In this week’s Friday 15, Andy & Brian discuss:

– A new dating app based on people’s credit scores (and what that has to do with B2B)

– The next generation of conversational commerce.

– The Affiliated Distributors Marketing Summit and all of the buzz around AI.

—

Brian Beck: Good morning, Andy Hoar. Welcome to Friday 15. 

Andy Hoar: Or good afternoon, from the east coast. 

Brian Beck: I’m Brian Beck, we’re with Andy Hoar, my partner in crime.

Andy Hoar: You’re on the road, it looks like. 

Brian Beck: I am on the road, sir. I can’t disclose my location, (laughing) my knock on the door, so I got my window escape here, I got the sheets tied up, I’m gonna crawl out. I’m in St. George, Utah on my way up to Park City, Utah, to the little skiing this weekend, so yeah, taking the drive up from Southern California, it’s a long drive, so we decided to stop halfway through and I’m here at the, we find us at the Holiday Inn Express. (laughing) Got my coffee. So welcome folks. 

Andy Hoar: I saw an interview with the people who are an Oppenheimer and to put you in good company when they were filming in I think New Mexico, all the stars stayed at a holiday in Express, we’re talking about A-Listers, Academy Awardwinners stayed at the Holiday Inn Express. 

Brian Beck: Oh, well they have good coffee, what my son likes is the pork sausage, he’s 12, (laughing) he’s like, you can’t get barely in California, you can’t get pork sausage where we live, so, you know, here in Utah, you can get pork sausage and he loves it, so, (laughing) you go back and stay at the holiday inn. So folks, welcome to Friday, 15, we’ve got some great stuff going on this week, some really interesting things, we’re gonna do a report from the field, but before we do that, Andy, we have some. (upbeat music) He knew, he just loved that sound. So Andy, you found this week, this is absolutely hilarious. 

Andy Hoar: I was looking for this, so let’s be quiet. 

Brian Beck: All right, we’ll talk to Cindy about this one, but go ahead. 

Andy Hoar: We occasionally depart from B to B, rarely. This one is just too good to be true, so there’s a new dating app called Score. Yes, it’s a double entendre, as you’ll hear in a moment, but-

Brian Beck: Is that SCOR or E? Score? – SCOR or E. 

Andy Hoar: But not what you think, it’s actually a dating app that is tuned to people’s credit rating. So, crazy. I guess somebody decided this was one of those missing variables in the evaluation of a potential mate, is you wanna know what the credit score is. I guess you don’t want some deadbeat, showing up on in a Trans Am for the 1980s but with a credit score of 320. So they can match you up with people with similar credit scores. I think that’s fascinating. 

Brian Beck:  Did you just say that someone driving of Trans Am in the 1980s? Actually, that is kind of cool. If you have a Trans Am from the 80s. (laughing) Anyway, when I was in high school, that was awesome. So, all right. Now, that’s a fascinating thing. So, what we’re gonna talk about today is actually more about AI. And one of the fascinating things, this story came out, Brett Taylor. He was a CTO of Facebook. He’s the chair of OpenAI. He’s on their board. He was the co-CEO of Salesforce. Everybody knows this guy in Silicon Valley, but nobody knows him outside, right? But he founded this company here called Sierra. And this was in TechCrunch just a couple of days ago, two days ago, I think. How Brett Taylor’s new company is rethinking customer experience in the age of AI. He’s got some interesting goals here, Andy. Tell us about him. 

Andy Hoar: Well, you know what? This is only a matter of time. Now, we’re seeing kind of gen two, maybe even gen three of what’s called conversational commerce. And the original bots were a little bit problematic. They made some improvements. They’ve gotten better. They still have some weaknesses. They had an experience the other day with one with e-trade. It was terrible. But I’ve also had really positive experiences with these. But I think Brett Taylor finally realized, when it meant now we can take this to the next level. Those before were not AI-based bots. This is a pure bottom-up AI bot. This guy’s got a lot of credibility. And you can see what he wants to do. It’s pretty ambitious. He said, “AI will become the dominant form factor that people will use to interact with brands, not just for the sort of current trends like customer service, but all aspects of the customer experience.” I agree with them about this. I do think they’re gonna be better. They’re moody. They have infinite knowledge. But they’ve always just been limited. AI actually takes the limits off. It takes the guard rails off of it. So it’ll be an interesting thing to see what happens here. 

