10 Best AI Use Cases for Ecommerce in 2026: Ranked by CEOs

22:57Sean Frank:Yeah. Look, this is an aside, different episode. We should talk about how expensive things have gotten. Because I remember in 2020, if you offered somebody a$60 ,000 a year job, that was good. You know what I mean? And now, like, it looks like the minimum you have to pay somebody who knows how to use a computer is$100 ,000. And it's like, maybe this is just me being a boomer now, me getting out of touch. But, like, I remember making$100 ,000 for the first time being like, this is crazy. I'm so rich. And, like, I think real inflation has probably been 10 % a year for the past six years. Because now I've got people -

23:31Matt Bertulli:Yeah, but 10 % for six years compounds real fast.

23:33Sean Frank:Yeah, I got people asking for, you know,$180 ,000 for jobs that I used to pay$80 ,000 for. It's crazy. Matt, you know, our friend Shereen had a great tweet on this subject. She said, the biggest threat indicator that AI can take your job, someone needs to tell you what to do. If someone needs to tell you what to do, how to do it and when to do it, you should be very concerned about your future and find a way to fix it. So that was, Matt, that was your last one. I think we've all kind of like maybe reconsidered our stance. Jason, you moved us around. That's good. Matt, what was your top one?

24:05Matt Bertulli:My top one is actually, so I don't know if you guys, I had a really hard time ranking. the top five because they're all so close in their value to me in the organ like in a company but i went with my number one was internal knowledge base because i think the more context and more memory that these things have the more powerful they get and i actually think that competitive advantage right now i mean we talked about this years ago with ridge that because sean has sort of enforced this like write everything down mentality because they're a remote company, that that has actually made them much, it actually enables them to go much faster in adopting AI because like they have an incredible amount of like internal intellectual capital that is documented in a way that is AI friendly, right?

24:54Matt Bertulli:Whereas like most things are in people's heads that's not very useful to AI. So for me, I chose this one because I think it sets up the rest for more success. personally like having a a Hermes agent with a G brain and a chief of staff that like literally has a it's like a twin of my brain everything I do digitally is stored in this G brain thing is incredible leverage right and then now doing that at a at a company level like standing up a new company from day one and focusing first on like getting context into the place that AI can use is making the content creation piece better, market research better.

25:35Matt Bertulli:Like even how do we want to set up data architecture in this company? Like we need to know how the company is going to run. Who are we serving? Like all that stuff seems to be getting better with more context. And then I think shared memory and shared skills and all these things like AI is a single player mode. No question people are getting a lot of value out of that. But I still think AI multiplayer player is a big unlock for companies. So that was my number one. That's why. Jason, what do you think about that? Dude, I just went for a 20 minutes. How about we ask Sean? Jason says 20 words. He's like, I'm done, guys.

26:13I'm done. I'll see y 'all next week. I'm just trying to let everybody get their chance here.

26:18Sean Frank:Yeah. Unfortunately, I think I have talked the most, but I'm happy to share that. Matt, we got so lucky that we embraced being fully remote. And I do not think in-person teams have any advantage in the age of AI, simply because by being remote, everything is recorded, everything is written down, everything is documented. And this is like, it's better to be lucky than good. We fell so backwards into that. But now I have everything for the past six years recorded that I can feed into an AI model. And that's why I said, I hooked it up to Notion. I just said, make me emails. I don't need to go hunt those down.

26:58Sean Frank:I don't need to go pull them from my favorite website or whatever, because I have six years of them documented in a folder somewhere. And every conversation, every Slack is just, it is a huge advantage to have the, everything be documented right now. And I fully agree that it's kicking ass.

27:18Matt Bertulli:Most AI tools and software don't give you direct ROI every month, but Postscript does. I run two Postscript AI products at Ridge. Both show up on our P &L. Infinity Testing runs continuous A-B tests on our SMS automations completely in the background. It's driven$446 ,000 in incremental revenue, not just attributed, incremental. That's a 32X incremental ROAS with click-through rates up 43%. Then there's Shopper, an AI sales agent that answers every inbound text in under 40 seconds. We're running 23X ROI on messaging costs alone. If you want SMS software that gives a proven ROI, go to postscript.io and book a free demo.

