← Marketing Operators 
Marketing Operators
E067: How Marketers Should Actually Be Using AI, with the Boring Marketer
01:26:35
Loading summary
Transcript127 lines
- [00:00]
James
We've got a special guest today with the best Twitter handle out of all of us at Boring Marketer.
- [00:05]
Connor Rolane
Thanks for having me.
- [00:06]
James
We're going to talk how ecom marketers can best leverage AI.
- [00:09]
Connor Rolane
When I think about agents, I think about an LLM or an AI model that has access to a bunch of tools already, and you're just entering the situation with like a prompt or a broad idea or maybe even nothing, and it's just going out on its own. Any AI tool, Manus or Lindy or Gumloop, Greg's probably popping up with YouTube.
- [00:28]
James
You can't learn about AI without watching a Greg Eisenberg video right now.
- [00:32]
Cody
Just constantly recommended if we could develop all of our paid media landing pages just using five coding, like, how big of like an unlock that would be for our organization.
- [00:40]
Greg Eisenberg
Not using AI is the same as, like sending physical mail instead of using email right now.
- [00:47]
James
All right, we're back with another episode of Marketing Operators. We've got a special guest today with the best Twitter handle out of all of us at Boring Marketer. James, thanks for joining.
- [00:57]
Connor Rolane
Thanks for having me. Excited to dig in today.
- [00:59]
James
100%. And I respect the Hornets hat. We're a couple days past the draft, so congrats on a great con caniple pick. Good Duke guy.
- [01:06]
Connor Rolane
Yep, yep. My wife's a big Duke fan, so I suspect we'll be at a. At a lot of games this next season.
- [01:12]
James
Dude, amazing. We do have to address Connor Rolane's background. It's a little bit more exotic than usual. Connor, where are you calling in from? I'm.
- [01:19]
Cody
I'm calling in from Medellin, Colombia today. We're. I'm at the. I'm at the hostel right now. You know, the.
- [01:25]
James
You've gone full drop shipping.
- [01:27]
Cody
I've gone full. Full drop ship. Full E. Com, bro, in South America. You know, this is my first time in South America and I got in Wednesday. I've been working since then. And the. It's so, so, so, so built for, like, people that do what we do. I want to travel but have a good work environment. So. Yeah, it's been. It's been fun, dude.
- [01:47]
James
Good. We'll appreciate you making it. Brought, you know, hauled the. The podcast mic down to South America. That's commitment.
- [01:54]
Connor Rolane
Mm.
- [01:56]
James
Cody, you're good?
- [01:57]
Greg Eisenberg
I'm good, man.
- [01:58]
Connor Rolane
Been.
- [01:58]
Greg Eisenberg
I had a few days off. I feel. Feel not refreshed because I was with a bunch of kids and stuff. So happy to be back. Saddle. Missed you guys. Excited to chat.
- [02:07]
James
Yeah. 100. 100%. I'm in the same boat. I've had one of those weeks. I've got like six mugs and three water cups, like spread across my desk. I'm like, I don't know what day of the week it is. Well, anyway, all right, cool. Today we're, we're going to talk how ecom marketers can best leverage AI. That's why we had, that's why we brought James. We begin. I want to thank our sponsors, Motion Rich panel after Self Presion and House. And remind everyone, please like and subscribe when you can leave a review. We've been reading them. I want to talk quickly about Motion's AI creative strategists. These AI agents are built by best in class DTC marketers including Barry Hot, Jess Bachman, Marella Crespi, Alex Cooper and many others. And what's unique is that you get to use agents to analyze your creative using real data from your meta ad account. Connor, have you played around with these at all?
- [03:28]
Cody
Oh yeah, absolutely. Our team's deep in there.
- [03:31]
James
I wanted to pull up too. So for those listening, we'll, we'll talk through it quickly. One, I think the implementation beautiful. You could see our top performing ad creative. Last week what I pulled up was Jess's critique this ads messaging played that progress ran took maybe two minutes and then this is what we got. And I thought it was great. The good fast paced messaging, clean product demonstration, smart focus on MagSafe, what we can push. And I thought this was, this was funny. You know me as a marketer. I'm very precious about our ad creator. But he's like Apple's must have accessory is presumptuous and lacks credibility. The smartest wallet you'll ever own is empty. Superlative marketing speak. I'm like, okay, Jess. He definitely, he's definitely roasting the ad. And what I thought was really cool were some of the next steps. He says reframe around the actual problem. This solves the daily friction of wallet phone management. So the AI has generated what this ad is about and actually reframed what the problem could be. And I thought that was really powerful. 2. I ran Motion's Trek this month's winning themes and you see exactly what we're talking about all the time. Durability that lasts a lifetime. Ditching the bulk for minimalist design. Lifetime offers create urgency and cultural partnerships that expand appeal. And I'm like, yeah, look, you basically get what we're trying to do as an advertising business. I think this is really helpful for creative Strategy and giving the team actionable insights that we can work on every day.
- [04:42]
Cody
Yeah, this is saving our team tons of time. Motion's also giving teams an AI adoption cheat code. It's not like they're just throwing this new technology at you and saying, go figure it out. And they are really giving teams what they need to utilize this and get the most value out of this. And last but not least, you do not need a Motion contract to use this. You can try these agents for free, test them out, see how they fit into your workflows, and start getting value out of them without even needing a Motion contract. So if you want to try out some of these AI agents, go try them@motion app.com.
- [05:15]
James
Actually, before we jump into the AI stuff, James, I did want to get. I've been following the Boring Marketer for a long time. You weren't always AI focused. Like, how did this. Were you pseudonymous at one point? How did start. Like, quick background would be great for everybody.
- [05:28]
Connor Rolane
Yeah, I'll give you the full kind of background story. So let's see. I've been, you know, building like startups and, and doing marketing work since around 2011. So left my job back then, started building apps and all that, got involved in, in kind of like the venture capital world. And I was, you know, working on an investment team and everything. And I, I was analyzing like large amounts of companies going deep into the weeds, understanding how they were trying to grow, what challenges they were facing. And, you know, I wasn't really like a spreadsheet guy. So I, I left that job in 2012 and I started doing growth consulting. So since 2012 to around 2023, I was basically a performance marketer. So I was running a lot of ads, you know, across the whole stack. Creative CRO, analytics and data analysis, all that good stuff, and got a little tired of doing that. And I have a friend, his name is Greg Eisenberg, and I don't know if you follow him on X or whatnot, but we, we linked up and we'd known each other for a long time, and he was like, hey, let's start something together. I was like, let's do it. And he kind of had an idea around starting an anonymous X account, which we did, called Boring Marketer. And, you know, it was me behind the scenes the whole time. And the first thing that we started talking about was like, SEO and organic growth and how to get traffic and things like that. And it wasn't super AI focused at the beginning. And I just started like, you know, seeing a lot of different tools and use cases pop up in the marketing world and sort of started to pivot our agency into really leveraging, like, AI tools for our clients. And up until about, I don't know, five or six months ago, we had like a big team. We were like 40 people, totally remote, like, all over the globe. And, you know, I realized that, you know, we could actually automate a lot of the manual work we were doing. So we started to dig into building these, like, workflows and things like that. If you open up X, you probably see all these, like, workflows that people are sharing. So we started to get into that and we realized, you know, hey, we can really kind of like, shrink our headcount, be more focused, automate a lot of the manual stuff. And that was sort of how, like, the term vibe marketing was born. I was in a group chat with Greg and a couple of our partners, and he was like, hey, you should write a thread called vibe marketing. Like, vibe coding was popping off. We realized we could, like, prompt the system or enter an input, and then AI would do a lot of the heavy lifting and give us something really valuable out of it. Kind of reminded us of vibe coding. If you log into, like a lovable or a bolt, I mean, you can build a landing page or an app. And we realized you can do the same thing for marketing assets. So that was kind of how it was born. And now I spend most of my time, you know, experimenting with AI tools around, like, marketing use cases, sharing what I know, and a bunch of other things that we can dig into.
- [08:29]
James
Yeah, 100, 100%. It's funny, I didn't realize Greg played somewhat of a role in the start of Boring Marketer. I knew it's like somewhat SEO focused at the beginning. And what's ironic about that is greg is crushing YouTube SEO. Like, you can't. You can't learn about AI without watching a Greg Eisenberg video right now. He's just constantly recommended. That's how. That's actually how I kind of got down the rabbit hole. I watched you guys interview from a couple weeks ago.
- [08:54]
Connor Rolane
Yeah, Yeah. I mean, if you search like any AI tool, Manus or Lindy or Gumloop, Greg's probably popping up with, with YouTube. So that gives us some good sort of like LLM SEO and some other modern SEO tactics to talk about. But yeah, it's really interesting to see the rise of, like, video and UGC and all of that in the search results.