Brian Beck: Well, to me, Andy, this also reflects what Amazon, the announcement we talked about last week with Amazon Rufus, or two weeks ago, where it’s contextual, right? It’s adding all kinds of different aspects to the, to now companies that can put these things on their own site. And answer contextual questions. I think that’s part of the promise here too, is not just answering routine, where is my order, but contextual questions about where do I, and how do I use something, application questions? What’s the best for this particular use case? 

Andy Hoar: Well, and we talked about this. This is particularly, talk about this last week. This is not only a nice to have in B2B. This is quickly becoming a must have, because we have a workforce that’s starting to retire and a lot of their domain knowledge is leaving with them. Because it brings up some 25 year old good natured person who’s gonna work at the will call desk at Granger and expect them to know how this particular part or piece is gonna fit into this. How do you bridge the gap? It’s gonna have to be something like this. 

Brian Beck: Yeah, no doubt. And so, yep, so it’s fascinating. And the whole AI theme, our topic today is the buzz from the field. And I’ve had a busy, you know, last week and a half, 10 days, Andy. You know, you’re sitting back drinking lattes and I’m running around the country here doing all the work. I’m just teasing you. But, but seriously, the buzz from the field folks, as you may suspect, is all about AI. And I was fortunate. Thank you to Caroline Ernst, the head of e-commerce at Affiliated Distributors. I was able to attend their 2024 AD conference and marketing summit in San Antonio, a very cool city, by the way. Andy if you haven’t been there. And, you know, AI, just looking at the agenda and sitting through, you know, the talks and the keynotes, a lot about how to use AI practically. And there was some high-level trend stuff, I’ll share in a minute. At the end of the day, a lot of it was about the, you know, these companies and how are they, how are they practically using it? So, AD for those who don’t know, is basically a trade group. It’s not an industry association. They’re really a buying group, but they’ve introduced all these, you know, sort of groups within it. So they electrical, plumbing, different categories, HVAC. And they bring people together and they started this e-commerce division, I wanna say maybe 10 or 15 years ago and it’s really about, you know, enabling their distribution members with tools. For e-commerce, for data management, for PIM, things like that. So you’ll see a lot of the PIM companies, e-com platforms at their conference. But Andy, it was really about AI, this conference. And it probably, I don’t know, several thousand people were there. And some of the use cases, you’ll find these interesting. I won’t, you know, go through everyone. For those of you who are listening on our podcast, I have a list here of a couple of different use cases that I’m sharing that companies talked about in their panels. The one I wanna call out here is, I thought was really funny, Andy, is APR Supply Company, which is a mid-market distributor. They’re using chat GPT to write policies on how to govern AI. 

Andy Hoar: I love it. (laughing) Hey, each of your dog food. In fact, we have a story come up a little bit later about this. It’s, if it’s not good enough for that, then what is it good for? 

Brian Beck: Yeah, I mean, but, I mean, other practical things too, like ingesting sales orders into the ERP from emails. So think about the inefficiency, right? Of re-keying orders or emails coming in. I mean, B2B, you’re still getting a lot of these manual orders, even fax, right? PDFs, you’re getting orders coming in through email, taking and using AI tools to extract that information out of the unstructured data that’s in an email or the PDF and then bringing that into the ERP. Writing code is another common one. Ideal supplies using it for incoming traffic, predicting personas so they can create a better experience on the website. I heard all kinds of things. You know, everyone says AI is all about, you know, just content, writing content. I heard, I’m hearing lots of use cases in other areas as well. So it was a fascinating, fascinating to hear this, you know, practical ways people are using it. But this is something which is really interesting, Forrester Research did a presentation on day two of the conference and she talked about something called B-Y-O-A-I. What is that? “Bring your own AI.” 52% of B2B decision makers are concerned about B-Y-O-A-I in which employees use unsanctioned AI in their work. Have you heard about this Andy? 

