Stop Writing AI Marketing Slop (Do This Instead)

It's not a fun time, I think, to run a traditional SAS company >> 100%. >> Every time Claude comes out with a plugin that's for cyber security or legal or financial services and suddenly you can hear the stock just going down for some of these companies. And I think people are adapting. I don't think all these companies are going away tomorrow, but my gosh, it's a hard time to be one of these companies and figure out what is your actual moat when we have generally capable models that seem to be getting better at light speed. Honestly, one of the things I like doing with agents, this is for more content writing, is providing push back. The quality of whatever you're writing is 10 times better by just getting it to push back on itself. >> AI marketing slop, it's everywhere out there. What the heck is going on out there, Mike? But no matter where you're at, you really should be experimenting with some type of agentic AI. Whether that's Claude Co, Claude Code, you need to start thinking about your own career. >> You're watching Marketing Misfits with Norm Ferrar and Kevin K. >> Senor Ferrar, how be you? >> Great. But look at you without your glasses. Oh, well, you know, sometimes I got to look smart. >> Like, what happened to the ones the glasses that you had that made you look like Junior on the Sopranos? I miss >> Junior on the Sopranos. I don't know. I I went to these uh these Warby Parkers. I know. And everybody always says, "Uh, oh, cool glasses." Uh, >> there you go. You're just trying to be like me. >> These are better than the ones you get at the dollar store or the drugstore for like two $2.50 50 cents or whatever that I used to wear cuz I had them all over the house. I had one like in every room. Uh so I always have my little reading old man reading glasses. >> See, all you have to do with that one though is grow a beard and we could be brothers. >> Uh we could uh we could but you know I always have to have the glasses so I can read your text cuz I don't want to miss you know when you send me one one of those texts you know late at night. I I don't want to miss it. >> Just one of the like I'm one of these single line textters. Kevin, next text. Da da da da. Next text. Single line drives some people crazy. >> Speaking of single lines, you know, single lines uh when it comes to like AI used to be the way everybody was doing stuff. You know, AI prompting single lines. Now, you've heard me talk about uh briefs. You need to really write a full brief. But I don't know about you, but you know, I've been playing this AI game. Well, I know about you. You've been playing this AI game since 2022. Well, before that, actually. >> Yeah. >> But when it really took off, when Chad GPT came out, November, I think 2022 is when it really it started this this whole thing. Uh, and we've been using it, you and I, we're marketers. We've been using it for marketing for a while. And it was okay. You know, help you write some emails, do a few things, but man, has it changed lately. And has it gotten so much better in the last hadn't even been four years. three three and a half years now I guess it is or is it four years? Yeah, what whatever it is. Um and now with Claude and you've heard me talk about this Claude Co-work I can bang out just amazing freaking marketing stuff and amazing stuff and now with the new chat GBT 5.5 with the imagery that it can do. It's evolved rapidly and it's going to be interesting today with our guest because that's what he specializes in is is AI mark AI marketing and it's gonna be interesting to see his take on what's happening and where this is going and and you know you and I both know a lot of people that are still misusing it or still not even using it to to its full extent. Um so what what are you seeing out there in the world when it comes to people using AI for for marketing? I I see a lot of people trying but not getting the knowledge and education that they need to understand it fully. Like I I've seen prompts that just suck. you you know you put into it you get back what you put into it and if you don't understand the uh like each one of the LLMs then why bother this you know what this reminds me of is Amazon back in the day when everybody that knew Amazon and how to manipulate it um they won the people that just threw up anything lost and they wondered why they are losing and That's exactly what's happening right now in AI. Um, every day it's changing and if you're not on top of it, you're eating somebody else's dust. Yeah. I mean, there's a lot there's no doubt there's a a lot of slop and a lot of crap, but and you know that in Nashville, I showed that chart of the number of people that are actually using AI. And we had a dot for like every 2 and a half million people. And I forget how many dots there were, but whatever the number of dots, 8 billion divided by 2 and a half um million is uh dots. And there's only one dot of all these, you know, I don't know, there must have been a thousand dots on this thing. Uh or not, actually, I should be more than that, but 4,ou almost 4,000 dots on this thing. One dot of the people that are actually using AI and in the proper effective way. And I I think that's going to be interesting to see today uh what our guest has to say about that when it comes to marketing because the people listening to this podcast are primarily using it for some s form of of marketing to write their emails to write their to create their landing pages to create their uh product listings on Amazon or or whatever it is. So it's >> poetry. >> Oh well that's that's Did you like that poem that uh Chat GPT wrote for you? >> Yeah. Thank you. Thank you. >> My name Yeah. with the gummy bears under my pillow and Oh, yeah. >> All right, that wasn't sloppy, was it? That was pretty good. >> All right, let's bring in our guest today, >> Mike. >> Hey, Mike. >> Hey, guys. How's it going? How >> you doing, Mike? Good. How are you, Mike? >> Doing great. >> So, >> as you can tell, this is going to be very casual. Mike, >> that's just my style. I like it. >> So, marketing slop there. It's everywhere out there on on AI marketing slop. What what the heck is going on out there, Mike? What's what's going on right now, man? >> Unfortunately, I think human beings are being human beings and we love taking shortcuts a lot of the time, right? So, it's I think a lot of marketers got real tempted to start using AI for all their writing without, you know, doing anything to train AI and how they sound and how their brand sounds. And it it's it's I couldn't agree more. It's a mess. We're seeing every social channel flooded with AI slop, aren't we? >> But there's so much crap that's coming our way. And it's not all crap. I mean, it's just shiny objects and new information that's coming in daily. Like, how do you even decipher where to go? >> Yeah, I struggle with it myself. you know, I've had to really over the years of keeping tabs on the AI and marketing