- [09:16]
James
Awesome. So before we get into some More like tactical stuff about how E Com brands and marketers can be adopting workflows. I'd love to maybe just get a quick lay of the land of some common terms. There was a discussion recently, agents versus workflow. I'd love for you to touch on that. And Cody is one question was I want to learn about mcp, so maybe we could jump in on that too.
- [09:35]
Connor Rolane
Awesome. Awesome. Yeah. So the way that I think about agents versus workflows, you see a lot of workflows and people actually call them agents. You know, really like a workflow in the most basic sense is kind of a series of like if then statements chained together. So let's say for example, like you, you have an E commerce store and you want to understand top selling products or something like that. Get a report, you know, you're accessing like something from Shopify, you're triggering like a slack message to be sent to you or something like that that requires you to set up the steps, provide the inputs. And it's not like autonomous. It's not going into Shopify looking at data, thinking on its own and then providing you information. Right. It's a series of steps that you've defined this workflow to accomplish. So when I think about agents, I think about like an LLM or an AI model that has access to a bunch of tools already and you're just entering the situation with like a prompt or a broad idea or maybe even nothing. And it's just going out on its own to find opportunities to provide you insights and maybe create marketing assets for you or something like that. Now when it comes to the marketing space, we're not quite there yet with autonomous agents. So the first iteration of like what I call vibe marketing is somewhat workflow based. So when I think about the evolution of this space, you know, you have kind of like level one and level one is when you know, marketers are using ChatGPT, they might be using Claude, they're using a couple other tools, mostly like in silos, you know, like, hey, I'm going to, I'm going to go to Claude and come up with like, you know, copy for my next email or whatever. Now the next stage is when you start to look at what you're doing and figure out how to chain these things together with workflow automation. So a couple of the common tools there are like gumloop, Lindy in 8 in, there's a new one called string.com that just came out. My friend Jordan turned me on to it. You can prompt to build a Workflow so you don't have to think about setting up all these complicated nodes. It'll pre populate them. So yeah, that's kind of stage two. Stage three is starting to unlock mcps, which gets into, into Cody's topic here and just a brief like high level overview of, of what an MCP is. You can think of an MCP as kind of like a two way radio between your LLM and a third party tool software or API. So for example, you have Google Analytics sitting over here, you have Claude sitting over here. Like the MCP allows you to essentially chat with that data in a way that the LLM can understand and format and you can just go back and forth with natural language. So yeah, and then beyond.
- [12:30]
Cody
Is that what Shopify. Is that what Shopify's new Sidekick or whatever is that? Exactly. I mean that sounds like that is what it is. Is that an MCP that they built?
- [12:39]
Connor Rolane
I haven't specifically really looked at Shopify's mcp. I think they do have one. But yeah, I mean most like software companies are creating MCPs. They're not that difficult to create. It's like you have an API, you define sort of like what tools you want the LLM to be able to access to, you know, give you data, give you information and what actions you can take from that. And yeah, you see the McSPlay MCP space like growing super, super fast right now. So yeah, MCPs are super interesting. Like it's one of my like favorite places for marketers to get like leverage and all that. And we can get into that when we talk more tactics.
- [13:21]
James
Amazing. Okay, cool. So I, I do want to get into like e commerce workflows that we think people should be doing today. But I'd love to maybe hear from Cody. I know this is. You tweeted it this morning. The year of, what'd you just call it? The year of efficiency operation. Yeah, the year of efficiency at Jones Road. Beauty AI is playing a role in it. You also seem to be between the three of us that are typically on the pod, maybe on the, the more cutting edge end of new tools. So like where are you guys at at Jones Road currently? What are you excited about in terms of implementing AI?
- [13:53]
Greg Eisenberg
So we did this thing. I, I had this, this tweet which was like has any like the motion ad that we talked about which was like, you know, has anyone done a topi? So I did that. I sent an email to my team. Those were probably a week and a half ago. May go long ass email. Like this is a future we're going all in AI. Like we're still a physical product company. Like we're not, we don't think we're a tech company but like in terms of being more efficient it's like I think the same way everybody has a phone now where everybody is using email. Like we're going to look back on this time as like using, not using AI is the same as like sending physical mail instead of using email right now or something like that. So just like everybody needs to know how to use these tools and use them daily and, and just starting with just chat, GBT and stuff like that. So just like setting our expectations, setting how we're looking at it. Definitely mentioning the concern of job loss and things like that and explaining our. That's not our goal with it but our goal is definitely to, to grow headcount slower as we grow, you know, our revenue and then obviously following up with training resources. So I led an hour all hands just showing like what I wanted to do. I'm sure everybody here has had that like red pill moment where you're like mine is blown. And we've probably had that like a hundred times. I just wanted to make sure everybo and my team had that. So I just shared a bunch of stuff that I thought was cool. I obviously just did a bunch of chat GPT stuff, showed all the connectors, shared deep research, little basic prompt engineering stuff. I shared a bunch of stuff I vibe coded so just did a bunch of stuff in Lovable or Claude some of that was Jones road related. Built some calculators and stuff like that. Some was like some stuff I've just done for fun on the side. Like I built like a workout app and like a training program that I've been using. I just, I just think I'm so curious about this stuff and it's such a fun time for, for people that don't actually know how to code. Shared some of the other tools that are like vertical AI tools like Nectar Social or Listen or stuff like that and just got people excited. And then we're going to bring in two to three speakers. We actually have our first one today with an outsider, Alex Cooper, if you guys know him from Twitter. Him and Jimmy Slagle are coming in. They're going to do two sessions for our entire team. So that's actually happening right after this and then I'm looking to bring in a few other people and then we have a $7,500 bounty for the end of the year just to One person, whoever makes the biggest impact. And dude, our, our procurement manager built five coded three calculators. She built a tariff calculator. She built some kind of like duty and drawback thing to understand like should we air or split shipments. So, and it's like in a week she did it and it already has saved us a ton, a ton of time, time and money. So I'm excited about it. But I think we've got like the whole range of people that are like, how do I even use ChatGPT to like vibe coding and deploying apps in Vercell?
- [16:29]
Connor Rolane
It's, it's, it's really interesting. Like you know, one of the ways that I think about it is like, I don't know, vibe coding something unlocks so much like individual efficiency. Like if you have a 50 person team and everyone knows how to like automate workflows or vibe code a solution and they can save 10 hours a week doing something that they were otherwise doing manually, that's what, 500 hours a week, which, I mean, what does that look like over the course of the year? That's like 25,000 hours that are like freed up for more like strategic thinking or other initiatives or you know, more experimentation. And I think that the way that, that you're approaching it is right, like empowering individuals within the organization. Like no matter what your role is, you can use these tools to do something like increase efficiency, get more leverage, you know, produce more valuable outputs. And something that I notice is it really closes like the knowledge gap, you know, like your procurement manager, vibe coding something about tariffs. Well, like there's no secrets around like what tariffs are, how, how much they could impact you. You could go and do deep research and find that out, take that information, throw it into cloud code and build a calculator like that. So like just the democratization of like information and the ability to act on it is just like insane right now.
- [17:52]
James
I also think from what I hear from Cody is like the proliferation of level one the way that you described it, James. Like just make sure everybody's got Claude and Chat, GPT knows the basics, can use it individually. I think that's probably where most brands are today. Like hey, let's just make especially for like top performing operators at companies. It's like they are the ones on top of it and they're like familiar with the trends. I am curious about that level two that you mentioned, like stringing together workflows for a brand that hasn't gotten to that point. Do you have any recommendations or tips on getting started and, and where you're seeing, you know, the most success.
- [18:30]
Connor Rolane
Yeah, getting started, like it's a similar process to level one, right. It's kind of about individual empowerment, finding like maybe an internal champion that wants to like, you know, start digging into some of these tools and then maybe sharing what they're learning across the org. And I think if everyone's sort of embracing some of these workflow automations, the automation tools, they'll be able to find something that they can like string together and automate. The tools that, that I recommend starting with are like, I think Gumloop is a great one. It has an easy to use interface. You can understand what's going on. There's other tools that are a bit more advanced. You see a lot of innate in stuff out there on X. But I find Gumloop as a very like beginner friendly tool to start using. So like really for me like it totally depends on like what your organization's doing in terms of like what the most valuable workflow is. But I think auditing, like what people are doing manually, having people sort of like, you know, share what their day to day looks like or maybe what their week looks like and sort of have a brain trust, it's like, hey, maybe we could use AI to do that. I've seen Cody talk a little bit about like getting reports in Slack or you know, something along those lines. So I mean you could easily, let's say connect Gumloop to Google Analytics, format that in a table and then have AI summarize the top, you know, 10 insights on a weekly basis and then throw it in a Slack channel. You know, that would be an easy example to get started with. And once you, once you kind of like build your first automation, it really opens your mind and then you can like start to see all kinds of opportunities around you. Like, why am I doing this manually? Like there's no reason to be, you know, going from this tool to this tool to this tool. So to get started I, I think it's about kind of looking at that, kind of auditing yourself and, and understanding like what you're doing manually. Sometimes I'll use like Chat GPT and I'll be like, hey, here's my current process. Or I'll record a loom video of like how I accomplish something today and I'll share that video with an AI model and I'll be like, how could we automate, you know, 90% of this using Gumloop? You know, if you share that into Claude Claude can plan out exactly how you can build a node based workflow and take you like 90% of the way there. Then you just go and you follow those steps and if you have an issue or you run into a problem, it's a prompt away. You go back to claw and you're saying, hey, I'm stuck here. How do I get past it? And it's kind of like just going through that process to figure out where those blockers are and how you can like hop over them. And after you do this a couple times, you get way more comfortable with it and you start to understand the logic and how these tools work. I don't know if any of you have ever tried to learn how to like code before, but it's a similar thing. Like it, it feels like totally foreign at first. It feels like you're never going to kind of climb up that initial hill to get your like first app out or whatever. But once you do, it's, it's like kind of a snowball effect.