27:56When we first started embracing AI, like the first thing I wanted to do, and it was sort of, it was this internal knowledge base. And like, you just have it, Sean, because you documented all that stuff like really, you know, really early on. I actually don't even know what the status of that project is. I gotta go check on it. I just asked some people to do it. But like, seriously, there's, but we're not going to get into remote versus in person here. I'm going to not do that. Well, we don't need to litigate that, no. But I do agree that it's a benefit that we were not thinking about two years ago.

28:30We've had a lot of discussion about this and I would say for sure, it's something that has changed the calculus. Even as somebody who's kind of been like, Sean, you're just wrong on this. It really has changed the calculus. One of the things that makes me think about, there was a point in Amazon's history where Bezos sent out a memo and basically the memo was like everything is going to be externalizable. Really what he was saying is everything's going to be API based. It's just going to be. And if you don't like that, that's fine. You're fired. And he just said, that's the way it's going to be.

29:03With everything inside of Amazon, it has to be API based. And that single decision led to AWS and probably the core of the modern internet in the cloud and i think in a same in a similar way in your company you have to put your foot down that we are going to document our context you have to because what is really depressing is when you look at your organization and you kind of say what am i paying i'm buying all these hours from people what am i actually getting and you realize like 40 or 50 of those hours are just people trading around pieces of information and for the first time you can really be like, like, for example, Jason, you're like, I don't know where that project is.

29:43We're hurtling towards a future where it's like, you just ask Claude where, or, or, you know, chat GPT or whatever codex. Hey, where, where are we on this project?

29:51Sean Frank:Yeah. And if you are in person, the one thing I would do, cause I had an in-person meeting yesterday. You have to record it.

30:00Matt Bertulli:Guys, these things, like that was why I said, same thing as Sean, what an unlock to just do that.

30:07Sean Frank:Just carry one around. yeah and like i would i would make a mandatory across every office every meeting granola notes because like we had a big meeting maybe a week ago we're talking about like uh some very nuanced product decisions like what type of product this is or whatever and we recorded the whole thing and then i just said hey claude turn this into a one-page visual chart and it did and then i can now i have like a little poster i can send everybody hey here's how we made these decisions It's a visual graphic you can easily understand.

30:37Matt Bertulli:Yeah, I mean, look, that is why I have one of these. This is Applied. It's a digital recorder. It does like automatic AI summaries. It gives me my transcripts. My chief of staff can just grab everything from this every single day and store it in my digital brain. I highly recommend that if you are in person, you get one of these. All right, Matt. So we've talked about an internal knowledge base. Give us your entire ranked list. Yeah, so, and again, I think all of us are a little colored by like our experiences in our own companies. So my list is internal knowledge base, number one, content creation, number two, market research, number three.

31:15Matt Bertulli:I'm a copy guy. I spent a lot of time thinking about personas and angles. So like, this has been wildly valuable just to be able to research what customers are saying. data and reporting was number four operations was number five god i had a hard time with this mike like operation is so valuable in supply chain uh customer service six website and cro seven paid me to eight nine was product development ten leadership and people yeah i had a really hard time i could rearrange like six of these and argue all of them yeah maybe that's an indication to know where we're at, that it wasn't, I mean, I think a year ago, if I would have looked at this list, I would be like, I feel really confident in one or two of these use cases.

31:57And even in this conversation, the ones I ranked near the end, I'm kind of being, I'm having my perspective changed on it. It's just much more broadly helpful. So Sean, let's go to you. I want to hear your list. What was your number one way to use AI in your business today?

32:13Sean Frank:Yeah, it's the same one as Jason, data and reporting. Both of us are on the Ceres analytics stack. And I've talked about that and what that means. Ceres is the data warehouse we use. All that means is they take all your data from all your sources and then clean them up. What data? Product type, like product SKU. you know it's very common to be selling one skew on amazon and one skew on dot com and one skew in wholesale and then they're all functioning the same thing with different names and different skews so saris comes in and cleans all that up um and then once they clean all that up they build you a tableau dashboard but i don't care about that at all they have an mcp that hooks up to clod so you have all of your company data that's incredibly clean and then you hook it up to clod when i say company data you have all of your sales all of your marketing spend by channel You have all of your expenses every single day coming in there.

33:08Sean Frank:And so you get perfectly accurate, perfectly malleable data bytes every single day that you can sort and cut however you want with words in Claude. And it has cut reporting timelines of Ridge down to literally minutes, right? I used to spend, I used to by hand do every presentation at the end of every month being like, hey, here's what we did. Here's what we closed out. And it would take me one to two days, right? Every quarter I would do the quarter summary and it would take me three to four days. And now it's 17 minutes. I just go in there and say, hey, rerun this for this new month with this data.