space, I've kind of, you know, gotten a good sense of which people are uh solid narrators of this stuff and which people to follow. So, that's trusted voices are always helpful. But for me honestly I in some ways I confess I use AI to cut through the noise now because like I need to train AI to actually understand what kind of content commentary advice that I actually care about. Um and you know I've had to really train muscles in my brain to ignore everything else. It's really hard. I I don't know how people do it. I feel like I'm drinking from a fire hose most of the time. >> Yeah. Yeah, I mean there's so much I mean now you have the AI slop teaching the new AI and so you got you got to be got to be careful on some of this stuff and it just it's it starts this cycle of regurgitated just crap uh and it affects it brings it all down. But I know the the guys that that are coding this stuff over at Anthropic and uh OpenAI and stuff they're putting in some bells and whistles and I know what is the latest 4.7 as the time of this recording opus. Yeah, >> that actually is you have to be ve very much more specific now. I think it's it used to be like okay you would write a you could write something kind of general and it would kind of like interpret it a little bit like oh I think he means this based on this and now if you don't do that um you get crap. Uh if you're specific and very specific you get better stuff. Um what are you seeing on on that aspect where whether the the models are changing to accommodate for some of this or try to balance it? Yeah, I think that's largely kind of how I see it going. It's actually really frustrating, right? To anyone who has uh fallen in love with how a model works for them, whether you know a lot of people like GPT40 way back when, people love Claude Opus 4.6. I was one of those people. And it's, you know, there are real changes in tone and personality and how these models approach your work. I think you have to really focus on just I you know it's you kind of really have to sit back and start being strategic about thinking about how specific you are with the different tasks and prompts and briefs you're giving AI and also what outcomes you're trying to achieve. And I it's why we always tell people like, you know, we've all got our favorite AI tools we're in love with. But you really do need to be if you can and if you can afford it, testing out multiple models regularly and cycling through those because things change faster than I would like sometimes. Every new update is exciting, but every new update can also torpedo what you used to like about a model and what you used to be doing with it, too. Well, you like Nano Banana, you know, Sora came out and that was the hot thing with video. Now Sora is is moon is sunseted a year later. It's it's like it was a hot thing and now it's gone. And then Nano Banana came out uh and then the the pro version or 2.0 or whatever. I get them mixed up, but whichever they they came out with even a better one that was that was a rage. And then Chad GPT put their 5.5 model out which blows it out of the water. So if you're not using I use about five tools and luck and the highest level on all of them. So I'm spending you thousand bucks or whatever it is a month but to me it's worth every single penny of that and and I bounce between tools depending on the task. >> U and so I don't just use I'm not just married to one and I know Norm you do the same thing. Right. >> Right. And you know, even outside of that, you're talking about Nana uh Nana Nano Banana and you know, these imagery LLMs. I like using uh I don't know if you've heard of it, but Open Art >> and you can bring in all the images or wherever you want to bring it in from, but um you can write the prompt, but it'll create a better prompt. So, you get even more photo realism or whatever that you want. >> And it's very inexpensive. And I find that right now uh that's the best thing on the market, at least for me, because uh again, we talked about writing a proper prompt. Well, this will do it based on whatever LLM you're uh importing into it. >> Is that the one you used for those sexy ab photos that you were putting out on? >> Yeah, of you of you in a bikini. It was awesome, Kevin. And then I did one of you like there is Marilyn Monroe going over the great dress down and pose. Yeah. >> Don't give the audience any ideas. >> And I do have a like another question on how do you know like let's say you're in Claude and you want to uh change the model. So, you know, you've got Opus 4.7 right now, but maybe you should be using Sonnet or maybe a a different model that's in uh Claude. How do you know when you should be using one or the other? >> You know, it's really difficult and I would actually argue, at least from what we've seen, a huge amount of people, probably the vast majority don't even bother to switch between models, you know, unless they're switching between tools and trying out the newest one that came out from Chad GPT or Gemini or Claude or what have you. But a lot of people don't really take the time to switch between models. I would say I think people are getting a little more aware of this because of all the issues with usage and compute is like anytime you have something that's a little less cognitively demanding definitely scale down to a smaller faster model that costs less in tokens and usage but how you do that is really hard. I mean how do you determine what is and isn't super cognitively demanding. I think that's why actually as a larger question we're going to run into a lot of really murky questions around like people hitting usage limits because like even if you told me today like I I use this stuff pretty regularly and if you told me like hey we have to cut down usage by 50%. like what tasks should we switch for a you know a lighter weight model I I don't know if I could tell you really well every day with the hundreds or thousands of things we are doing with AI like where that breaks down so I think we're going to run into some real problems with that but yeah I think generally if you're doing more simple things especially around things like content generation where these models have gotten very very good at especially shorter content generation like I would probably not be using the absolute a state-of-the-art model if you're worried about usage limits um or if you don't want to see what it can do. >> Man, things are getting tough out there. Ardu are getting squeezed. Things are changing. AI is playing a big role in what's going on in the e-commerce world. And there's a lot of people that are they're pretty worried. What What should they consider doing? >> Well, I think one of the best things they could probably do is go to an expert that understands the market sentiment right now. First one that comes to mind is is Quiet Light Brokerage. And here's why. They're going to build you up. They're going to understand your company. And at the end of the day, you're going to know