- [21:30]
James
Totally one. Oh, go ahead Cody.
- [21:33]
Greg Eisenberg
I was just gonna say yeah, I used to use Zapier like a lot like past role and past life and stuff. And like, yeah, it, it takes a ton of time to learn and then you get like good at it. And I'm sure they're building stuff with a, like they all seem pretty similar but yeah, I don't know if you guys are the same but I'll just see stuff in Slack and I'm just like, I know that can be automated. Like I know that would save this person so much time. I just don't know how to do it or have the time to do it. But like it would be so cool if it's like, if there's like a follow up reminder from a project manager person like hey, where's this? Do you have to do like I am sure that can be automated. I just don't know how.
- [22:03]
Connor Rolane
Yeah, yeah, I mean the way that I kind of dig in is like I lean on AI like very, very heavily. You know, like if, if I don't know if something can be built or something can be automated. Like I' right to Claude and I'm starting a research process. I think that you know a lot of people when they're trying to identify like valuable workflows. Like the place I recommend most like marketers to start is kind of on like the research and the analysis phase. Like if you arm an LLM with the right tools, the insights you can get are pretty crazy. Like you know, to go and manually look at like your top 20 competitors and everything that they're doing on a weekly basis is nearly impossible for like a human to do. But LLMs give you the power to go out and process so much data and information and get insights from it, you know. So this, this kind of leads to MCPS. 2 MCPS I would recommend everybody to look into and you know, you can sort of go to like Claude or Chat GPT and ask hey, how can I set these up? Get Claude desktop. Okay, ask Claude. Hey how can I set up the Perplexity and the Fire Crawl mcps? You'd be amazed as to what you can accomplish with those when it comes to like competitor analysis and things of that nature. And we'll talk more about it when we get into the tactical stuff, but that's a good place to start.
- [23:29]
James
Yeah, I love the meta of, you know, how do I use AI? Well, you ask AI if you're listening to this podcast, I know that you're working hard to drive revenue with upsells and across your E comm site generally. But if you're not monetizing your post purchase experience, you're leaving easy profit on the table. Most brands and everyone listening are optimizing the transaction journey, but stopping at the thank you page and that is a big mistake. After Sell is changing that. We've been using their Rock thanks feature at Ridge and it's a no brainer for us. Here's how it works. You show premium non competing offers from brands like Disney plus and HelloFresh on your order confirmation page. AI driven personalization picks the best offer from hundreds of reputable advertisers. And brands like True Classic, Blendjet and Jones Road Beauty are already using it, seeing 30 to 70 cents lift per order. No disruption, no fees, just extra cash in the bank. And the numbers speak for themselves. After sale customers generate 20 to $30,000 in pure profit per 100,000 orders. There's a 16% average engagement rate which is way better than traditional ads. And it's a new revenue stream without disrupting the checkout. It's an easy win and it's completely free to try. Plus After Sell is hooking up marketing operators listeners because you can sign up today and you get their entire upselling suite for a year which we also use very heavily. So visit aftercell.com operators to try rock thanks and see the money rolling in. Stop leaving money on the table. Turn your thank you page into more revenue with After Sell. One thing you said quickly that that I we've been talking about at Ridge and I've been advocating for, for a long time is like an internal champion, the level one and I really like the framing that you had level one table stakes. Everybody should be using ChatGPT in the same way everybody uses like Google Search. It's like super approachable to be extremely valuable without really knowing all that much. I find it hard to believe there's actually no way that like everybody gets good at stringing together automations. And that's why I'm like, oh, if it's an internal champion or. Our director of remote has taken along a lot of responsibility in terms of like identifying what workflows can be automated via AI. And we got super lucky. We hired this guy in the Philippines. I hired him at the end of last year before kind of the AI automation craze took over and he was really into building the second brain, like that whole thing like just like non AI automations. I'm like, oh dude, this guy's just like a nerd for automation and process. He's going to be super valuable. So they end up building out a lot of our flows and that becomes extremely valuable because I, I almost liken it to like Excel. Right. Most people at the company are not good at Excel. They'd be way better at their jobs if they did know how to export data, format it, run pivot, do the whole thing. But it's just unreasonable to think that you're going to get 100% adoption. So you get the like 10% of the workforce that are extremely good at Excel or AI automation and then they're going to empower everybody else at the organization.
- [26:24]
Connor Rolane
Yeah, it's a, it's a really good point. And you know that, that those sort of levels have been kind of like my progression. But I've, I've watched other progressions as well. And you know, something that I've kind of come to the realization of lately is that most workflows can be vibe coded. You know, like I'm not actually the best at building a step by step like workflow automation, but I'm really good at vibe coding. So like when I want to automate a process, I'm much better at just building that into a micro app or a micro tool or something that I can use internally. So like I did one for like creative analysis for meta ads and I was trying to think of like how many different steps in this workflow this workflow would require and it was like overwhelming. And then I would have to go into these little, you know, nodes in the workflow and figure out how to like format the data and all that. And it does feel like, you know, trying to go really, really deep into something like Excel. It's not very fun. So the big unlock for me was starting to work with Claude code. Now, Claude code is not claude as, as we know it. Right. Claude code is a development like agent that Claude built and used internally and you can install it in your terminal. Okay. And what I've realized is that Claude code takes something like, hey, I want to do competitive analysis on, you know, these 10 companies, meta ads or whatever. It will take that overarching goal, break it down into step by step sort of tasks that it can go and accomplish. Now it's not going to be able to just one shot that out of the box. You know, you're not going to just have a tool that is able to go out and scrape and understand ads with one prompt. So you have to go back and forth and work like it. Work with it like you would like an engineer. Hey, I like this. I don't like this. It'll say, hey, you need this like API to go and scrape these ads or whatever. And then it's like, show me how I get that. Where do I find the API key? How do I provide that to you? And what I've realized is that a, it's a much more fun and enjoyable process to like Vibe code these workflows and these processes versus building these like somewhat static automations. You know, some people may like to go and, and do the automation, like, workflow stuff. I really love working with, with Vibe coding and it's much more like, I think it's more powerful than building like workflows because really, like, over time you're becoming like an AI engineer. And I think that once you get to that phase, which is kind of like, you know, level three, I like to think of is where the real magic starts to happen. So, yeah, some people might prefer Vibe coding, some people might like, you know, building workflows. Both can add efficiency. I think this is more powerful and that's where I like to spend my time.
- [29:24]
Cody
Yeah.
- [29:24]
James
100. So, Connor, I'm curious to get you in the mix here. In terms of the different levels, where would you say hexclad's at? And are there any things that you guys are investing in right now?
- [29:35]
Cody
Yeah, we're. I mean, I'd say we're very level one. We just had an AI. We just had an AI expert come in and kind of do some basic training on how to use these tools. So, like, right now we're using it to create like, like many brands within our orgs. Like, for example, brief writing is something that, you know, we have varying levels of, I'd say, competency at and like, not also like some templatization in certain parts of our org and not another. So that's one of the things that we are trying to do from the jump, is like build a GPT that we can input all the information it can spit out a standardized brief. So I'd say that's like our. We're at level one. I think that's kind of like us tiptoeing into level two and trying, trying to create some of these custom tools that we are five coding and like figuring out like, where do we need more standardization across the org and how can we create a custom GPT to provide that? So we're, we're pretty early on. I, I really like the idea of having like an AI champion. That's something we'd like. We have certain people in the org that are like, doing it on their own. You know, like, we're using AI voiceover tools and like, you know, chat GPT and cloud to write briefs already within our creative strategy. But it's not uniform across the org. So I think it'd be cool to have someone that can kind of like almost parachute into different parts of the org and like, figure out what they're doing and be able to like, implement stuff like this. But yeah, I mean, we're definitely, we're definitely like in that early phase of like, trying to get out of love, not get out of level one, but like, keep getting better at level one and like getting into level two.
- [31:06]
Connor Rolane
Totally.