Andy Hoar: I have, you know, reminds me of the first wave of innovation with all new technologies. We had the same thing with storage technologies, dropbox, one box, if it even still exists. That’s what it was. People would just, you know, do whatever it took to be more productive and then all of a sudden CIO’sI would realize it’s thousands of people using DropBox. DropBox then showed up and said, hey, you want to do a corporate license? This is going to follow a similar pattern. But it does show you that this has penetrated the consciousness of workers. And they realize that it can make them a lot more productive. Companies, however, are scared to death. Right. This could become the wild, wild west and open them up in all sorts of negative possibilities here. 

Brian Beck: Well, and for good reason, Andy, I mean, think about, I think the quote unquote, sort of risks to sort of security and information, privacy, and confidential information are significant. Check this out, Andy, less than a day ago. I think about 18 hours ago found this leak from an Amazon internal document or an email that this is reported in Business Insider, Amazon’s internal documents warn employees not to use generative AI models for work. Now, they’re not talking about Amazon’s, right? But they’re talking about using external AI to use it for putting internal information into so that someone can create notes or a summary document or something else using ChatGPT. Amazon doesn’t want their employees doing that, right? They’re worried and the key of the worry here is info security. 

Andy Hoar: But is there something ironic about this? Don’t use Gen-AI because we can’t trust it. Now, go out and sell Gen-AI because you can trust it. Right, right. Right. 

Brian Beck: Well, I mean, they’ll think about it too. I mean, they remain competitor, Microsoft, invested heavily in OpenAI, right? So ChatGPT. So they’re concerned about it from a global competitive perspective. What do you think about it? What if all of Amazon employees start throwing internal information into ChatGPT because it creates a nice summary of my meeting the other day. But then that meeting the other day was a whole bunch of things about what Amazon is doing, and then competitively, right? And now Microsoft gets all this information. So anyway, there’s a lot happening here around info security. 

Andy Hoar: This is the world of bartering information. We’ve had early incidents with this. We’ve all been in this scenario where a company is trying to give them the email address and exchange for a discount. This is no different. It’s bartering for value. It’s like, hey, put something in here. And what we’ll do is provide back a meeting summary, some synthesis of this, but you’re right. Once it goes into that system, it’s in the system. And you could be exposing yourself. So this is once again, another, just elevating the issue about the value of data. No question. And it was, I heard the Forrester person say something else, interesting at AD. This is what she said during her keynote: Humans aren’t going to be replaced anytime soon, right? So any time– so it was the anytime soon, please, they got me on this. We have this kind of the theme I’m hearing at these conferences is that, oh, it’s not going to change. Everyone will be OK. Everyone’s going to have a job. No, I don’t think so. 

Andy Hoar: I don’t know. I call these bedtime stories. You know, the fairy godmother saves the day, and everybody’s better off. Now go to bed and enjoy your slumber. This is nonsense. That is purposefully why they’re building these tools. What was Brett Taylor’s company Sierra about?  Working side by side with humans? I don’t think so. It’s conversational commerce, et cetera. It’s about taking what humans do, building it into a bot, having the bot handle it, not partially, but wholly. Now, we may play a game for a couple of years where it’s side by side in partial, but don’t kid yourself. The end goal here is for this stuff to replace humans in because it’s supposed to be better, more efficient, faster, more knowledgeable, et cetera. 