how to maximize your valuation. So, the very first thing you need to do is go and get your free confidential valuation at quiet.com and then, you know, let the games begin. >> Awesome. What What was that website again? It's quietlight.com. >> Awesome. I'm going to head over there. So, if you're creating a wireframe, for example, you've got an idea, you you you're using co-work and you're creating a wireframe for something, you know, Opus um 4.7, but everything else like all your copy or anything that you're asking it to do or alter or revise, that could be all done one notch down, go to Sonic. Yeah. >> And the other thing that um uh what was I going to say? Oh, go ahead. I'll remember it. I'm old. I'm a fossil. Kevin, you could ask the next question. >> A fossil? I was wondering what that petrified look was in your eyes. >> That's just uh not being able to remember what my next question was. >> So, so yeah, like like Norm's saying, you know, going down to a different model or switching model. I mean, I use Claude primarily for for writing. >> Yeah. Uh, I use, uh, now I'm I was using Gemini for images, but that's switched now to to OpenAI. Uh, I use Plexity if I need to, you know, I want something that's going to optimize a lot of stuff or complexity computer if I have a lot of agentic stuff that I need to go on. And I use Claw Co-work for stuff that's on my desktop. Um, I was just telling Norm the other day, I mean, this is just this is not a marketing task really, but or in a way it is, I guess. Uh, I have Beehive, which is my newsletter. uh they just added a podcasting platform uh inside beehive. So I I have been actually creating a using AI uh in 11 Labs. I created a voice clone of myself. I create a audio version of my newsletter that goes out twice a week. I send it to a guy in the UK. Um and for 300 bucks a month, he does what's that 10 10 so like $30 an issue. I don't I just don't want to mess with it. and they take take it and they summarize the entire newsletter. They they use AI, I'm sure, to summarize it and write a script and have me read it and send it back to me. Well, I'm turning those into podcast and I put that in the newsletter as a a list as a as a uh a link, but also I can get those up on YouTube and Spotify and other places, and that's just lead, you know, it's just more uh uh lead generation for me, as well as people that just want to listen to it rather than read it. And so we're we're creating that, but it had to be MP3s uh and to upload to to Beehive. Um and then I I use audiogram to actually put the little audio stuff there. So there's a picture of it and and it's it's a complicated process, but not complicated, but uh several steps, but it's pretty cool. But I I need to convert all these wave files to MP3. And I was telling Norm I went online, found some free tool, and it let me do 10. I had 39 of them because I was going back to the first year. So it let me do 10 for free. And it said to upgrade. Wait a second. Claude Code can freaking do this. What am I even talking about? So I just directed Claude Code to go to that directory, convert all these MP3, and you know, for basically next to nothing or part of my $200 a month plan. I guess it probably use a few credits there, but I had them 3 minutes later done uh completely. And so a lot of AI, these AI tools, you know, Norm was talking about the image creator that he likes. idoggram was a hot thing for a moment and now idoggrams basically put out of business. I mean they're still around but they they're they're hanging on a thread u because now all every time these uh big LLMs come out with an update they put a thousand tools out of business. >> Yeah. It's not it's not a fun time I think to run a traditional SAS company is my sense. No, someone today in our space, um, one of the guys that advertises in my newsletter, his video, his, uh, email that he sent out today, um, it's an advertising email that I do. It was all about using Claude, uh, to do keyword research for Amazon instead of using Helium 10, a $100 a month tool. He's like, you don't need Helium 10 anymore, >> and let me show you how. >> And he he showed a basic prompt. He said, this basic prompt gets you what Helium 10 does, but now let me show you how to put gas on it and get you way beyond what Helium 10 does. And it's a little 14-minute YouTube video. And I was like, so SAS companies like are in trouble um big time. >> Yeah, it's it's pretty crazy to see. I mean, we've seen it in the stock market so far this year and there's a lot of factors behind it, but yeah, every time Claude comes out with a plugin that's for cyber security or legal or financial services and suddenly you can hear the stock just going down for some of these companies. I think people are adapting. I don't think all these companies are going away tomorrow, but my gosh, it's a hard time to be one of these companies and figure out what is your actual moat when we have generally capable models that seem to be getting better at light speed on it. >> There's a lot of dinosaurs out there, you know, and yeah, I just remember Kodak when we were coming into the digital age. Nobody thought it would go down, but I did remember my question and I think this could be a gold nugget. All right. when you're working in, let's say Claude, for example, >> when you're prompting it and you've only got so much information before your session runs out if you're using the $20 a month, even when you're doing the pro model. But, um, is there a way that you can prompt it to use less memory so your session can go longer? Like I'll usually end be concise. Yeah, I think that's probably the key for me would be just determining what is the best format for its responses overall in the conversation that in in a way to conserve usage. Um, I confess I'm I'm I'm I have an embarrassment of riches over here just having a clawed max like 20x account. So, I'm just uh you know wasting water over here probably with AI. >> Can I be your bucket? Right. Right. So, but yeah, I think you're on the right track how you how you talk to it. I think using any type of files you're using with it, like markdown files are quite a bit more, I believe, token efficient. So, that's helpful. Um, but again, it kind of comes back to that point like, yeah, there's a couple strategies you can definitely follow. But my gosh, if you're routinely hitting usage limits, it becomes just a real pain and I don't know how you get around it except for running your own model or, you know, we're all going to end up paying these companies more and more, I guess. >> Well, they're bleeding money right now, too. I mean >> right >> I mean and they're built but the the infrastructure >> was it 490 billion or maybe it's more than that is committed to infrastructure right now for AI between the between Meta, Amazon, Microsoft and Nvidia I think but just those four uh and I think yeah I think it's $490 billion in the next uh under construction and in the process right now for AI power uh