- [31:07]
James
I, I've been complaining about this a bit because, like, ChatGPT is like, really good at data analysis. So we're like, we've gone to the point of like, encouraging people to do data analysis with chat GPT, but what that requires right now is like, everybody's logging into the Shopify admin separately, downloading different reports, getting big CSVs, uploading this chat GPT and doing analysis from there. And when you deliver the analysis, like, okay, awesome, you might have some nice charts, it might be accurate, whatever. But I'm like, like the, the margin of error there is massive. I'm like, what we need to be doing is level two, where it's like, how are we actually stringing this together so we're all working off the same sort of resources and workflow. And I think we're probably just a couple months away from like, turning the corner there.
- [31:49]
Greg Eisenberg
All right, jmt, what would you do in that scenario? Because, like, we're using Shopify Sidekick inside. But, like, if we wanted to get all of our Shopify data or all of our data, you know, into an LLM, like, what's the best way to do that?
- [32:03]
Connor Rolane
The best way to get all your Shopify data into an LLM? So yesterday I actually, like, Vibe coded a Shopify site. I was trying to learn about, like, you know, how to work with the platform and things of that nature. And I built a really nice site. Like, it looks great, like, fully with Claude code. Just kind of talk about my process and then I'll try to answer, answer your question as I go. So at first, like. So I've never built a Shopify site, like for, for instance, right. Like, have very little experience in the platform. When I've tried to go in there, in there before, I'm like, this is, this is a lot. So I, I kind of like, just shelved that. But what I realized once I sort of unlocked like, Claude code and, and cursor and stuff like that is, you know, a Shopify theme or a Shopify website is, is just code like any other site, right? So I tried to start with like a free theme from, from Shopify. I downloaded the theme, I loaded it into my Vibe coding tools, and I started to work on it. And what I realized was there's a lot of stuff in there that prevented me from getting it, like, on brand, so to speak. It was kind of hard to make it look good. It wasn't really functioning right. So I, I said, hey, Claude code, let's just write our own theme from scratch. Like, knowing what you can go find out about, you know, the, the Shopify documentation and the files that I've already uploaded. And it built a theme from scratch. I connected GitHub to Shopify. So within my vibe coding, you know, IDE or development environment, I say, hey, push it to GitHub. It automatically updates my Shopify site. So, and, and like, I can go there and look at it. I can show it to you. It looks really, really good. Today, before we hopped on here, I was thinking, like, about some of the use cases that we might talk about. And I was like, you know, I assume that a lot of brands spend a lot of money with like, agencies to build, like, landing pages and, you know, all these types of, like, assets or whatever. Can I build like a really nice landing page with Claude code and push it live to shopify that was kind of like my problem statement. So within Cursor, which is what I use, Cursor is kind of like the next level vibe coding tool beyond, like lovable and bolt and replit and all that. Okay, I went into Cursor, and within Cursor, I can chat with some MCPs that I have set up. So I used Perplexity. I said, hey, use the Perplexity MCP to look at Hexclad, Jones, Road Ridge, AG1 and Grunds. And I want you to tell me exactly how I could build a long form direct response landing page. All right, so it went out, it did a bunch of research. It took a couple minutes. It came back with like a super detailed roadmap on how I could build a page. I copy and pasted that into cloud code. And I said, hey, build this page. Use. Use the brand system that we already developed for the main site and, you know, build it and deploy it. So I did that. I went back and forth a little bit. It took about 30 minutes to build a really nice landing page that was like totally on brand. So, like, I think that's pretty crazy in the sense that, you know, you can access your Shopify site. You can create like a. Like a clone of your site, not like the production site. You can create a branch on, on Shopify that's maybe a preview or something like that. And you can get in and you can change copy, you can update CTAs, you can deploy landing pages and stuff like that. Seems to be a really good way to go like 0 to 1 and maybe share that final output with, say, your agency who has full context on the production website and stuff like that, you know, so the way that I accessed Shopify was through GitHub, but you can also install a Shopify CLI, so that lives within your terminal and that allows you to deploy product pages, access product information, and really get the full range of data and information from your Shopify site. So if you have Claude code or something like that, you can install the Shopify CLI and then it'll ask for your credentials and you go and log in and then, you know, it's a matter of, hey, Claude code. Like, what. What's everything you can access from my site? Like, tell me what my products are like, all that stuff. Like, list out all the descriptions for every one of my projects, give me some ideas for how we could optimize, you know, product descriptions or something like that. And it has full access to your code base and it can. It can do that, which is pretty crazy. So, yeah, you can access it via the cli, which lives in your terminal. You can access it programmatically via Shopify's API and pull that data into your Vibe coding tool of choice and work with it. So there's many ways to access Shopify. I think those were three that I just mentioned right there.
- [37:14]
James
Yeah. So awesome. I've got one. I love hitting some of the tactical stuff here, so. Well, one. I don't know if this is what you're laughing at, Cody.
- [37:22]
Connor Rolane
A little too in the weeds or technical.
- [37:25]
Cody
This is.
- [37:25]
Greg Eisenberg
So I just, like, don't know what half of those terms mean.
- [37:29]
Cody
Like, yeah, this is such a pain point for us though. Like, like, development resources is. Is really hard. Like, we really struggle with balancing all the different development projects we have. So I'm like, I'm like drooling over here because I'm like, if we could develop all of our paid media landing pages just using Vibe coding, like, that would be like the. I. I can't state enough how big of like an unlock that would be for our organization.
- [37:55]
Connor Rolane
Just the volume that you could test and the speed at which you could iterate is like, unreal, you know, So I, I think that it's a huge opportunity. It's similar. Like right now I. I think brands are doing that for like, creative testing. You know, they might be generating like a bunch of image static ads with. With GPT or maybe even some like, videos with. With VO3 or something like that. And, you know, you can get some real data on like, what hooks, angles, formats, or whatever might perform. And then you could go and take that, you know, into your normal, like, creative workflow or something like that. Same thing with like, landing pages and other assets. You know, if you can go and test something and get some data on it and go from zero to one, and then you can validate it and say, hey, this is what works. Like, go and build this. You're saving a lot of time doing that testing process with an agency that's probably charging, you know, 200 bucks an hour or something like that, you know?
- [38:50]
James
Totally.
- [38:53]
Cody
I keep hearing the same things from people. Lately. I'm being asked to do more with less. I have big goals to hit, but my budget is tight and every marketing dollar needs to work harder. I know I'm hearing this at hexclad. I'm sure a lot of operators are hearing this at their brands. Super relevant right now. Incrementality testing is the best way to figure out what's actually moving the needle. So you can move around your ad spend in a way that's backed by science. All three of us use House for incrementality testing. We all love it. And House is now working with more than 40 of the top 100 DTC brands, which is pretty insane. I can speak personally for Hexclad. The amount of insights that we've gotten, especially on our view based channels and how those are driving impacts and efficiency and revenue for our business is not only super valuable, but only possible through House. You know, channels like Connected TV, YouTube, TikTok, these very view based channels that don't garner a click the same way that Facebook and Google does have been really only measurable through geo incrementality, holdout testing and the actionable next steps from these tests in terms of deciding to scale up or down or keep spending where it's at is is really amazing. House helps you run experiments on your channel so you can confidently answer the questions you've always wondered about. Things like what channels are actually driving my business? How much should I be spending on each channel? What's the impact of my ads on Amazon or retail sales? How should I structure my Meta and Google accounts to make sure they're spending dollars in the right way? The beauty of House is that it's built for marketers. The science under the hood is rigorous, but the platform itself is simple. So you pick a question, launch a test in minutes and get real results fast. Every customer is paired with a measurement strategist, someone who understands growth and brings a clear strategic point of view. They will get to know your business, help you build a roadmap of impactful tests and guide you as you operationalize incrementality in your day to day. Even if you're brand new to testing, you're not doing it alone. It is very much no in no way like a go sign up for the platform and figure it out on your own. You have a partner in these customer success reps from Houzz who's been in your shoes. And now more than ever, you need to make every marketing dollar count up. Level your measurement with House by going to house IO forward/formators, that's H A us.IO/operators and start allocating your budget with confidence.
- [41:10]
James
So where we're at with with creative generation is interesting because we're using, we do quite a bit of it, whether it's ChatGPT generating images or VO3. We're doing a ton of the AI voiceover, but it's like, it's just like a relatively thin Layer of like the whole creative process. Do you have any examples of going like upstream? Like, I loved your point around the perplexity research on landing pages or even downstream in terms of like. Oh, because the other point that I'll make is we're just generating like the image and then we're putting that into Photoshop with a designer to overlay headlines and things like that. So can we go. Do you have examples of going upstream? Upstream for research and how you would do that and downstream for like, how does it actually become a piece of ad creative quicker?