33:47Sean Frank:So when I say it's like the number one use case, it has saved me personally the most time. Everyone in my company can now pull whatever data they want. I think data is a big bottleneck when you're trying to make a decision around product upsells, product pricing, MSRP, sell through. literally seconds, literally one ask of Claude because it's hooked up via MCP, it's totally changed the entire department. That is my best use case. I am the most passionate probably about this. I didn't rank this number one. I ranked it number two. But one of the observations I would make is that all of us tend to use AI in our wheelhouse first.

34:22And my wheelhouse is like analytics. And what I would say is that I am using it now to produce outputs that are better than what I could do, regardless of how much time I had. And that was pretty shocking to me. I thought that I would probably be able to automate and produce similar quality outputs to what I could do individually. I could just do it a lot quicker. Let me give a really simple example of this. So with Trevi, one of the biggest things we're trying to do is we're trying to predict how many returning users are you going to have every day? because these consumable businesses are all about recurring users and understanding your LTVs and what you can pay.

35:02So it's a pretty big part of the business. And before, I'm trying to do it on cohorts, and I have kind of these retention curves, and it's taken up all this space in Excel. And it's frankly not very good. And over time, what I've done, first, my model has moved from Excel, where I update things, to its Google Sheets, and it updates every single hour and everything's automated. It's constantly pulling in the most recent data. And what I did is I stood up all of these individual parts of the model. And so returning usage behavior on Amazon would be one of those parts of the model. And I'd stand it up and be like, hey, you know, we want to build a model for this, build me a model for this.

35:43I was working with Codex and so we built a model. And then I'm like, okay, great. Now I need a 60 % improvement. And you should consider these three or four factors you might not have considered. And what they'll do, the models will do is they'll build a kind of like a theory of like, hey, here's how this could work. And then they backtest it. And they have very mathematical ways of backtesting it and saying, if this were the model and I backtested and I went through all the data, how much better of a prediction system is there? And then I just beat the thing. I'm just like, okay, great. Now make it 50 % better.

36:15Okay, great. Have you considered this? Make it another 30 % better. Okay. It's not good enough. Do it again, do it again, do it again. And I basically built this thing to where at this point, what it does is it looks at every single user. It looks at what they've bought, how many sticks they bought, how much the total card size was. Didn't they buy on deal? Did they buy on a weekend versus a weekday? What flavor did they buy? All of this stuff. And it projects out every single user and what they're going to do in the future. And then it rolls it up into one forecast and it does it in 30 seconds.

36:44And mathematically, it just blows out of the water what I had. And it's just like a small, like micro example of how like our ability to understand our numbers for all of our businesses is just going to go to an unprecedented level because the amount of compute, the amount of intelligence that you can throw at any individual problem is just so much greater than before. If you had told me that there would be a tool that would get me where I didn't use Excel, I would have said, I don't believe that will happen in my lifetime. And my Excel usage has dropped off a cliff because I'm doing so much more of it through the AI now.

37:23If you're scaling an e-commerce brand today, ads alone aren't enough. After Sell focuses on the one moment that every brand already owns after checkout and turns the post-purchase moment into more profit. Monetize every order with post-purchase offers and thank you page experiences without disrupting checkout or hurting conversion. Enterprise-grade tech used by Gap, Ticketmaster, Macy's, and Target, now driving results for brands like True Classic, Hexclad, Ridge, and Jones Road. I would know. This is the reason I ended up buying three pans from Hexclad instead of two. After Sell has already generated over$1 billion in additional revenue for e-commerce brands, revenue that doesn't require more traffic or higher CAC.

38:03So check out After Sell and tell them that the operators sent you. Sean and I have very similar use cases here because we both were early adopters of data warehouse because we were smart enough to see the future. You know, I think Sean has had like major, you know, I have a major unlock with, like I get a daily report. I get it basically because, you know, I just get it, I get a daily email. I get a weekly email. I get a monthly email, like all the performance effects that just pulling straight out of the AI, out of Saris, and with MCP, with Claude. We also, my finance and accounting department is now really starting to leverage it as well to get better reporting.

38:47So, I mean, that's just huge for us.