infrastructure but what what about the the so Let's talk about it from a marketing point of view though using AI a lot of people are still using AI as like Google they go and ask it questions >> there's nothing wrong with that but when you're coming using it as a marketing tool um what are some uses I mean there's there's several pieces of software out there that there's one I don't know if I want to give the name away because Norm and I use it in in our our company Dragon >> let's not do that >> I don't want to have everybody have it but there's one there's one particular tool where >> instead of you going to one of these to let's just say Chad GPT is the easiest one and saying, "Hey, write me an email for uh you know, I'm I'm doing this promotion or whatever. Write me an email to a three-part series email to do this." And it it'll spit out something and sometimes it's not bad, sometimes it's depends on what you how you how you prompt it. But this other tool, you put in, you give it the URL, it'll create an avatar uh uh based on that website. It will then go out and create all these pain points of that customer. And the more you feed it, the better it gets. And it's got all these different tools built in. And so you basically build an infrastructure first say this is my avatar. This is what they like. Uh this is their pain points. And now they may write an email uh a three-part series email. And it's got some that are in different tones. And one that's a click drivers. It's just meant to get the click. So it's a very short to the point one link type of email. Others are like long stories like how to get in somebody's head if they want to read like a 20 p 20 scroll email. A long long form email I guess you would call it. Um, and it'll do that. But it's all trained on like 150 years of psychology. So, they've gone in, they've actually chosen like these these copywriters are the great copyriters. There's Oval, it's this and like, okay, and taught it's kind of like a sublm and this is the waste and then it's trained on they they're copywriters themselves. So, it's trained on their stuff and then it writes in a whole different way. I test it against Claude, Perplexi, Gemini, and and it almost always wins in AB tests. >> Um, and it it's that's a powerful use of AI in my opinion to to save a lot of time and a lot of money and do a lot of research. >> Are are you seeing more like specialized tools come out like that for marketing and the people that are using them? You know, Benson was another one. Uh John Benson had his little thing and there's the uh there's there's some others, but what are you seeing specialized versus just the what the the general LLMs are doing? >> Yeah, I think that's actually a really good correlary to what we just talked about with how much trouble some SAS companies are in. think there's real opportunity, at least today, for some of these marketing and sales providers to be creating tools that are fine-tuned and customtrained for these really specific purposes, especially if you've got that domain expertise in something like copyrighting because you absolutely can get better results from tools that are created that way versus your generic LLM. Now to be fair with enough time and energy you could easily I think or relatively straightforwardly train a model to a general model to do the same thing with the same output but the key is nobody wants to take that much time to do it and it takes a lot of work and energy to figure that out. So I think there's a real advantage here. I don't know if I'm seeing an explosion of like specialized third party tools. In fact, there's probably I mean there's probably quite a few out there more than when we started you know tracking this stuff back in like 2016 but I think you know even pre pre- chatbt it was all these interesting specialized third party tools that were using AI often in a more predictive sense to do very specialized things and then Jarvis or what was >> exactly yeah was there's yeah quite a few that were Jarvis I think they became yeah like Jasper they became Jasper yeah and they're still around and they're one of definitely one of the leaders, but there used to be, you know, like a hundred different tools like just like that. I don't know if all of those have made it, but I think there's real value in some of these really specialized third party providers. But also, you see some of the major platforms are adapting at least. I mean, the market leaders or traditional marketing and sales companies, everyone's racing to bake specialized AI into what they do now with varying degrees of efficacy, I would say. But I think so I think you're seeing there's still a robust market for these kind of specialized solutions and there's real value to using them I think. >> So what AI skills should every marketer have right now or start to build on? >> That's a great question. I would say every marketer my best piece of advice I realize like everyone is always at such different levels of kind of their AI journey and AI literacy but no matter where you're at you really should be experimenting with some type of agentic AI whether that's claude co-work even I mean you could classify something like deep research in chatbt or Gemini is technically agentic even if you're just starting there you need to start thinking about your own career. Not only what AI can do in a chat window, which 99% of people have not fully exploited yet, um, but also what AI can do when it starts doing work for you. And your job is less about chatting back and forth and more about orchestrating tools to go do things for you autonomously. It's a subtle but important shift. Like I, for instance, me personally, I've since late last year, I've basically exclusively kind of gone all in on cloud code. I mean I still use have accounts and use all the other tools but just for what I do like having more agentic capabilities has been an absolute game changer. You don't have to like learn everything all at once and build all your own agents right out of the gate though that'll be interesting to do moving forward here. that you start shifting how you work and how you think about what's possible when you start having to manage instances of AI working autonomously for you for several minutes or you know some people even have them running for several hours or you know codecs from OpenAI is another version of this co-worker way to use this but you got to start experimenting with these tools because I think it's pretty safe to say however long it takes agents are going to be the future and it's going to be cropping up if it hasn't already in every single marketing platform you use. Agents are going to fundamentally unlock new capabilities in marketing. So that would be my biggest thing is if you haven't done it yet, start getting your feet wet with that. >> But to use an agent, like you said, you got to orchestrate it. You use the word orchestrate, you got to basically be a manager. >> Yep. >> So if you