- [41:53]
Connor Rolane
Yeah. So I'll just kind of describe to you how I would go about this and I've, I've built some tools that, that I use to, to do it. I, I posted a tweet, I don't know, like a week or two ago where I did this for cuts versus true classic. And my process looks something like this. So, you know, the, the first thing that I think about with creative generation is I, I feel like everyone wants to get to like the sexy output. They want to get that cool static they made with AI. They want to get like the video with VO3 or whatever. And they're kind of basing that off like existing performance data, which is fine. Maybe they have that in their head or they've got a dashboard that says, hey, this was a top performing ad or whatever. But I think what most people are missing are new, like unique angles and opportunities. All right? So I like to focus 80 or 90% of the time on identifying what those look like because creating the actual asset is the easy part once you have the right information. So I do something called gap analysis. And the way that I would approach that for this use case is, you know, let's say I'm, I'm, I'm gruns or whatever. And you know, what I'm going to do is I'm going to identify, maybe I'll use Perplexity and I'll go and identify, you know, 10 competitors to my brand. You probably already know who your competitors are, but let's just say I went and identified those. The next step that I would do is I would probably go and scrape all of their Facebook ads, at least from the last like or meta ads, at least from the last like 90 days or something. Now you can use tools like Apify, which has a suite of scraping tools and they have a Facebook ad scraper in there that, that you can utilize. Now again, I do this with vibe coding. So I'm here within like Claude code and cursor And I'm saying, hey, like, I want to use this scraper from Appify to go out and scrape the ads over the last 90 days for these brands. Like, that's how I do it, just prompting back and forth. So once I get those ads, it's going to come in like a messy format. You got to do some work with your, your coding agent to kind of like, you know, whittle that down to some clear, some clear assets. So in this process with Cuts and True Classic, I got a big dump of ads. And then I was like, wow, a lot of these are videos. Like, how do I understand what's going on with videos? Like, the AI can't just understand those out of the box and watch them. So I transcribed the videos. I used whisper from OpenAI, which is a video transcription tool, to programmatically go through and transcribe all the videos that Cuts and True Classic were putting out. Now I had the transcriptions and I also had snapshots of, like, what those ads looked like. I had the copy, I had the CTAs or whatever. I think I. I scraped, I don't know, three or four hundred ads, you know, something like that. So I've got this baseline information. I can do the same thing for my ads. I can scrape my own ads and I can sort of have three tables here, here's mine, Here are my two competitors. Here's how we're all sort of trying to advertise and address the market. So that's cool. I have that information. But what I really want to do is figure out, like, hey, what are some unexplored, like, angles and hooks that I can go out into the market with? So what I did was I started to use like, MCPS and scraping to go onto, like, Reddit and YouTube and Amazon reviews and all these other data sources where people are buying shirts. And I was like, hey, what I want you to do is go out and find all the pain points that people are talking about that are unaddressed currently with, with these ads, you know, and what I found was it found, like, I don't know, 10 unique angles that no one was talking about in their, in their meta ads. You know, they were all kind of talking about similar things like both of these brands. And I think the reason for that is because, you know, you can go to any brands, like Meta Library and, and look at, like, what they're doing. And I think in this space, a lot of people are just kind of like, remixing the same things and the same concepts and the same angles. So what it, what it found was some angle around like shirts that don't like pill and don't lose their like color when you wash them. And that was like a pain point that people were actually discussing on, on Reddit. So from there once I have that insight, then I can build a brief. You know, right. It's not remixing someone's existing asset. It's a brand new fresh insight that who knows, I can test. Maybe, maybe it'll work, maybe it won't. But you know, once I had the brief, I was able to create a video with VO3 and it gave me like an 8 second video that I thought was pretty good with one shot. And then what I did on my tweet, I like put all this together, like all the analysis that I found, the angles that were unexplored and then like a couple previews of the ads that I was able to create in this complete, you know, workflow. So yeah, I know that's a lot, hopefully that kind of talks about or like gives you an idea of how I would approach this. I, I think that everyone wants to produce the thing but the real values and like knowing what to, what to create.
- [47:11]
Cody
I, I wanna, I'm so that's awesome. I think that's like this amazing workflow that if any brand could implement, that's like a huge value add. How does a brand learn how to do that? Like you know, we talked about like you just mentioned a fairly sophisticated process that had a bunch of other tools plugging into it. Like yeah, what, what would a brand do to actually make that something that becomes muscle memory and a competency of theirs.
- [47:36]
Connor Rolane
Right.
- [47:36]
Cody
Like do they need to hire someone like James to come in and do like a workshop and then have like recordings that they're going to refer to and then over time they get good at it like that just, that'd be hard to be like I'm going to go figure this out on my own. Which you obviously did. But you've clearly spent hundreds of hours playing with these tools. Like what's a practical path for a brand to that workflow or the landing page dev workflow or any like practical marketing workflow where we're trying to actually like use these tools to produce deliverables that are going to perform for us and drive results for our business. Like what is a brand? Where does a brand start to figure that stuff out?
- [48:09]
James
It you do you just ask AI is that the answer?
- [48:14]
Cody
Chat GPT.com by James's course.
- [48:16]
Connor Rolane
Yeah, yeah, yeah. So I'll, I'll kind of talk about, like, my experience over the past 30 days because I think it, it might give people some good tips. So 30 days ago, I, I had a, a guest that I interviewed and he showed me Cursor and it blew my mind. And I was like, I cannot believe you can build these things and do this research and produce these assets all from one platform. You know, Cursor is a vibe coding platform, right? And, you know, it's got a very smart agent, like I said. And it just kind of blew my mind as to what I could do in there. And like, 30 days ago I was like, I'm going to learn how to use this. Like, I think this is the future. This could be the interface where analysis, coding and producing assets all comes together in one place. This is way beyond building a workflow, etc. So I just started with it 30 days ago. And this, I built this workflow after a couple weeks. Now most, most people don't know what to create. The problem isn't like, necessarily how do I do it, it's what should I do? You know, so, like, there's a lot of logic and experience that goes into knowing, like, okay, I'm gonna, I'm gonna try to find unique angles in the market and here's where I'm gonna look and things like that, you know, So I think there are some things that like, experience can't necessarily replace. But I'm sure that all of you guys have more experience than, than me with, you know, figuring out like, angles on, on meta ads and stuff like that. So it's just kind of like, like trying to unpack what you already know, right? And then any of you, if you, if you spend enough time just opening up Cursor and installing Claude code, could probably build the same workflow. It might take a week, it might take two weeks. It might be frustrating. There might be a back, a lot of back and forth. You might need to ask like a developer, hey, what's happening here? You know, something like that. But I, I just think that there's, there's really no like, gatekeeping anymore. Like, anyone can go in and figure this out. And I, I preaching that there's two things that you really need to know, and it's like you need to know the right questions to ask. So like, zeroing in on like, hey, like, would this work with, with this kind of logic, or why is this breaking here? Or whatever. And then what good looks like is the second thing? So what good looks like comes from experience. It comes from research. You need to know, Try, try to try to know like where you're trying to get to. Otherwise you'll just be in this endless like loop of like vibe coding. So I, I recommend everyone like installs cursor and tries to get Claude code in there and does something very simple. So I mean like it's just a matter of trying to invest the time. And I realize like, not everyone can do this in 30 days. Like this is my full time thing. Like I basically experiment with these tools constantly, but over the course of like three months, if you spend an hour a day trying to figure this type of thing out, like after those three months you're gonna, you're gonna be like almost like an AI engineer. And the opportunities that you have in front of you are just like kind of mind blowing. Now it's at the point where I'm like, what can't I build?
- [51:34]
James
You know, I think I love the point around cursor and that this feels like ex. Like there's some alpha here for sure for people. I had a similar. I just had the mind blowing experience. I haven't had the chance to dive in yet with Jacob Pozel showed me cursor for like an hour and a half and I was totally unfamiliar. I hear about cursor all the time because people are doing coding with it. Right. It's like a web development first tool or a software development first tool. What you described earlier is the first great example I've heard of. Like, I don't know if you'd consider it knowledge work being done within cursor, which feels like a big unlock and like maybe really early in the adoption stage.
- [52:12]
Greg Eisenberg
I didn't know you can. I thought it was just a development tool. Yeah, yeah. We're about halfway done through the year. H1 is almost over. We're prepping for the second half of the year. We're also prepping for Q4, which is huge for us. And our budgets are going to be the highest then. And when our budgets are the highest, we're going to need the most granularity and the most confidence knowing where we should spend our dollars. And that's why we turned to prescient. Prescient is an MMM. And most MMMs, they use 60 year old regression models. They're not really built for D to say prescient. You can get readouts instantly. You can get readouts really quickly if something is changing, which again, T2C is really volatile, so you can't rely on an outdated mmm. That you're only gonna get one read out a quarter and you're able to see one thing I love about pression so much. You're able to see the halo effect. So upper funnel spend, which is again is huge. We love to fill the funnel prior to a peak moment prior to Q4 holiday, something like that. We're really able to see the halo effect that has. Because sometimes you're not gonna see great attribution from YouTube campaign or from TV campaigns. Well, prescient plugs into all them and it can actually tell you your base plus your halo. So it's been really helpful for us to understand and actually have confidence to invest in some of these upper funnel channels that are harder to measure with other ways. So I love Prescient. I can't tell you all the technical stuff behind it, but I can tell you it gives me and my team more confidence know where we should put our budget, especially in some of these pesky upper funnel channels that are much harder to measure. And again, don't wait till Q4. It's going to be too late then. We're halfway through the year. But we've really got to all lock in for the rest of the year for Q4. There's a reason we use it. Hexclad, holo Talks, Coterie and 100 more leading brands so highly recommend checking it out. Go to prescientai.com operators to book a demo today.
