This week, Andy & Brian discuss whether a recent court case means Amazon will be treated like a distributor and be responsible for the quality of the products it sells.  

Then they dig into the new Amazon AI chatbot called Rufus, and whether Rufus will put the salesperson out of business.

Brian Beck: Welcome to Friday 15, everyone, with Master B2B. My name is Brian Beck, I’m here with Andy Hoar. We’ve got some good stuff coming today. We’re talking all about Amazon. We got some news about Amazon this week, which is really interesting and exciting we’re going to dive into that in just a minute. But Andy, I’m excited about some breaking news, which of course is all about Amazon.

Andy Hoar: Hey! You messed up the breaking news. – What’s the other guy look like?

Brian Beck: What other guy? (laughing) – You know, oh my nose, yes. Well, for those of you on the podcast, you can’t see my nose, but I got smacked by a volleyball last weekend. My wife and I like to play volleyball. I’m 52 and for some reason, we decided we wanna play against high school volleyball players who like to hit the ball really really hard.

Andy Hoar: Are you talking about all girls though?

Brian Beck: Oh, shut up. It’s okay, by the way, girls can hit the ball really hard too. (laughing) – Yeah, they can, they can for sure. Anyway, it’s fine. So, I recovered, luckily, didn’t break, but yes. So, our breaking news, I’m going to go back to our breaking news. Amazon could be held responsible for third-party products sold on its platform. So the US Consumer Product Safety Commission has a case in front of it where they’re ruling, they may rule that Amazon is a distributor, and if they rule Amazon is a distributor, that means that it would be liable for third-party products, including the safety of goods it sells, it sells for outside vendors on its website and ships for them through its logistics network, not just the platform. Amazon has historically claimed, “Hey, we’re not responsible for those products. If a third-party seller puts them up on the site.” And you have cases, Andy, where some stuff was documented in the Wall Street Journal over the past few years, where bicycle helmets used by kids that don’t work or motorcycle helmets, I think it was a motorcycle helmet, somebody died from a helmet that was not actually compliant with federal safety regulations, and they bought it from Amazon. They bought it through a third-party seller. So this is a real issue, and it has implications, I think, for the folks who manufacture and sell products on Amazon, particularly some of the big established brands. What are your thoughts?

Andy Hoar: I’m 50/50 on this one. On the one hand, I think they can’t be absolved of responsibility for what they sell. On the other hand, it will significantly curtail what’s available. They won’t take as many risks. And so there’s good and bad about this, but no doubt about it this would be a monumental shift. If Amazon had to be responsible for this stuff, I’m sure the distributors are thrilled about this idea, because they are, if Grainger ships the exact same thing it shipped on Amazon. Amazon’s not responsible for it. Grainger technically is. They want an even playing field, but yeah, it’s going to be interesting when to watch.

Brian Beck: Well, I think too, if the other side of this too, is those companies out there that stand by their product, that’s the traditional manufacturers who, they stand by the quality of their product, it’s their key differentiator in the marketplace. This could actually be good for them, ’cause what Amazon would do, if they have to do this, there are billions of products on Amazon, Andy. And they would have to really go in and make sure that the sellers, they’re going to have to make some broad stroke decisions on some of these sellers. Just the order of magnitude of the challenge of addressing this issue, they’re going to probably have to make some significant changes in how they manage sellers and tighten that quite a bit. And so I think in a lot of ways, that would be good for those traditional brands, and ultimately potentially good for the customer of Amazon too.

Andy Hoar: Well, final thoughts, it might actually finally force them to clean up for the last time-ish, their rating stuff, which has been highly suspect. And so it might raise the bar for that too. So yeah, it would be interesting to watch.

Brian Beck: So today guys, we’re talking about – will AI change the game for B2B on Amazon? This is a really fascinating topic, Andy, and again, all about Amazon today. Well, for those of you not watching our on LinkedIn Live or watching the recording, those of you listening on the podcast, there is, we’re showing a picture of a dog. What the heck, why are we showing this?

Andy Hoar: That’s any dog too, a royal dog. These are corgis. – That’s a very impressive dog. – I think it’s a corgis. – A royal dog, that is a corgis.

Brian Beck: Yeah, short legs and really cute. Why are we looking at a dog? Well, Amazon, when they started, had a dog, well, it was an Amazon. It’s one of their former editor-in-chief and principal engineer, had a dog named Rufus that he brought, it was a fixture in Amazon. He brought him to work every day. He was called Amazon Shortest Volunteer Worker. He was, you know, actually called that. And he was in the office for, I don’t know, 15 years or so, and really became a part of Amazon’s culture. So what does this have to do with AI? Well, funny enough, Rufus is now the name of their new Gen AI-powered conversational shopping tool. Rufus is wandering around going, “Oh, I thought I was a dog.” (audience laughing)

Andy Hoar: Yeah, just like people know, we’re not in front of a live studio audience. That was a, laughtrack, very clever Brian.