relate that to the human world, there are managers and then there are worker bees. Yep. >> So, how are the worker bees that, you know, if I'm a manager of a copywriting team and I got five copywriting cubs underneath me that are learning copywriting and doing some of our stuff and I'm managing them and kind of guiding them, how how I can probably make that shift to a gentic a little easier because I understand how to delegate and how to manage and how to what's what's good and what's not. But those people, they don't know and they're not maybe cut out to be a manager. They're cut out to be a worker be they they're not they don't have entrepreneur spirit or leadership spirit in them. what what happens to them? >> Well, I think that's probably why I'd counsel people to start trying out this stuff because it's going to be an uphill battle for some people. I don't think you're totally, you know, screwed if you if you're not kind of managerially minded, but you're going to have to start exercising some different muscles. You know, we're actually in a couple days here, we're doing a virtual half-day event, our AI for writers summit, which is uh something we do every year. and it's a a virtual event with a free registration option. And as part of this, you get a lot of individual contributors trying to learn about AI. And it's a very foreign thing for them to say, well, wait a second. I'm not just doing the work. I'm going to actually have to be managing instances of AI. That doesn't really sound like what I like to do or what I'm good at. And I think I would just say there are hopefully ways you can do your job moving forward that play to your strengths and to the things you enjoy about your job and the things that create value. But you're going to have to learn this. I just don't, just to be perfectly blind, I just don't see a path where you're going to be an individual contributor in marketing 24 months from now and agents are not some part of what you do. One of the things I like doing with uh agents, and this is for more content writing, is providing push back. >> So, I'll create a I Kevin knows all about this, but I'll create a research team. It'll go out, research the trends, it'll come back, it'll provide the information to the next agent, which will uh be the writer, but the writers looks at the draft and gives push back and so it forces the research agent to go back and rewrite it, come back and then it brings it to the writer. Then the exact same process happens to the editor and it just your the quality of whatever you're writing is 10 times better by just getting it to push back on itself. >> I've seen that push back sometimes it becomes worse. So some have you >> I have not experienced that Kevin. >> Not always. No, not a lot of times it does get better, but I've had times where I'll push back on something and actually the second or third version is actually worse and I go back a version. >> You know, it's interesting. I've actually had both ways happen. I think it sometimes comes down to you do have to watch out a bit for the sick of fancy of these things. Like if you keep saying like, "Hey, I'm pushing back. Shouldn't it be this? Shouldn't it be that?" It's like going to kind of sometimes agree with you and then do something worse or do it in a way you didn't anticipate or not keep the stuff you liked. So, it's all the all the fun of these models that operate on probability is not determinism. >> Yeah, >> it's all math. I mean, it's just um Have you heard of first principles thinking? >> Mhm. >> So, I uh Norm's heard me tell this, but uh the podcast listeners haven't, but I was just at a mastermind about a month ago, and one of the guys uh that's co-founded the mastermind is big in the AEO space, answer engine optimization space, or G some people call. There's slightly a difference between them, but um he he said, "Hey, I want you I'm we're going to do a little workshop. I did a full day workshop around this, but um I only got 3 hours here this afternoon. We're going to do uh hands-on." So, everybody pull out your computer, your phone, go to your favorite LLM, type in uh ask it a question of something you want to solve. So, maybe it's some business thing, maybe it's a personal thing, whatever it is, just ask it a question and and and see what it says. So I went in, I said, "Hey, how do I get more subscribers uh to my to my newsletter?" He said, "Use a tool that kind of knows you, that you have a history of uh so it kind of knows you." And it spit back, "Oh, well, Kevin, you should do this, this, this, and this. Here's some ideas." And they weren't bad. Um, they're pretty good. And he said, "All right, now we're going to go through this process of first principles thinking." He he explained what that was where it's breaking things down to their fundamental parts and then building them back up and not accepting anything as as the way things work as approaching it from a different point of view. He said, "I want you to download this uh MD file and go over to Claude uh it's going to ask you some questions um and go through that for the next 30 minutes or so and then if I I'll let you know when you need to tell it, hurry up. Uh we got to finish this up." And so 30 minutes he says that and it it went through a series of some of them personal questions. what hap what do you think what do you do when this happens what happens this or what is your number one goal about this or how and so then spit out a file at the end of this and he's like all right now take that file and go to this other uh prompt that I have for you this load this other MD now it's going to go through another process went through another process a similar type of questions and then he said all right now add it's going to make a another MD file now add that back to the original tool that you used ask it the exact same thing again and and and put this you add this MD file to it and he said now to go and look at the difference. It was mind-blowing how much difference it was when I applied first principles thinkings that were related specifically to the way you think and what you're trying to achieve. >> And then then he said, "Now go do it on two or three other tools uh and you're going to get slightly different answers because each tool approaches different or knows the different information about you from past chats." and and now combine now use this prompt and copy this prompt and use whatever tool you want, but combine the three together and that's your final master. And it was night and day difference on what it spit back out. But that's a level that I can guarantee 99.99% of the people weren't doing. But it also illustrated to me the freaking power this thing has if you know how to freaking use it. Um, and what what do you what experiences do you have in anything similar to that that can be applied to marketing? >> Yeah, I'm I'm so glad you mentioned that because one of my projects over many months uh that's mostly come to fruition is