- [53:54]
Connor Rolane
So I'll just give you like I'll share my screen in a moment and then what I'm going to do is just kind of show you kind of a knowledge work, use case and kind of how you can use cursor outside of coding and what the power of it really is while you're doing that.
- [54:11]
Greg Eisenberg
I've been playing with this string thing that you mentioned. Dude, it's awesome. It's great. I'm definitely going to use this. It's prompt to build zapier agents and their workflows and stuff.
- [54:21]
Connor Rolane
Exactly. Yep. It's a lot easier than trying to figure out how to go through node by node and stuff like that.
- [54:30]
Greg Eisenberg
I have a question, but before you get into it, something like this, I know that AI, there's a lot of costs and variable. Obviously it's cheaper than a person doing it. What, how all these different tools, like what's a pricing model? Like is there anything like I need to be aware of or my team needs to be aware of? Because I see people post screenshots that are like coding with Claude and it's like a thousand bucks a day or something like that. Is there anything that people need to be aware of?
- [54:51]
Connor Rolane
A lot of these like screenshots that you see are like humble brags of people like trying to flex how much they're spending like or how much engineering they're replacing with AI. So like you, a lot of these folks have like five or six different instances of like Claude Code, like running at any given time, like doing a bunch of tasks. It's all, it's all token based within Claude code. I mean you can monitor your cost. It's a slash command, slash cost and it'll give you a full breakdown like daily report on what you're spending or whatever. I definitely spend a lot of time that on schools, just kind of like experimenting flex on them.
- [55:34]
James
Flex on them. What's your, what's your Cursor bill?
- [55:37]
Connor Rolane
Cursor? I think my, my highest bill was like I don't know, 600 in a monthly period or something like that. Claude code. I think it's actually a lot cheaper than Cursor's Agent. So been saving some money there. But like you know, I, I probably have, I don't know, $3,000 a month in different like AI tool subscriptions and what I'm spending and tokens and stuff like that. But for me like it's like a, it's almost like a marketing investment because I show off what I'm building and things like that, you know. So you know, this is, this is kind of like a blank workspace here and I'm just going to show you Cursor Agent. Okay. This is Cursor Agent and you know you can select various models here. Now I'm just going to have like Cloud 4 Sonnet selected and one of my favorite kind of knowledge work use cases is actually keyword research. It's a pain, it takes a long time. A lot of the out of the box tools are very expensive and if you open up those SaaS tools, it's just like what's happening in here. How do I even get to like, you know, 20 keywords I want to focus on for my like organic traffic or something like that. So let's see what, what company do we want to use as a use case here? I'll let you guys pick.
- [57:02]
Cody
Definitely hexcloud.
- [57:04]
Connor Rolane
Okay.
- [57:04]
James
They're big and successful.
- [57:06]
Connor Rolane
Okay. So here's how I would do some knowledge work. So I am the head of growth at hexclad. Step one, I want you to fully understand my brand. Use the Fire Crawl mcp to scrape the site and then produce a summary of the company in a MD file. So I'm just going to enter this prompt and Fire Crawl is an MCP that I recommended earlier. It's one of the ones that I use all the time. So this allows you to go and scrape pretty much any website that you can think of. So right now it's calling Fire Crawl via this prompt and Firecrawl is working in the background. I hope to go and scrape Hexclad. So sometimes it takes a little bit of time, but we'll wait and see what happens.
- [58:08]
James
I've got a quick question here for you, James. The Firecrawl MCP is like. Is that almost the equivalent of installing an app? That's something you put in cursor someplace else so that you could reference it?
- [58:18]
Connor Rolane
Yeah. And you want to know the secret to setting up an MCP easily?
- [58:22]
James
Yes.
- [58:24]
Connor Rolane
That is Ask. Ask AI. So like, if you're, if you're working in cursor, you can just go and ask Cursor Agent, hey, I want to set up the Fire Crawl mcp. What do I do? Cursor Agent will set up a, basically a file called MCP JSON. And that's just a fancy way of saying it'll set up a file that's stored on your local environment where you go and enter an API key. So you'll go to Firecrawl, you'll go get your API key from the main dashboard. It's like in the upper right hand corner say, hey, Cursor Agent, share the document with me where I can enter this and tell me exactly where to enter this information. Or here's my API key. Go and do it for me. That's it. It'll update that file. You'll have the MCP installed. You go to Cursor Settings, you should see the Firecrawl MCP and Cursor settings. You can switch it on and then once it's on, you can reference it with natural language here with Cursor Agent. So it's a very quick process to get an MCP set up. All right, so it's looking at a bunch of stuff here and it's creating a brand analysis for hexclad. So, you know, success with a lot of these models is about giving it the right context. So typ, typically I want to give it the right context on like, what I do and what my brand is, just to get started. And then we'll build off of that sort of like systematically and, you know, it's brand knowledge. I can, I can reference this document and things like that in, in the flow. So I'll show you what that, what that looks like. All right, so it's, it's telling us the key findings and stuff. But we're going to open up this document that I was able to create here and you know, this is a document that I can now share. I can referen for the agent. This is context building, right? Like I'm not in Claude having to go search the Internet and then re remind it like you know, what the context is and what we're working on and stuff like that. I have this file in here that I can at mention to my agent. I can say hey, make sure you reference this when you're, when you're approaching the next thing. So let's just go through this. So position sells as a premium hybrid cookware brand. Strong direct to consumer focus. Leveraging celebrity chef partnerships and patented technology. It's got the core product lines, technology differentiation, pricing strategy, growth opportunities even we didn't ask it for that. Celebrity partnership strategy. All kinds of stuff, right? Warranties, recipe content, hub education, what kind of educational content you have, what your competitive advantages are, brand challenges, growth recommendations, like all kinds of stuff. Right? Right. KPIs to monitor, etc. So this is just from one MCP call with Fire Crawl to get an understanding of the brand and we're having Claude Force on it, sort of summarize this. That's it, you saw the prompt. So the next step is okay, knowing what you know about my brand, use the Perplexity MCP to identify my top three competitors.
- [61:42]
James
I've got, I've got another question. While we let the, the Perplexity MCP run, do you have a perspective on the future of AI? Like we see a lot of purpose built tools coming out, right. Like you look at like icon for ad creative generation or Triple Whale has a really cool demo out now. Like if you were to speculate on what the future looks like, is it more bespoke tooling done with cursor and mcps or are we seeing brands adopt like really well built, you know, purpose driven solutions?
- [62:14]
Connor Rolane
There are some things that you can't just like vibe code in house. Like you know, some of these kind of like, like Triple Whales example for whatever. Like I, I've seen some, some videos of that. It looks great. Like that's, that's really amazing. Like I couldn't build that, you know what I mean? So like I think some of these tools. Yeah, you're gonna have These expert engineers that are building like awesome sort of products that you can't create yourself. Otherwise, I think there's a whole layer of SaaS that is just gonna be replaced by Vibe coding internally. You know, I think that most SaaS becomes a node in a workflow or something along those lines. It's not like a destination where you need to go and do something. So I think, yeah, there's going to be a layer of tools that are must haves that you can't go and replicate. And then beyond that, I think most things you're just going to be able to build and do yourself. Now, the next evolution of all these tools is all heading toward autonomy. So right now I have to know what questions to ask and what good looks like to get the most out of AI. Eventually agents are going to understand this. You know, my brand and my preferences and what I'm looking to accomplish, and it's going to go out and automatically do the work for me. So we're not there yet, but. But it's coming, right? Okay, so now we've got some information on a few competitors. Connor, you can tell me if these are relevant or not, but we've got Analon X, Cooksey and Freeling Black Cube.
- [63:48]
Cody
I'd say Analon X is very relevant. Cooksy and Freeling Black Cue, probably not as much.