38:49Sean Frank:Connor, my CMO, said the Saris MCP will go down as the biggest thing we've done this year, just because it has saved the executive team across the board, easily 40 hours a month of work. And then everyone downstream saving 100 plus hours because a lot of their job is reporting, pulling in graphs, checking data, seeing sales, making those tweaks. And just to having it instant, like I really can't express the magic. And it's so hard because I want to show people, but it has my sales data. So it's really hard to go out there and do a Twitter post being like, look how cool this is. But anyway, here's my real sales data, which I'm not going to do.

39:32Sean Frank:So anyway, it's pretty awesome. You have to have a data warehouse. And there are so many downstream benefits of that. I think that's one of the reasons why this is obviously at the top of the list. And the gap between me having an answer to a question, even a very complicated question, is now seconds or minutes and not days. So, Sean, what was the worst thing on your list, the least helpful in your experience? I chose product development. It kind of sounds like you guys are treating market research also as product development. I was treating it more for marketing. When I look at my list, I'll go top to bottom.

40:10Sean Frank:Data reporting, I think, is a must-have. I think customer service is a solved issue. Half the tickets this year would originally be answered by AI or more. Content creation, we've battled that out. I think statics are entirely solved. internal knowledge base i think it's incredibly important and with hq which is like i don't get into like specific tools but notion is a great way to build ai context but like how do you share that across users i think hq could do that and then i have market research and it's amazing for finding out you know what to sell who to sell it to who your customers are um but then at the bottom of my list i have i have product development is my number 10 it's because ai is not good at cad files yet it is not good at um you know like actually making tweaks inside of product files like it could probably do it but that is those are very fine details you need to actually do right it can help you maybe think about a category to sell into but it's not going to help you think the you know the the way ridge needs to approach that angle and every time it tries to do it it's very ham-fisted.

41:17Sean Frank:So it's my least favorite one. My experience with that, Sean, is that coming up with a concept is much quicker now with AI. But the problem probably here is that there's so much context that AI doesn't have about the actual physical world and making things for us. So like, for example, there's all these like rules about tooling and all of our lids come through tooling. And so we'll come up with ideas all the time that we'll send it to China and they'll just laugh. They'll be like, yeah, that's cute that you would think that would work. And like, I think this is a good example of AI. AI can dream up stuff that doesn't really make sense in the physical world because it just doesn't have the context of the constraints that you have around the tools.

41:59Like it might make sense in a CAD file even or conceptually, but it just like doesn't practically work for whatever reason. And, you know, like you can design the coolest lid ever in AI, but if it leaks, like who cares? if it's not durable, who cares?

42:12Sean Frank:Totally. And we've all had that experience working with, you know, there's product designers and then there's design for manufacturing and it's different things because the product designer will be like, it's one perfectly molded piece and it's all made out of aluminum and it shines, right? And then the person who actually does design for engineering or design for manufacturing, they're like, yeah, so it's going to be$80 ,000, but it's eight cents if we make it in four pieces. It's like, we've all had that. So Matt, Jason,

42:40Matt Bertulli:either of you using this in product development actually no i'm with sean in in in our world it's borderline useless outside of i think ideation uh so like our i mean our product team is sort of broken up into like there's the weekly design drops so like what else are we putting on the cases so i think for that it's been actually really helpful like to explore the fringes and edges of like interests online but i think that's less like hardcore pd which sean is hitting on uh but no like mechatronics uh electrical engineering like all that stuff no nobody is genuinely using it in their i would say like their actual work right um it's just not good there yet yeah and jeff bezos

43:25Sean Frank:just raised like a bunch of money to go out there and build like the ai model for cad so like maybe maybe this is all coming, but like, yeah, we use it to make a lot of color choices or whatever. You know, my bottom of the list, in number eight, I have operations. Number nine, we have leadership and people. Number 10, I have product development. I want to talk about operations because it's Mike's number one. And I'm like, how are you getting value out of this? Teach the audience. Well, here's my thought process in putting it number one. In operations, so we're talking about things like fulfilling POs and moving things around with logistics and placing your purchase orders.

44:05I mean, I would put inventory and like planning kind of in here, Sean, which probably would change your perspective if you grouped it in there. But it's like all of the blocking and tackling to keep items on shelves, to keep things in stock, basically, is the way I think about it and getting product to the customer. And the thing about that is, Number one, there's really robust digital context, basically on all of that stuff at this point. There's APIs. You can get all the information. So you don't have a context deficit. And there's not super intricate judgment going on, usually. There's not a lot of qualitative pieces that you can't teach the AI to have.