uh I do in my work not only marketing and content work but a lot related to some pretty complex strategy problems for our business, for our marketing, for our audience, for everything that we do. And uh I have over months been training uh and iterating on a skill that is literally intentionally designed for first principles thinking and how I think and approach problems that Claude code has continually added to and updated anytime we do any type of project thinking through any type of problem to solve and it has just become this incredible weapon like anytime that I can bring to a problem where it's like I'm not you know outsourcing my thinking I'm literally doing what I do best but having a model augment my approach and make it so much more robust. The results I've gotten in terms you know I used to do pretty good strategy work before but like it's 10x better easily from this approach. So I think to translate that to marketing, I'd be really seriously thinking about uh almost for lack of a better phrase like what are the meta skills that you use all the time in as a marketing leader, as a marketer, like it's not just about training AI to write like you or write for you or you know go create an email campaign or do data analysis. It's all super helpful, but like what are the things you apply your brain to most often? How do you think personally? You can have literally have AI just interrogate you. They ask me a hundred questions one at a time about how I think and how I solve problems. It won't take you that long to type back some answers or speak back to it. And by that point, you've got a bunch of raw material about exactly how you as a marketer approach marketing work and problems and then drop that into a model, create a skill based on it. Guess what? You don't just have something that does stuff where you have something that thinks with you and thinks like you. And I think that can be broadly applicable across basically no matter what anything you're doing. And I think that's to your point like that's a real level up for any marketer who's listening to this. If you haven't done something like that to create whether it's a GPT, a skill, um a project that just has that information as its instructions, if you haven't done that yet, >> MD file, right? I mean, >> yeah, easily. Yeah, just a markdown file. Yeah, easily >> file. What about What about should I be doing that? Norm and I have a an agency. Should we should our clients we be giving them to do this and then we take that uh file and if if we're writing emails for them >> y >> and you know we're using this fancy tool that I just mentioned earlier may and we go through a series of questions when they on board that we ask them and we use some of that data to actually augment the tool and augment our process but it might actually behoove us to actually have every single one of these clients now we're thinking about this do exactly that create something that's a 100 questions or 50 questions or make it quicker and say all right we need you to go and answer all this and we can just automatically have that spit out into an MD file uh with the mark the markdown and all the stuff and then we attach that to everything we do and it should put some rocket fuel on what we're doing. >> You know, Kev, in theory that's great, but in reality very very few clients will take the time to answer a 100 questions. I know just >> understitting in in my past just getting people for web design or SEO work asking them to give you a boilerplate >> could set you back months you know giving >> answering the 27 questions now maybe we just rephrase that uh make it 20 question you know put get the core stuff you know what's your email address you know the basic uh contact kind of stuff what's a link to your products and then say hey we got let's cut it to 10 or 12 I agree with you 100's way too many me and you will sit there and do a 100 cuz we know the power of what it'll do for us. But maybe it's 10 or 12 that go in that get put in there and then we just create it or something. But I could see how that could really make a difference. I >> I I think so. But Mike, what do you think about that? Like you you're dealing with a bunch of clients. How would you handle that? >> Yeah. I mean, I've been out of the agency world for a little bit, but I'll tell you right now, I think that's actually a fantastic idea. to your point, it's hard to get clients to do stuff like that, but I if you can, then absolutely you should be a try to build have them in their own words essentially help you build, for lack of a better phrase, their personality profile, right? Um, that can be really helpful in the work you produce for them. But I'd also suggest perhaps even if you can't get them to answer something like that, I would have killed back in the agency days for something like this because, you know, you run in a bunch of different clients and you know, they all have their own nuances. I know this guy's going to get upset about this thing or cares about this thing. It's like even just internally with like my account teams or myself being able to like chat back and forth with almost like a simulated version of the client and say like, "Hey, we're about to email them about campaign performance. it's not great. Like how do we frame this for their personality even? I think could be really interesting too to consider or at least the context of the client account all being in one place that can be referenced regularly as long as you know it's information you're able and willing to share. But having that almost like second brain for each client account, I think could be a really interesting idea. Are you looking to quickly boost new Amazon product launches or scale up existing listings to reach first page positioning? The influencer platform Stack Influence can help. That's right. Stack Influence pushes high volume external traffic sales straight to Amazon listings using micro influencers that you only have to pay with your products. They've helped upandcoming brands like Magic Spoon compete with Cheerios for top category positioning while also helping Fortune 500 brands like Unilver launch their new products. Right now is one of the best times to get started with Stack Influence. You can sign up at stackinfluence.com or click the link in this video down in the description notes below and mention Misfits. It's m i sf ts to get 10% off your first campaign. stackinfluence.com. Uh yeah, I mean they give us some of them have a brand guide. They give us, you know, the PDF of their brand guide now or they'll give us a founder story. There's some basic stuff. But I think if we >> if we just went to uh Claude or whatever and said, "Hey, create this is who we are. This is what we're trying to do. This is um >> we need first principles thinking. We need uh you know they're not going to spend a lot of time on this. What are the key 10 to 20 qu give me 20 questions uh that are the key things that's going to help you achieve