- [63:55]
Connor Rolane
Okay, cool. All right, so how do we take this maybe like a step further? We've got some competitors, we've got an understanding of our own brand. I mentioned keyword research. So use the Data for SEOMCP to identify the top 25 keywords I should focus on to generate more organic traffic than my competition. So this is another really good mcp. It's called Data for SEO. I've canceled a lot of expensive, like keyword research and SEO platforms after I was able to figure out how to use this. So right now it's going and looking for like what you're ranking for on Google. Then it's going to go and look for like what your competitors are ranking on. It's going to go identify some new keywords that we can go and focus on. On. So I like to think about this from the lens of what are some good keywords that I can go and attack without having to write a bunch of blog posts or, or things like that. I want to create like LLM friendly content as well. Now LLMs favor things that they just can't answer in the chat window. All right, so like think about tools, calculators, you know, things like that. So Oftentimes I, I try to design like a programmatic SEO strategy. Maybe I have, you know, comparison pages for all the major pans on the market or something like that. Maybe I have, you know, like a pricing calculator or you know, something along those lines. You can't get the same output in the LLM, so it will cite you as a resource to go and utilize it to answer your query. So that's a little hack for, you know, thinking about how to, how to optimize for LLM discovery as well. It also helps you with, with Google discovery too. This one can take a little bit longer sometimes, but it's, it's trying to work.
- [66:05]
James
One thing.
- [66:06]
Connor Rolane
Go ahead.
- [66:07]
James
I was gonna say one thing I like just watching you do this is that you're so involved in the process, which I think is a little bit different than how people are thinking about AI now. They' they're looking for like autonomous solutions or even, and you can correct me if I'm wrong, But like the N8N stuff is like they're strung together and it's running. And I like that you're like, it's like a AI exoskeleton, right? Like you're just supercharged and you can edit the research if you need to, or you give feedback on the competitors and things like that. And that's what I think feels interesting to me, was seeing this play out in real time.
- [66:39]
Connor Rolane
Yeah, yeah. I mean, again, like, I think there's this whole, now there's this whole kind of almost cottage industry on like X and social channels around like these mega complicated, like in it, in workflows and stuff like that. What if you need to maintain that? What if you need to change something? What if other people need to go in and utilize that somehow? It becomes very difficult and these workflows are pretty static in their design. You know, you might have 20 different things on there. And how does someone know where to go and where to tweak? One thing I would rather lean on AI that has an understanding of the code base and the logic and I can just prompt with it to make an update, to optimize, to change something, to build a front end for it that people can actually go and access and use. Another problem with a lot of these workflows is that where's the output live? How does someone interact with it? Is it going to dump everything into a spreadsheet? Then what do I do with that? You know?
- [67:34]
James
Right.
- [67:34]
Connor Rolane
So the way that I think about it is this allows me to go and build like, like purpose built outputs as well based on the data. Maybe I can build a nice report for my boss or my team or something and send them a link and they can go access it. Maybe I can pipe that directly into Slack. So I think of it more as a co pilot with unlimited opportunities with how I want to format that data and close the loop into actually deriving action from it. So I don't know what's happening with this right now. I'll go back into it in a second. But like just with this use case. Right. Like I built this Shopify site in cursor with Claude code, I'm able to do keyword research right here and now if I wanted to. I could also go and execute the SEO or the LLM discovery strategy right here by building out the pages that it's recommending tied to each of those keywords and then I can push them to production if I want. Right. So like the ability to go from insight to action here is, is unmatched by any other platform or any other tool on the market.
- [68:37]
Greg Eisenberg
That's cool. That's very cool that you have them connected and kind of skips a step.
- [68:41]
Connor Rolane
Yeah. So we'll just see kind of what's happening here. So. Okay. In the process of analyzing the opportunities, this has taken a little while, but maybe I can find an example.
- [68:54]
James
Well, look, you know, on the flip side, it would take Connor Rowland days to complete this. As far as time goes, we're through it.
- [69:01]
Connor Rolane
Yeah, yeah, yeah, yeah.
- [69:06]
Greg Eisenberg
If you know me, you know how cheap I am and how much I love saving money. This year is all about cutting costs and staying lean. You've seen me talk about it on Twitter. I cannot stand watching great D2C brands, even mediocre ones, get ripped off by SaaS vendors in 2025. So many of them are charging way too much over promising and under delivering. And so I can't stand to see it. So I don't want to see it happen to you. I absolutely won't have it happening to me. You've also seen that I'm all in on a. And both those reasons are why I love Rich Panel so much. It's about half the price of the beautiful software that we used before, and it was built specifically with AI in mind. We made the switch to Rich panel right before Black Friday last year and I almost regret not switching sooner. We had zero downside. The implementation was pretty immediate, which actually shocked me. I was a little worried about it, not gonna lie, but it just went off without a hitch. We saved money. We leveled up our CX at the same time. So here's our numbers. We cut our CCX team from 18 to 10 agents since not before Black Friday, but since onboarding to Rich Panel. With our order volume staying the same or even going up, our CSAT has stayed rock solid at 4.2 out of 5. I'd like it a little bit better, but that's, that's stuff that we can do. AI is handling 70% of the tickets, which is awesome. So our ticket to order ratio has gone way down from about 40% to 14%. So we have fewer full timers. Our revenue employee has gone up. We don't need as many people to get back to people and able to be a little bit more strategic with where we allocate our resources. So if your current CF CX software needs more people, is more expensive than you want to pay, and has some janky AI solution, I highly recommend that you check out Rich Panel. Rich panel guarantees a 30% ticket reduction in the first 60 days or your money back again. It's super easy to switch. They take care of data migration, staff training, all that good stuff. So you can go live in under 14 days. So check out richpanel.com and tell them you heard about them from the marketing operators podcast.
- [70:52]
Connor Rolane
Here's an example of one that I did the other day. So I was doing it, I have a website and I wanted to figure out like a keyword strategy for, for me. So what, what this recommended was I do like SEO and AI tool comparisons or something like that. So all this is real keyword data, real growth information, competition, cpc, all the stuff that typically you'd have to go into like, like Semrush or Hrefs or something like that to, to get. And again those tools, like they're hard to use. So I would much rather just be able to ask an AI and, and get the output. So here's what it gave me 25 high value commercial keywords for AI SEO using data for SEO's keyword data. And it's segmented it into these like clusters of different keywords. So the first cluster is like AI SEO tools and software. It's got the target audience, it's got the campaign focus. So this is focused on product discovery and comparison. It gave me a few here and then you know, it's got a bunch of other clusters. And what I was able to do from this was build out like a programmatic SEO page that compares tools or whatever. And I think I arrived at like, I Don't know, comparing like Claude and ChatGPT as kind of an example landing page for marketers. And it built it out based on what successful, like, you know, pages look like when you're trying to accomplish something like this. And now I have a template that I can go and replicate a hundred times, 200 times, 300 times to target these long tail, low volume keywords that are out there. And, you know, what does that mean in terms of incremental traffic? I mean, depends on the data here, but, you know, if each of my 100 pages are able to get, you know, 100 a day, like, that's a lot of incremental organic traffic that I'm able to capture from Vibe coding, you know, and I, I think the big thing here is again, like this, this is a closed loop workflow where I'm able to get the information and actually deploy something that can go and generate a business result. And to me, that's, that's what Vibe marketing is all about.
- [73:05]
James
Love it. Cody, did you have anything?
- [73:09]
Connor Rolane
Anything?
- [73:10]
Greg Eisenberg
No, my mind is blown. It's.
- [73:11]
Connor Rolane
It's okay.
- [73:12]
James
Cody's having his. Cody's having his cursor mind blown. Moment live on the pod.
- [73:17]
Connor Rolane
So I'll just show you a couple other things. You want to see some stuff? Yeah. All right, cool. So I'll just show you, like, again, I have never worked in Shopify. Like, I just wanted to see if I could build a basic Shopify site. I don't know, this week, week. And so I just came up with a random idea. I think within Shopify they have this like, boilerplate that has like snowboards or something like that when you're like, just looking at their, like, free theme that they give you when you sign up for a partner's account. So I just like started there and I was like, all right, it's already got some products loaded up. I can play around with this. So I created a logo in ChatGPT. I have Claude Code working on this site. I went to Unsplash, I grabbed some images for it to use here, and I had those in a folder and I was like, hey, Claude Code. These images are in this folder with the logo. See what you can do. That's it. And then, yeah, it designed a pretty nice little site here. I went into Chat GPT and I made some like, example product images just to make this look good. And, you know, it did a pretty good job with very minimal back and forth. So just the fact that you can go and like create this is, is pretty cool. So you know, and I, and I think this is just kind of scratching the surface. Like I, I think the next phase is I'm gonna try to find a friend that has like, you know, a real like site with a bunch of data and a bunch of information. I'm gonna see if I can essentially clone their site in cursor and play around with it it and see what I can create, you know, but you know, it's, it's, it's pretty neat.
- [75:00]
James
Did you see Shopify Horizons, their latest release and what they're doing with themes?
- [75:06]
Connor Rolane
I haven't seen that.