Brian Beck: So Amazon named their AI after Rufus, after the dog, right? Why, hey, why not? It’s pretty creative. I think it’s a little too close to Dufus. (laughing)

Andy Hoar: I agree. I think that if it doesn’t work, all people are going to call it Dufus, right?

Brian Beck: Anyway, sorry, Rufus, but it is what it is. Here’s some context for everyone. Amazon traditionally plays in what’s called the tailspend or spot buys. What we mean by that is the customer, when a customer knows what they want, a B2B buyer, they want it in small quantities, the orders unanticipated, it’s uninfluenced by the procurement department or the parts of the business, and they need it fast. That’s where Amazon has built a lot of its volume, both in B2B, and to some degree, B2C too. A lot of B2B buyers will buy through standard Amazon. That’s 15 to 30% of overall B2B spend. We have some data we’re showing on the screen here from the Hackett Group from a few years back, but it shows, you know, kind of the, you know, where spend lies for B2B companies. And a lot of it, 40% is influenced by strategic sourcing or influenced by the buying desk where it’s what’s called more of the sort of the planned spend or, you know, more consultative or involved selling. So the question, Andy, is, you know, really around this new tool, does it have the potential to help them shift this. So tell us a little bit about what this Rufus is, Andy.

Andy Hoar: Well, you know, in the tradition of chat GPT, and BARD and the others, now Amazon has its version. Maybe a little late to the game, and as a matter of fact, but it’s a Gen AI-based product finding tool that’s been trained on Amazon’s product catalogs, reviews, information, et cetera, that allows shoppers and B2C buyers and B2B to ask contextual questions now. Conversational commerce, as they called it a couple of years ago. And to your point, what makes this particular interesting is it seems like it’s now Amazon also throwing down the gauntlet that they now want to get out of just being a finding company where if you know what you said, what you want, you go to Amazon, you can find it. They want to actually add to that the ability to discover products, which traditionally they have been weak on. We’ve done content about this before, where Google’s stronger on the discovery side, Amazon was just stronger on the finding side. Both want to get into each other’s businesses, but Amazon now with Rufus, we think, has the ability to do that, getting the discovery stuff. It really is kind of a recommendation engine on steroids, that’s the way I sort of think about it. Again, conversational commerce, and there’s a significant opportunity here for Amazon and the advertising side in B2B. This is the part that has been sort of flown below the radar, I think, with a lot of people in the last couple of years. Amazon has dramatically increased its revenue associated with advertising. So once you become a, once you reach critical mass, now you make money off of selling people’s stuff. You can make money off of sponsoring sales of things. Now you control the Buy Box, and so Rufus could do this, and we found an example of this. So I was just thinking in preparation for this, what if Rufus was looking for a coffee drip maker? And so there was an example of that where somebody asked Rufus, what’s the best drip coffee maker out there? Rufus instantly replied with Mr. Coffee, showed a picture, an image of it, and recommended it. Well, I thought, what if I ask Alexa? So last night I asked Alexa, and Alexa says, “are you shopping?” I said, “yes,” and Alexa said, “the top rated coffee maker by ratings is…” and it listed the name of some Italian coffee drip maker. Interesting that Alexa seemed to be more objective and unbiased about it, whereas it feels like at least in the outset here, Rufus was pushing a sponsored product.

Brian Beck: Well, I think we’ve seen precedent for this too, Andy. In Amazon, I mean, clearly, as you look at their most recent earnings report, just came out this week. You know, advertising is a key area for Amazon, and I think they’re thinking about ways to monetize this. So it’s going to be a fine line, to truly deliver this contextual, quote unquote, unbiased. But I mean, are traditional sales processes unbiased, Andy? No. So maybe that’s okay, and maybe that’s part of what the AI is going to capture is that. And by the way, Amazon collects so much information, behavioral data, obviously purchase data, but even think about things like Amazon Prime Video, and other behavioral signals that they can collect, they have an advantage on this kind of conversational commerce, quote unquote, like we would talk about it five years ago, but this sort of next generation moving up the funnel, getting into more difficult and complex sales, they have an advantage versus really anybody, because even over at Google, certainly over at ChatGPT…

Andy Hoar: They’re going to have to establish something, and B2C, hey, if they say the coffee drip maker, I bet I don’t like it, so what, right? But if you’re a B2B company, and they recommend a product that’s sponsored, that isn’t actually the highest quality product, maybe a knockoff from Asia, for example, if they’re not responsible for what they’re selling, and it breaks a machine or you get fired, that’s a whole other level. So trust is really key. So Amazon’s going to have to demonstrate that Rufus is a trustworthy advisor.