44:43And so, for example, like inventory and planning, which I know is one you've talked about, you can just kind of solve that. Like, hey, here are all my constraints. And here's all the pieces of context. And I want you to come up with a plan. And it's just not hard to move your organization towards the way that we're planning out our inventory, the way that we're placing our purchase orders, the way that we're putting it on ships, the way that we're moving it to 3PLs, the way that we're sending it out to POs to our vendors is just like clockwork. And it's much more automated. And the reason why I think that's number one is because it has to happen.

45:20Like your business basically dies if it doesn't happen. that there's real money to be saved and that it removes all of this complexity from running your organization. Like the more that you can just focus on product and marketing with your organization, those are the growth factors, the better. And it just totally, it totally takes operations to much more of like a solved game. So I've been amazed at our team internally, the tools they built, the automation, just in the last few months. We have an internal app builder that we basically have like a context layer. And then on top of that context layer, you can very easily build apps.

45:57A lot of the most useful apps have been built by our operations groups. So I'm just very bullish on this. And I think you're just going to be able to run world-class operations with less people. You're going to be able to run them much more inexpensively. You're going to be able to have much more clarity on what's going on in your supply chain at all times. And those are kind of the lifeblood of a business. I agree with you, Mike. You know, my team in ops, look, fulfillment, you know, is a big thing, right? And we are, we're actually trying to, we're constantly trying to take down our percentage, you know, of COGS or percentage of revenue fulfillment costs.

46:40And we, we're global. We have many distribution centers. We are having to get stuff in and out of Amazon. You know, we've got many, many, many containers on the water because our products are large. And there's a lot of stuff moving around. And I know that my operations team has found like major, major savings and major, you know, just efficiencies through this, through Claude, through MCP, using Fulfill to do that. it's been like a huge, these guys are thrilled. Like seriously, these guys are so happy. I love to see it.

47:27Matt Bertulli:Most AI tools right now are all promise and no delivery. You know exactly what I am talking about. Super fancy launch videos. This is why I'm loving what Rich Panel is doing. They have AI and support smashed together, made it practical, made it useful, made it valuable to brands today, not on some future promise. And instead of asking you to spend weeks writing prompts and uploading help docs and babysitting, they flipped the whole script. Their AI builds your support team for you and everything it needs. I've watched it. It works. It's freaking amazing. RichPanel even guarantees 50 % of your support volume will be automated by AI within 30 days or your money back.

48:04Matt Bertulli:I call that a no-brainer offer. The pricing is also kind of wild, about 20 cents per conversation instead of most other vendors being like$1 to$3, which is a little nuts. If you are interested, go to richpanel.com slash demo, not for some generic demo, but to actually book a call and watch them build your actual support team.

48:22Sean Frank:I didn't think about the inventory planning or buying. Because once again, I think that is a solved issue. Like inside of your organization, I was even thinking about it as like a data problem. But you're right, if that's bucketed under operations, I was thinking I'm like pick, pack, and ship. I'm like, how's AI helping you with that, right? And one of the themes I think that's come up in this conversation is AI has not punctured the physical world yet. and that that is coming. You know, like the next phase is going to be when it really gets into the physical world and it really can help in product development.

48:54And you do have robotics and there is this extension, but we're not really there yet. So I would agree with that, Sean. But yeah, I think there's a lot of, I guess one way of thinking about it is if you say data and reporting is really valuable, it's conjoined twin as operations, right? Because what you're doing with the data and reporting is usually you're turning that into operational decisions. And I kind of see them as two sides of the same coin. And I think that that's why I put them one and two is that it's like the information about how we need to, like what we need to actually do and then the doing of those things.

49:28If you can use AI to be better in both those camps, like, man, you're going to be in great shape.

49:33Matt Bertulli:It has, and the operations thing, like, so we run our own factories and we do demand planning for us is almost daily, right? And Claude is now doing that just because we have data warehouse we have all this stuff like we used to our process was sort of like our factory managers let's just whatever that title is every monday they would sort of have their own demand plan for the week it's like these are the tools we need to run and these quantities and these colors and that was how we kept the whole pila like on-demand inventory thing going is we were all kind of always trying to be like a day or two or three ahead on what we thought would be needed right by our various channels ai is doing that now and it's far more accurate our efficiency is actually quite a bit up at a factory level uh so i would i would bucket that as operation sean um but it it is like it is like mike you're right that is more data it's just that it's capable of processing a lot more information than our team could do it like humans could do it before.