- [75:07]
James
It's, it's pretty cool. But they're trying to make like AI native themes. So because what you just described, you custom coded the theme. But how most Shopify brands start is you just buy something for 200 bucks off the theme store or whatever and the, the latest ones are built in a way. I don't know the technical details of it, but you could go into the theme editor and just say, hey, when you hover over this image, I want you to jump to a lifestyle image. And then it will just like create that feature or that change, which I think is incredibly cool for brands just getting started can all of a sudden have all the features that they want, all the specifications without having some big custom theme built.
- [75:43]
Connor Rolane
Absolutely. Yep. For sure. Yeah. My, my, the challenge I had was trying to work with an existing theme. So I tried to have Claude code work on an existing theme, but it was, it was sort of tough so I found it easier to like build my own. Now I'm sure Shopify their own tools have, you know, the ability to work better with like, you know, existing sites and existing themes, but I think it's a huge opportunity for them, just generally speaking. So like, okay, so let's say that you're the person in cursor or whatever that figured out the workflow, but then you want to like empower your team or something like that. So I, I built this tool here that essentially runs that same process. I was trying to lead you through with the agent, but it timed out or something like that. So what happens here is, you know, you enter your URL, the type of business you have, maybe it's ecom or whatever, you enter in a couple API keys and you know, then you start an analysis and this will spit back, you know, a bunch of high priority keywords that you can go and, and accomplish. Now in terms of like, use and user friendliness, I would much rather interact with this if I'm not the one you know, operating some crazy workflow versus like a Google spreadsheet or you know, going into N8N or something like that. So just another example of how whoever is the one like building these solutions can kind of close the loop with internal tools. I think internal tools are really powerful. You don't have to worry about a lot of the things you have to worry about when trying to get them to like production or something. And they're shareable within the organization, which I think is pretty neat. I built a. This is a Google Analytics tool that I built. No one likes going into GA4. It's hard to find insights, it's hard to find information. I can chat with my analytics data right here. I vibe coded this with Claude code. I can schedule reports to go out to whoever I want. So I build a report scheduler. It's got like a preview email that looks pretty good. So imagine any data source that you have, as long as you can store that data, you can work with it with Claude code and do something similar. So, so noob, Noob question, but how.
- [77:59]
Greg Eisenberg
Did you get the data like Google Analytics data connected to this?
- [78:02]
Connor Rolane
Yeah, so I set up Google Cloud Console which is essentially like storing my Google Analytics data. They have some integrations with Google Analytics so that you can pipe that in pretty easily. I used AI extensively to kind of go through there and ask questions and figure that out. And then once I have those connections built, all the data is living in Google Cloud Console, which is just the database. And I can say, hey, cloud code. All my data is living in this Google Cloud console database. Let's work with it. Here's what I want to build. It'll say, hey, what's your like API key? Or whatever? I throw that in there and then it's connected to the data, you know, and then you can build on top of it. So I said, hey, I want to have a chat based system so that I don't have to scroll through this. And then, you know, it's got a bunch of different data broken down for me if I want to schedule these reports to go beyond email and go into Slack or something like that, that, that's very easy as well. Getting this into Slack is a Slack API call away. And the way that you do that is by asking Claude code or, or whatever your tool of choice is for how to, how to get it to accomplish the task. Yeah, like this is the, the cuts versus true classic thing that I was telling you about. So again I went and I scraped all their ads. I found competitive gaps I created some videos and I built this landing page report off of it and I'll just kind of show you like what this looks like. So it's got like an executive analysis or executive summary with like, you know, summaries of the brands and some key opportunities goes into a brand positioning analysis for both of them. Key strategic differentiators. I found it really interesting that Cuts is primarily targeting women as gift givers as like an icp. So that was kind of an interesting nugget. I didn't really realize that they might be doing that. And then yeah, has kind of their typical scripts here that they go through in all of their video ads broken down. Very useful information. What are the pain points in the market? So again this uses like Reddit, YouTube and Amazon reviews to identify things that people are actually talking about out there that might be unaddressed or you know, under explored by both of these brands. So summarizes all these pain points that, that we went out and found. And then it talks about the gap analysis that I mentioned. So what are the, what's the brand addressed today? But what are they missing? I kind of zeroed in on this true classic win, like long term durability, white shirt opacity. So I think I mentioned that before but like people were talking about when they buy shirts they were like pilling or losing their, their color after multiple washes. So I thought that hey, that could be an interesting gap to explore with like an ad. So anyways, it goes on and on and then you know, okay, like analysis is great, but then how do you actually like, you know, turn that into something I can use right now with VO3 you can create like 8 second previews, you know. So I didn't prompt this to create like an 8 second video output. I was just like, hey, create something so that I can like, like, you know, make an example. And I thought it did a pretty decent job just like out of the box at, at creating a couple concepts that then maybe I could take to, you know, like let's say an agency or my internal team or something like that.
- [81:35]
Cody
So I bought this six months ago.
- [81:37]
Greg Eisenberg
Worn it every week, still get compliments.
- [81:41]
Connor Rolane
So anyway, like you can see some.
- [81:43]
James
Like this true classic shirt, six months old, worn every week, week still looks brand new.
- [81:50]
Connor Rolane
So you know, like you can just get an idea of like where you can go and how you can develop some concepts and briefs and everything. And I don't know, like here's the process. Scrape, extract, transcribe, research, analyze, create and deploy and to do this kind of analysis manually, I mean it just doesn't compare. I was able to do this in a couple hours.
- [82:11]
James
100%. These are some fantastic examples to kind of close us out here. We had some questions on Twitter. Andy Rosenberg. We've talked about a lot of different ways that brands and marketers can be implementing AI. I'd say so far this year, from my perspective, and maybe Cody and Connor, you guys disagree. It's really just been around asset creation, the VO3s, chat, GPT getting really good. That's where brands have been spending a lot of time. Andy Rosenberg asks what will be the next biggest implementation of AI in the second half, half of 2025. Do you have a prediction, James?
- [82:46]
Connor Rolane
You know, I have a prediction for, for EE Commerce and DTC. Like the thing that I keep thinking about is paid media and media buyers. Like, I mean from my perspective, like Meta is just going to create agents and tools that are going to replace most of that work. You know, I, I think the way they've been heading over time and, and you guys can like correct me if I'm wrong here, here, but it's kind of leaning more on the algorithm. You know, people aren't like hyper targeting like specific interests and things like that anymore. It's mostly leaning on creative and kind of like, you know, having a good creative mix and you know, launching a bunch of ads and letting the algorithm do the work. So what's the role of like a media buyer look like 12 months from now? Are you even going to need to go in and log into Meta or do you just give it a budget it and say get to work? So I think that's, I think that's one thing that will develop and that's sort of a microcosm of the larger macro direction where I think autonomy is just going to be the next major shift. Like right now I'm going into these tools, I'm chaining these things together, I'm identifying what I need to do and what good looks like. But agents are going to be able to take all of that off my plate and I'm just really going to be orchestrating these agents that are going and doing things for me. I'm not going to be executing the tasks, doing the work, coming up with the approach, building the thing. It's going to be about directing teams of, of smart AI agents. And yeah, I think that's super exciting.
- [84:23]
Cody
Yeah, 100% we're doing that. By the way, we have, we're using a tool in one of our market just one to test it out and see how it goes over the last like three months where, yeah, we're not buying media. Like there is no manual media buying happening. It's the tool we use as an API connection with Facebook and they are moving budgets, they are moving bids, they're doing it on a daily basis. So it is very much like kind of going back to like the 2021s of like more manual day trading and media buying, but now using an AI agent to do it. And like from what we're seeing in the data, like if you look period over period from like the moment we started doing this versus our manual media buying, if you look at the year over year comps, both blended and in channel, like, it seems pretty good. We've had some stock issues which has made it like kind of hard to see where it could have taken us without that. But yeah, we're there in our UK account and it. I think it's only a matter of time until that's more broadly being used.
- [85:21]
Connor Rolane
Awesome. Very cool.
- [85:23]
James
Yeah, I love it as a second half prediction. Well, I think this was a great episode. Cody or James, anything you guys want to wrap with?
- [85:29]
Connor Rolane
I mean, I think just the opportunity to dive in and learn and work with these tools is there and they can look crazy at first and scary and stuff, but you just got to dig in and play around with it and tinker. I think that just dedicating a little bit of time every day to this can really compound and I don't know, I think it's the best time ever to like be a marketer, to be a builder, to be embracing this type of stuff and super exciting.
- [86:01]
James
100% love it. And everybody can find you on X at Boring Marketer for all the latest insights.
- [86:07]
Connor Rolane
That's right.
- [86:08]
James
Awesome. Well, cool. James, thanks for coming on.
- [86:11]
Connor Rolane
Thank you. Thanks for having me.
- [86:12]
James
All right. We hope you found that episode actionable and insightful. You can follow James at boringmarketer for more advanced advice. As always, thank you to our sponsors, Motion Rich panel. After Self Prescient and House and like and subscribe, let us know what you want us to talk about in future episodes and we'll see you next week.