this goal that we're trying to achieve for them and then okay for the people that don't have time to 20 what should the first 10 be that they don't make it all the way through it. Um >> right and and so I think that it's worth testing that I think that could be super powerful. Now, I'm just thinking out loud here on the podcast and giving other people ideas, but that's that's okay. That's what we're trying to do. But I think back on the your two too on on this, and this kind of plays into it, the moes right now, uh Perry Belchure just talked about this and I've I've been saying that something similar for a while, but he succinctly said it like there's only two modes in this AI world for for marketers really, and that's distribution and data. So, so if we have the distribution, which we do with Dragonfish, we we help them with their email. We also get intent based email using AI and stuff where in market that never been to your website, we can actually capture them um that just typed in something into Google and they're looking for whatever it is. And that's if that's the type of thing that you sell that solves that problem, we can capture them without them ever coming to your website a lot. Uh but the so that's the distribution side but then the data side is is keeping those lists like how do what's the source of that or this data like what we just talked about this uh these 20 questions having that that nobody no other nobody else has and the the way we use that that comes to the data side of things and those are the two big things. Do you agree with the data and distribution or the in an AI world are the two most powerful things like back on the S SAS tool where I mentioned Helium 10 a keyword research tool for Amazon sellers. Maybe they need to switch um and I know some SAS tools are doing this like open up their API to an MCP and like okay now our new model is not you pay us $99 a month you pay us X amount to access our data because we've been scraping Amazon for 15 years. Claude doesn't have that most likely and we have all these trends, you know, we have all this stuff that you can then layer on to your queries that the LLMs can really do. So that I do you see a shift kind of coming in and in and those kinds of things or what are your thoughts on on all that? >> Yeah, I would I would largely agree. I think distribution and data for sure are the things that any whether you're small, medium, big company like that's going to be the secret sauce that helps you get much more value out of these tools and also yeah build a competitive moat because again it's like if you're just clicking a button and the tool is doing something for you that's amazing but if it's not based on any particular data or context or information you have personally anyone can do that right so then it becomes like okay well maybe you can get it to a bigger audience and that's a valid moat as well and that's Great. But I would say yeah, it one of the pieces of advice I give a lot of people too, especially people that are maybe earlier in their AI journey, it's like everyone's always like, "Oh, what tool should I use or should I use this thing versus that thing?" This is like fine questions, but before you go chasing the next tool or model and you probably want to look at what data have you been able to connect to already because I realize there's a lot of questions at companies around like what data we can and can't use. But you really can unlock wonders if you take even a tool that's not your favorite and you get it connected to HubSpot, to your email, to your calendar, to your documents, to your project management system. Like that alone is a huge unlock and if you're not doing that already you're missing out on a huge bit of value and then yeah to your point thinking about the next phase of this what are the tools you can connect things like MCPs or via API to to start giving to these models what accounts and access and data do you have that is being not used I'd also as kind of a final point here encourage people to have a really wide view of what data is right we're all thinking probably and rightly so like analytics, customer data, all this stuff that's really really valuable that fuels our business. But in the age of AI, everything in language is data too. Comments on your social media are data. Your meeting recordings, your calls, every single word you speak at your company, every framework you have, every way of thinking, all the knowledge locked inside your head, that's data. And that's an advantage if you can articulate it. And kind of like we just talked about, if you can get that into some form where AI can use it and act on it, that's potentially a competitive advantage as well. >> So you've got all these connectors you were just talking about. Let's just keep it simple like workspace. So you've got a profile, you connect to uh all the Google mail, uh Gmail, uh the drive, the calendar, whatever else. >> What about security? Because, you know, we we heard a lot about security at the beginning and then when OpenClaw came out, we heard a ton about security and what to do and how to protect. But what can a person do right now to make sure that your information is secure? >> Well, one, if you're confused or struggling, like so is everybody because this is a mess right now. Um, I am by no means advocating go turn on every bit of data you've got into AI. you need to vet this at your own company. You need to make your own decisions. We still have things turned off or figuring out different angles of how we want to enable this stuff. But like yeah, you have to understand what does each connector actually share with a tool because we've just been looking into this for a few different tools and you know sometimes every single connector can be different in what it decides to share. Whether it can only read your data or write to your systems is pretty or actually take action in your systems is an important point to ask. I think like right now where we've landed and again it's literally company by company dependent is you want to be going for like the enterprisegrade licenses with these major platforms because if you're not on those types of licenses at the very least you have some guarantees there and it's still not perfect and I've seen plenty of IT teams be like there's huge problems with this which is their prerogative. So yeah, I think that going for the highest enterprisegrade account you can with something like Google uh Claude or Chat GBT is the first place I would start. I think if you're considering anything like Open Claw, like super interesting, but good luck because that is a a disaster when it comes to security. I don't think there's like a safe way to do that. There's probably like safer ways to do that, but you cannot especially the whole that's with just connecting systems to something like chat. Once you get into the agentic stuff, the connection might be safe, but what the agent goes and tries to do with your information