50:35Matt Bertulli:Like it can look at every single order. Whereas before we would look at like aggregate data because that's how a human would process it. Right. It's like, show me the averages, show me the summaries. This thing is looking at like every single order and every single item in real time and making better choices for our for our factories. So I have a take on this, and that is that everything in your business will eventually be like a prediction market. The prediction markets are these amazing things that suck in all this information and can say at this given point in time, this is the most probable outcome.

51:07And we've seen that they are really accurate. And a lot of what we're trying to do with our business, like with our demand forecasting, for example, is basically that. It's just a really inefficient market that is not doing a very good job of sucking in information and bringing it up to the second. And with these systems, you're just going to be able to be like, yeah, plug into all these data sources and then just give me like as of right at this second, what is the best prediction of the future and work around that. And our businesses have not been anywhere near that agile, but that's what's coming.

51:37So as I said earlier, my last one was leadership and people. I rank that as my worst. I think that talking to Jason, if I think about it from a different angle, I can see that differently. So obviously using AI to manage people is probably terrible. Don't do that. But obviously, it's pretty profoundly going to impact the way we think about leadership and people. Here was my ranked list because there's one of these I wanted to call out. I know we've discussed as a group a lot. Number one, operations. Two, data and reporting. Three, customer service. Four, content creation. Five, internal knowledge base.

52:13Six, paid media. Seven, market research. Eight, website and CRO. Nine, product development. And 10, leadership and people. But guys, I know that for each of you, I think customer service has been the most transformed to date by AI. And it's just really remarkable what we've been able to build internally. We have a bot and just the volume of our tickets it does, how quickly it does it. You know, one of the sponsors of the pod, Rich Panel, has unbelievable offerings in this area. Has that been as true for each of you and your organizations that customer service has been dramatically improved by AI?

52:49Oh my God, yes.

52:50Matt Bertulli:It's not even, it's freaking insane. It just like, even just think of it this way. Even if you didn't have AI that was capable of answering all of the basic questions, which it is, it can handle like 98 % of our tickets. It's still gonna make all your agents way smarter. Like, you know, Mike, you've already said, like you can ask it a question or Jason, you can ask it a question and in seconds, it's giving you an answer. That applies up and down the org. So why wouldn't it impact customer service the same way?

53:15Sean Frank:Me and Rich Payton have a bet that by the end of the year, they'll be doing over 50 % of my tickets with AI and they're getting, I think they're there right now, right? So totally AI, top to bottom, resolving tickets for us and customers seem to like it more, right? The AI chatbot has like a 98 % CSAT score. So like everyone's having a good time talking to the AI chatbot.

53:37Matt Bertulli:It's because it's just giving them money, Sean. That's what's happening. It's like sending out$100 bills without you really knowing.

53:42Sean Frank:I should dive into that. Maybe I do have a free money bot. Well, I said this before, but when we first implemented this, The biggest concern is, are we going to be able to nail voice with this AI chatbot? And very quickly, we're like, wow, this thing is way better than we thought. It's like you said, Sean, it's getting better CSET scores than we would have ever expected. Sometimes like somebody wrote in about, you know, like an Oklahoma City Thunder item that we had that was sold out. And the AI chatbot noticed that they were writing from California and was like, yeah, we know it's tough to be a Thunder fan, you know, with all those Warriors fans around.

54:18but we're going to get this back in stock later. I mean, it was just like riffing on things in ways we wouldn't have expected. And we actually learned that like the worst part about this transition to AI for us was like this thing had no chill. So it would be like somebody would send in a support inquiry and like two seconds later they would get a response. And so it was like, hey, you know what? We call it Hallie. We're like, Hallie, you got to chill out. You know, like we got to give them like a few minutes before you respond. And so if that's your biggest problem is it's being too prompt and responding to customers.

54:49That's a great problem to have. I want to talk about AI job displacement for a second.

54:53Sean Frank:You know, so let's say American hiring was a curve like this, right? And like, it was a steady line on the graph. And with the rollout of AI, maybe the curve has inflected down a little bit. So like we're hiring less people than we would in a world without AI, but job openings are still happening. I just told you that we're gonna hire 10 people here, right? So like there are still new people being hired. Maybe it's at a less frantic pace than if we didn't have this AI boom or whatever. But where that's definitely jobs are going to be lost, it is international CX agents. Those jobs have already left America, right?