Brian Beck: Well, and there’s going to be new jobs created, no doubt, but there is going to be an impact. So this last just two days ago, Andy, we had this fabulous event in Los Angeles. So again, report from the field here. We’re at the 51st floor of the City National Bank Tower at the City Club LA. We had almost 40 executives from the LA area. We got together. We had a very robust conversation, I guess, about what? Guess about what, Andy? AI. We had really the who’s who of the LA area. We even had a couple people fly in from out of town for it. It was a great event, followed by dinner or cocktails. But in terms of what we shared and what we talked about, again, I heard a lot of practical use cases and how it’s being used. A number of different things here. A lot of folks talking about using it for content creation. There were two that were really interesting that came out. One is scanning thousands of customer service call recordings for sentiment analysis. Talking about how they’re doing this for some of their clients to really understand and also improve customer actual care quality, making sure the agents are answering questions correctly. He was describing thousands and thousands of calls a day for one of their clients and how they’re using these tools to extract information and trends and sentiment. The other one was really interesting. Pooja from Motion Industries was talking about online chat bots. And she gave this example Delta, someone apparently, a chatbot, a Delta apparently answered some customer’s question with some non-existent loyalty program. But had made it up. It was hallucinating. And the customer took Delta Airlines to task and said, your customer service promised me this stuff. He didn’t deliver it. I’m going to sue you. And they won $600,000. And then Delta had to end up adopting the loyalty program, it’s insane. You have to be really careful about what goes into these systems so you don’t get hallucinations. 

Andy Hoar: But you know there’s a solution to this, which is AI is going to monitor AI. I’m not kidding. So that was– you don’t just turn it loose and let it do whatever it wants to do because you’re right. It’s going to hallucinate. That’s the downside. But you can have a check in place to make sure that it doesn’t make up a new loyalty program. And you can use standard software AI to do that. That’s the piece. That’s the layer that was missing here. But I wouldn’t through the baby out of the bath water and said, hey, just because it hallucinated this loyalty program and Delta got burned, if they just spent a few more dollars, a few more hours, and had something double check, would it go insane in real time? They would have never had this problem to get the best of both worlds. 

Brian Beck: Yeah. No question. We also– one other interesting point Andy was the whole– the guys from Slope were talking about this whole sort of– making this whole process of credit approval much faster and approving people much faster. But also, interestingly, some use cases around suspending, proactively suspending accounts. Knowing if someone is going to default before they even purchase. I mean, it’s using AI to predict the default of someone so that they could actually say, hey, we’re putting your account on hold. Well, wait a minute, 

Andy Hoar: And how do you know why we talked about Score? Because this ties right back into that. 

Brian Beck: There you go. It circles back to Score. Yeah, people– maybe they should all work together. I don’t know. So we asked our LinkedIn community, what’s the most practical use case right now for Gen AI and B2B commerce? Far and away, number one, was 67% was marketing and product content followed by a distant second with customer service chatbots, writing code, and at the end here, technical documentation. I think everyone sort of sees the use case on marketing and product content. And a lot of people testing that. But this is really across the board. And I asked a question at our roundtable Andy, how do you prioritize? And somebody said, well, you do what Jeff Bezos did, right? You look at the customer first and you start there. And I have to agree with that. What pain can you solve for the customer using these tools? And there’s so many applications here for AI. It’s incredible. 

Andy Hoar: Yeah, wherever there’s high level thinking and taking place by humans, there AI is going to fit in there. And even some low level thinking. The thing about product and marketing information that makes AI so powerful is that if it exists anywhere in the known universe, AI can find it. That’s a weakness to AI. If it doesn’t already exist, then it might just hallucinate something. Again, that’s going to get better. But what would you do? Pay a human to scour the internet to find all this stuff, do a bunch of searches. Why don’t just have AI do it? I do think you’ll see these numbers change, though. These are the very early use cases. They’re a little safer. It’s a bigger problem. There isn’t much of an alternative. That’s why marketing and product content is so high. But I bet if we do this in a couple of years, customer service and anything around the customer experience, will start to rise because that’s the next problem that people are just realizing. They can’t get talented people. There’s people making mistakes. It’s an annoying process. It’s a friction in the flow. You don’t contact customer service unless you have a problem. You don’t come off and say, just wanted to thank you guys for such a great process. So if you can eliminate friction with AI, you sure is how it going to do it.