business that has very few individuals that truly understand developing marketing strategy for persuasion, as we talked about on our last episode, to convince someone to buy something. So what is your guess? Just yes or no. Do you think computers will ever be able to replace the human mind to be able to do that very thing I just described?

Lorraine Kessler 
No, I think they I think this will all be a modeling experiment that might get us closer, but it's never going to be full proof or right it what it's gonna what it will get wrong. I think will be really wrong and Here's what here's what you said. They're judging the sentiment the problem is and I'm going to reference Quentin Tarantino.

Mark Vandegrift 
good reference.

Lorraine Kessler 
Quentin Tarantino in Making Inglorious Bastards had the hardest time filling the role that Christopher Waltz filled. think that's the actor, right? Christopher Waltz, he's the Nazi. He's the Jewish hunter. And he said because he wanted someone foreign, but people who were German or Austrian, they didn't understand, I guess the word I would use, he has a better word. I don't remember what it was, but...They didn't understand sarcasm and they didn't understand humor and they didn't understand syntax. I have a hard time understanding how AI is going to sort that out, just taking sentiment and flatlining it and then saying this is the sentiment. And so I can't imagine it's going to be much more statistically reliable than quantitative research that we do in old school ways. So I think it might be good as kind of a first blush, but I would want to verify. I would want to trust, but verify. want to like, let's say, for example, it came back and said 65 % of the people would intend to buy this product if these things were promoted. I would want to then do my quantitative on top of that, real people talking to real people in a statistically reliable way. So I'd want to have matching data points because I think there's just a lot we don't know what AI and syntax does not capture. Now this is way above my pay grade, so this is my best shot at this.

Mark Vandegrift 
Well, I mentioned in our last episode about the Grok AI saying that Elon Musk was a distributor of misinformation because the majority had indicated as such. And so that was the importer to AI. so AI was the arbiter of truth, small t. And so Elon Musk acquiesced and said, I'm not going to censor that because we all have to get used to AI being the arbiter of truth. Okay? How reliable do you think that's gonna be?

Lorraine Kessler 
It's not reliable. I mean, that's the frightening part, right? It's a bias. We are already talking about bias is inherent, right? And it's biased. And I think you've said this. I think the best AI can be as a strong collaborator. It's never going to be a replacement for human knowledge. We saw this in the so-called response analysis age, where we can track all the emails and when they open and what they open and where people go. And we would sit down with our people who looked at this data and asked strategic questions and they couldn't answer them. All they could do is say, well, know, 800 people opened it and they went here. And then, well, what they do then? And what does that mean? What do you think that means? And I think that AI is never going to be get to the meaning of stuff, the substance of stuff and the significance. It's only going to take in what's fed in and at best give you at least as a collaborative tool that to work against. And so like, let's take the thing about Musk. We could now do a quantitative survey that goes to people that are the phone. And I just learned this from Lori Dixon at Great Lakes Marketing. They're finding mail surveys are the most reliable and most answered. Literally mailing you a survey. Yeah, we're back to old school.

Mark Vandegrift 
Interesting. M-A-I-L, not M-A-L-E.

Lorraine Kessler
Yeah, I'm talking about in the mailbox. Yeah, I'm talking about in the mailbox. But you might find that that's held by certain people who are very socially active and that there's a whole slew of people who aren't that. So that might be 20%, 25% of the population where the other 75 % are not socially active. They don't get their news that way and they have wholly different opinions. So the biases always have to be asked about from the human side of the equation.  Who did we talk to? What's their habit? Where does data come from? What do they use to communicate? Does that create a bias? Because I think we talked about this, the old famous Dewey One headline from the Truman election. The reason that research was so wrong was because when they dug through the methodology, they found out they only interviewed people by phone. And in that time period in the 40s, only wealthy people of a certain economic and in urban areas had phones. So they missed most of America. And so that was a biased result and it was really in error. And so I think we're going to have the same thing with AI if it can be that easily biased by simply who's inputting, who's engaging with the communication that they get their data, that it gets its data.