55:31Sean Frank:They've moved over to the Philippines or whatever else. And now those jobs will be totally eliminated by AI. So do not hold Manila real estate is what I'm saying right now. It's like there's - Hold on. Let me divest. as we're speaking, not investment advice. Or, you know, we just talked about how powerful it is with Excel. And Mike, a top 1 % power user, hasn't logged in this year, right? And, you know, that's going to be very, very bad for the knowledge economy of places like India, right? Where big banks have outsourced lots of their, you know, computing work, right? Their manual data work to India, Pakistan, wherever.

56:10Sean Frank:and now they're just going to do it locally with Claude for probably the same price, right? It won't be cheaper to use Claude, probably the same price, but it'll happen instantly, so. And it'll be better quality, you know, like, and you have all your contacts. Like, it's an overwhelming value proposition, Sean.

56:26Matt Bertulli:Is it an on-shoring of productivity? Is that how you guys look at that? Like, so much productivity was outsourced, is that re-onshoring? AI is not going to lead to job loss. It's just going to shuffle the deck. And I think if you want to get hired at a bigger company, I think your job prospects in the next 10 years are worse. But I think you're just going to see so many more people start ventures of their own. You're going to see so many new smaller firms. And part of the reason for that is that if you think 50 years ago, it was difficult to compete with the incumbents because they could just do things that you couldn't do.

57:03And there's actually a great post, Aaron Levy, great, that's his name, right? Box. uh aaron like so there's this javins paradox thing and he talks about the evolution of marketing agencies and that basically for a while like you had to go with one of the big boys if you really wanted the best stuff because they had these huge creative departments and you couldn't get the level of creative output if you didn't go with one of the big boys because the scale like created a competitive advantage and that you would have thought that the invent of photoshop would have led to less designers, but it had the opposite impact.

57:39It just exploded the number of designers and it made it now where smaller design firms could compete with the really big boys. And this is exactly the same case. To have a really great customer support department 15 years ago, you had to be at whatever size, X size. Now you can be one hundredth of that size and have that same level of customer support because of the tools. And all of this favors small businesses, basically, and entrepreneurs and new venture creation that I think we have great customer support. But you know what? You could start a new company tomorrow and probably have a similar level of customer support because the tools that are available and that there's just so much opportunity in that.

58:21Matt Bertulli:Black Friday and Cyber Monday are coming. Are you ready? Not in the normal marketing calendar offers prep sense. I'm talking measurement and attribution? Are you even going to know what winning and losing looks like when the big Q4 season hits? Reality is the businesses that win set up their measurement solution when the times are slow, not the night before the big weekend. There's a reason for this. You see, with NorthBeam, you're getting a totally independent, completely new set of ad performance numbers with infinite look back windows. That means that the longer that the data grows, the more valuable it becomes.

58:52Matt Bertulli:So when CPMs are spiking in November, you will know precisely where every dollar is coming from while your competitors are guessing. This is important. Black Friday is one in the summer, not the day before. Book a demo with North Beam today. The link is in the description. Go check it out.

59:11Sean Frank:Matt asked if it's going to be an on-shoring of productivity. Well, it depends where they build the data centers because that productivity might go to space. But no, I do think the job loss narrative is a lie because those jobs have already been lost. They've already been shipped out of America. And I definitely agree with Mike that Facebook will have less employees in 10 years than it has today, and the company will be way bigger. They're going to get hyper... I'm talking about$10 million per person at Ridge. They're probably going to talk about a billion dollars per person or something crazy.

59:45Sean Frank:They probably won't just because of the real estate investment, but I do think they'll have less people. So it's, the economy will be bigger, it'll be more dynamic, and it's going to be a way more smaller firm. I believe in that future.

59:58Matt Bertulli:And where do you guys think that we're going to be competing as brands? So where is competition going? Six months from now, 12 months from now, 24 months from now, like, and I'll give you a bit of a frame on this. I had somebody send me a message on X around creative, you know, and his question was something like, if six months from now, 12 months from now, AI is capable of making video, right? Which it's probably going to be in such a way that everybody has the ability to make hundreds or thousands of pieces of creative all at the same time and upload those into ad platforms. It's like, where is the competitive advantage in getting attention and getting distribution?