Brian Beck: That’s right, and then you get back into the first part of our session today, talking about being now held responsible for or liable for third-party product. So yes, you’re absolutely right, B2B has a lot of implications. We’re showing on the screen a couple of examples of Rufus in action here, it’s actually, and you made this point, it’s not integrated into the search experience, it’s actually a sidebar application. So it’s not forcing customers to use this, at least not today. But is this pickleball paddle good for beginners, provide some back and forth? What are the best toys for a dinosaur-obsessed five-year-old? These questions being put into this AI tool, are contextual, and there’s lots of other examples. “What do I need for cold weather golf”, or “what to consider when detailing my car at home”? I mean, these are questions that aren’t, “hey, find me this product with this SKU number”. So that begs the question, I pulled up an article from a few years ago from Inc. Magazine, Andy, that says, three reasons why artificial intelligence will never replace sales jobs. And what was interesting, Andy, this is a few years back, and they were citing your research when you’re at Forrester about the death of a salesman, B2B salesman, and this is a quote from the article, I’ll read it for our podcast listeners, “While some purely transactional sales positions move to the way self-serve model, jobs that involve high consideration products through a complex sales process will be enhanced, not replaced by AI.” That’s what they talked about, but as far as, what we’re talking about here though is AI really helping with that more complex sales process, potentially, is it there yet, no? But is it going to get there with this Amazon piece? And is this, in fact, the beginning, truly of the death of a B2B salesman, Andy, as you predicted 10 years ago, or whatever that was? It’s interesting, ’cause we keep going back to the death of a salesman research, I keep seeing it everywhere that you did, ’cause I think it was prescient.

Andy Hoar: Well, I remember when I was writing the research in 2015, I had no idea that we were going to have AI, or pandemic, or any of that stuff, right? I mean, we sort of knew AI, but it seemed like science fiction at the time, but it didn’t take a genius to figure out, as software gets more sophisticated, recommendation engines were getting better. Now you have this AI, and I think everybody was thrown for a loop last year with chat GPT because it took it up on an order of a magnitude level. I mean, it was that much more impressive. People saw this, and then it seemed almost like magic. Well, take that and apply that to the circumstance we have right now, and it does not string credulity to think that within a couple of years, you’re going to be talking to a chatbot, powered by AI, that’s going to be able to recommend products that have great ratings, that they’ll be able to bring in social signals, be able to tap into what’s going on in communities, not to mention, like you said, just droves of information, that they’ve had over sales, over use cases, by use case over years. So no humans can be able to compete with that. Now, what I think is going to happen is AI is going to augment or enhance, like the Inc article said, initially, but I don’t agree with that statement that that’s an end state. The end state is it will replace what humans are capable of, it just has to because it’s so much better. It’s so much more sophisticated.

Brian Beck: And think about this too, you know, if you’re a B2B application, let’s say you’re installing an HVAC system and you have a specific application for that HVAC system, as Amazon deepens, its reach into B2B with Amazon Business, you’re going to get more and more customer feedback, Q&A sessions, purchases by specific types of customers. Amazon’s going to mine that information, to then understand, across millions and millions of transactions, to then understand what is the best application and be able to communicate that through AI, to a buyer who’d ask the question, “What is the best system or component for this application in my HVAC system”? So, you know, it has enormous sort of implications, I think, for the B2B cycle.

Andy Hoar: Oh, it does. And actually, there are a couple of interesting parts about it. One is that I think, initially, it’ll be humans that are backed up by AI. So, humans will make recommendations, and AI will check the humans to see if it’s correct. But over time, that’s going to flip. And eventually, it’s going to be the AI making the recommendations and humans will be double checking it to make sure that it works. Now, the reason why this is almost like good timing, it’s not good timing for people who want to have these jobs, but it’s good timing for B2B because many of the people in B2B are trading out of the workforce that knowledge they have about applications is disappearing as these people leave. And so, you can’t hire a 25-year-old who’s never installed insulation and teach them the nuances of decision making, as well as the guy who was retiring and had that. So, AI is necessary because I hear this all the time from companies, they’re really struggling to find A) enough young people who want to do this and B) enough young people who have knowledge in the space. They’re really hard to find.

Brian Beck: I wanted to just get through our LinkedIn poll. We asked our community, “Will new generative AI tools allow Amazon to capture sales of more complex B2B products in its marketplace?” The answer 75% was yes. So, clearly the community agrees that, at least to some degree, you know, this is going to impact Amazon’s ability to climb up the funnel into some of those other categories of spend we talked about earlier.

Andy Hoar: You know, I think if we go back in history, but what have been an interesting thought experiment. If we’d asked this question a year ago, what would the number have been? It probably would have been much lower, but a year later of Chat GPT and BARD, et cetera has changed the way people perceive this.