Mark Vandegrift 
Yeah. Well, to draw this episode to a close, I'm going to throw out one of our favorite phrases, the sea of sameness. We love to talk about the sea of sameness. you know, Meta rolled out their AI digital assistant called Meta AI in 2023 on Instagram, Messenger and WhatsApp. And Zuckerberg is expecting one billion people to be using it by the end of this year. And then making inroads into businesses to the point that he says in the near future, We believe the majority of businesses will have an AI interacting with customers on their behalf. So we see these big players like Meta, and they're predicting huge growth, and they're making headway into businesses. Don't you see this as devolving into a sea of sameness? And if not, what can businesses and organizations do to avoid sounding like their competition?

Lorraine Kessler 
Yeah, I guess I don't see AI as a differentiator any more than I saw lean business practices, you know, from Deming as being a differentiator. It can be if you're first to the first to the tape, but people are going to catch up because what makes AI better and can be learned and can be adopted. So it's going to become to me, table stakes and these big companies that you just talked about, you know, they're like the general stores, I mean, of all info, right? So they have to do it. They're the big general stores. If there is an opportunity for AI, for it to differentiate you, I think it's gonna lie in specialization because infobesity is an enemy. If I'm trying to be AI, all things, all types of searches, all types of people from bass fishing to cardiology to, you know, where to travel in Japan. That's broad. But I think that if you are a company that can narrow your focus and you can perfect an AI relative to a very special, whether it be cardiology, whether it be travel to Japan, and you really hone it. So it's better at that than anything else. I think we're just back to, you know, it's the same script, specialty stores beat general stores. And that's the way I would play it. I would tailor my AI, use my divining stick, if you will, to find a spot that's where there's gonna be water or opportunity and dig deep. Because a generalist can't afford to go that deep when they're that broad. So it's the same thing as, you know, why Victoria's Secret won over women versus Macy's, specialist, specialty store over Macy's when those things were viable. And neither are now, but they were once. The principle outlasts the example.

Mark Vandegrift 
Yes, differentiation. So going back to that point about marketing to AI, I found this really interesting. We have been seeing leads come in from an AI search engine. I'm not going to name it because I don't want to have it diminish our results. Because of our SEO in many service areas in which we have created SEO pages to talk about our capability in that. We have received very specific lead requests for that service because we are listed as number one or number two in that service that we perform. And how else would they get it except for the SEO? They've never called us. And we're literally listed one and number two. I can guarantee you there's other agencies out there that do these particular services way more than we do. In fact, they're specialists in that. As we know, we're about positioning and then we do all these services, right? These listings in this AI is saying that we are a specialist and we're either number one or two in these services.

Lorraine Kessler 
in a particular like tactical area. Yes. Okay. Wow. Well, you know what?


Mark Vandegrift 
and it's all based on SEO. folks, SEO is your way to market to search engines. Call us, we'll help you do that. There's my tag.

Lorraine Kessler 
I love that. I love that. I love that because there's lots of companies that while they stand for one thing, might want to grow their business. I'll give you an example. Hilscher-Clarke. needed an electrician. went on Facebook here for Jackson Township and said, hey, what electrician would you recommend? And I had a bunch of people recommend Hilscher-Clarke and we did their website and all that, which was great. And my thinking, they added in residential, but the perception still is, well, yeah, that's residential, but do they do really small jobs? Right. So if they wanted to grow residential and that without kind of disrupting their brand, which they're really great at the hard stuff and the commercial stuff what a way to kind of have your cake and eat it too, or eat your cake and have it too, which is the original way that that was said by the way. Makes more sense to me. I learned that from the Unabomber. True, true story. But my point is that you can kind of backdoor this, right? And grow that business all under the umbrella of we really do the complex, difficult stuff without hurting your brand. Kind of an interesting time.

Mark Vandegrift 
Yep. Well, good. Well, let's end it there. AI is certainly on the minds of everyone and we will have plenty to discuss about that in the future, I'm sure. So let's wrap up today's episode of the Brand Shorthand Podcast. Thank you to Lorraine for joining us and to our wonderful listeners for joining us again this week. Don't forget to like, subscribe, tell your friends about us, your coworkers, your family. Tell them to subscribe and like and share. And until next time, have an amazing day.