1:00:39Matt Bertulli:Like, that's kind of the game we're all in, right? It's like, how do we acquire a customer profitably? How do we make our unit economics work? I'm curious if you guys have any takes on this, because that is, I didn't have a good answer. So, you know, uh, help me. What is, what is your take on this, Sean?

1:00:55Sean Frank:Well, my take is if we're in a world where anybody, so Chinese factories or black ad dropshippers can make any beautiful creative they want, they can impersonate any celebrity they want, and they can put their product in those ads. How do you continue to get sales and transactions, it comes down to, do you have more budget? Because what we're forgetting is there's still a delivery mechanism. The delivery mechanism is Facebook ads. And Facebook wants to charge a higher and higher CPA and CPM every single year. And the problem with this competition coming in is, will they have everything in place to support a$100 CAC?

1:01:44Sean Frank:And that, like, the maturity, the reason why you have to lower OpEx to put as much money into ads as possible, right? And like Jason's talking about lowering cost of shipment to make sure he gets as much money into ads as possible, is I think we're going into a future where capital is the thing that lets you win, right? Which sucks if you're a new, if you don't have a business and you're listening to this right now, but at least in the consumer space, you know, we're going to a more nationalized system. So there will be tariff borders going up. It'll be hard to get direct to China shipments coming in.

1:02:14Sean Frank:So you have some sort of protected moat there. And then just naturally, those people are going to try to win on price. But if you win on price, you can't win on CAC. And I just think we get to a world where there's a floor price of everything sold on the internet and it's just going to be higher than you think. Yeah. Scaled economies and some of the traditional moats, I think, are going to come back to mattering. And like one example in our business right now, there's parts of our business where I'm like a little bit concerned about, you know, how competitive forces impact it. And there's other parts where I'm like, I feel great.

1:02:46Perfect example. We're doing, we're embroidering like a thousand units a day to 2000 units, somewhere in there right now. And we just bought four more embroidery machines. I think we have something like 12 to 14 embroidery machines. That's just a competitive advantage, right? Like that's literally like, you got to have the machines. There's very few people that embroider are at scale in the United States, ones that do often have like one month ship times. And so like, that's just an asset that is not easily disrupted by AI because it lives in the physical world. And actually what I'm thinking about is how do I do more embroidery?

1:03:21Like I've got this competitive advantage. How do I spin up more businesses around it? But you're going to need to be thinking that way. I think Matt is that like some of the things that were really helpful in the last 10 years won't be in the next 10 years. And then the one other thing I'd say is like, I think brand's going to be harder than ever to build. And I think it's going to be more valuable than ever. I mean, brand is the ultimate equalizer that somebody else is selling. Yeah, it's like, I mean, everybody can go and build a water bottle. Hell, Sean did it, you know. I'm just kidding. I wouldn't cut that up.

1:03:56But the, like, anybody can go and order a water bottle from one of any really good manufacturers. And so like I said this to my team before, if this means something, if this SM at the bottom means something really bright future. And if it doesn't, we're screwed. We'll eventually go out of business. The end, you know? And so like brand building is going to like e-commerce, it kind of bifurcates into you've got kind of drop shipping, kind of like what is the opportunity of the moment? Hit it and get as much money as you can really quick. That'll always be there. But if you're like talking about like enterprise value creation, it's got to be brand, right?

1:04:32Sean Frank:And I think we should do a whole episode about mean lunch and water bottles because I would recommend absolutely nobody do it right now. It's like, how are you going to stand out? It's the craziest market on earth. Like, you know, it's a knife fight. But Ridge can do it because it has people showing up every day on its website. It has some sort of attention already. We've been talking about use cases with AI. We ranked a list. And you're probably wondering, hey, where did you get that list from? Well, actually what we did is we took the operator's knowledge base, which you can go to. It is now online.

1:05:07We will leave the link in the show notes. And it literally has cataloged everything we've had on this podcast, everything you've heard on Mops and any of the other operators properties. It is incredible knowledge base where you can go and say, hey, I'm wanting to know, like, what have the operators said about this or that? And you can immediately pull all the quotes, look at everything we've said over the years. It's insanely helpful. I advise you take a look at it. But we went there and we just said, what are the top 10 things we've talked about when it comes to AI? That's how we got the list.

1:05:37That's what we ranked today. Find ways to apply AI. You're going to drive more value. Keep coming back to the Operators Podcast. We love having you with us. And we will see you next time.