could be totally unanticipated and like totally in terms of consequences as well. So, I'm just kind of thinking out loud on this, but if you had if you opened up a new Google account and you gave you had a different identity, a different profile and you logged in on that profile. So, Chrome profile, workspace profile and you only gave the access to that Gmail drive uh workspace and then you could only give it partial access to your main profile. So, your banking information or accounting or other things would not be part of the new profile. Is that something that you can set up? Yeah, I believe that's the way a lot of people who are kind of on the forefront of this are setting up these like more test environments or sandboxed environments so that these tools don't have access to all that stuff. But then it gets like chicken in the egg, right? We just talked about data as a huge advantage. Well, what happens? Like I could set up tomorrow probably open qua to be relatively safe with a totally new Google account, but like nothing is in that account. you know, nothing all this stuff that's sensitive is also stuff that's help really valuable for AI to be able to help me with. So, it is a real balancing act. But yeah, what you outlined, I mean, I believe that's how especially for test purposes, how people start um trying to safely set this up, but also good luck. Good luck figuring that out. I mean, I would have to spend a considerable amount of time to figure out exactly the right setup. And even then, it's like, you know, I've heard horror stories already. Not to not to say like avoid stuff like Open Quad, just like we're not creative enough in understanding how badly some of these things can go wrong, right? >> Think of think of this. If you're storing client information, I'm thinking from a legal standpoint, >> you put and you have an in-depth client profile with all of their information, possibly even their banking information like for transfers. If that goes out somehow, >> Yep. >> you are done. >> You're done. It's exist an existential uh event. Yeah. >> Didn't that just happen with some guy where his agents wiped his entire database? >> We just covered this on our podcast that came out this morning. Yeah. Someone who ran like a I think it was like a small or mid-market software company. It was quite successful, like very formal business, not like a side project, I don't think, or anything. Like had employees, had a bunch of clients. he uh had some I think it was cursor the coding agent um had actually and you know there's debate still over like whether he had his data set up right or whatnot but basically it found something it shouldn't have accessed things it shouldn't have and deleted all of his codebase in like 9 seconds and again maybe maybe it was his fault for the way it was set up but like it wasn't supposed to be able to do that in the first place so good luck like trying to antic participate, right? Exactly. When And there's plenty of people working on this and especially if you work at a firm like talk to your IT team, talk to your legal team. Like you're going to have to get these people involved. There's experts out there who can help with this, but it is a really messy messy world right now when it comes to this. So, >> what's the new org chart going to look like in our marketing department with Agentic? Like you said that you're doing it. What what what's it going to look like? Is it going to be one person for every five agents or something instead of one person for every six people? Uh or what what's what what's going to happen, do you think? >> That's a great question. I don't know exactly where we're headed with this, but I would fully anticipate at some point that we are going to be having people acting as orchestrators of agents. So you might still be a marketer who has a particular skill set, but I can almost guarantee you at some point you're going to be eventually managing and orchestrating a team of either only agents or humans plus agents. As a human, you may be working with agents who are your colleagues. So I don't know how I think it gets flatter just because like you're enabled to do so many things that are kind of outside your wheelhouse with AI. Like as one example, this is kind of interesting. This isn't strictly um marketing, but you know, the CEO of the the fintech company or crypto company Coinbase released a a post today on X saying they're unfortunately doing some layoffs and they're reducing by like 14%. One of the reasons for this is AI, but he goes into talking about like we have to revamp how we do things to be AI native. And as one of the things I've got this pulled up right here. He says one of the things we're looking at is basically AI native pods is what he calls it. He says we'll be concentrating around AI native talent who can manage fleets of agents to drive outsized impact. And here's the point that I want to mention kind of getting at what I'm trying to get at. We'll also be experimenting with reduced pod sizes including oneperson teams with engineers, designers, and product managers all in one role. So, that's pretty extreme. We'll see how that works out. But, I wouldn't be surprised if you have someone who's just a really good marketer with really good marketing sense doing five or 10 different specialized jobs in concert with agents because the agents at some point are good enough to help them do these very narrowly specialized domains within marketing that you might have hired for. >> We've got the past. >> We've got a buddy right now, Kev Andrew Wilkinson. He's got a whole company running on 12 agents. >> Yeah. I mean, just I I ran a big event uh my last last uh month in uh in Nashville and just some of the stuff I had to do, spinning up websites, just landing pages for testing things, just uh going through 41 speakers and creating their on their walk-on music and uh creating graphics and a ton of just stuff that I would have had to gone to Upwork before >> and hired hired somebody and paid them, you know. whatever rates that they were. Sometimes >> he doesn't pay well and >> sometimes it's I pay some people really too well. But to be able to do that and then to knock that out, I think it I estimate it's a $100,000 that I probably save just using the tools and weeks and weeks of time of having waiting for someone and back and forth and whatever. And I know enough like like you just said to orchestrate. I have enough skill sets. I'm not just in one thing. I know enough about enough things like where I can guide it. I don't know the details of how to, you know, make this happen, but I'm like I know know what it needs to look like or needs to do and so I'm a able to direct it in that way. It's like a director of a movie or something. You know, they they can't act, but they can tell you, you know, what you need to say or where you need to be >> or the emotion they want they want. >> And that I think that's going to be >> a major skill that people that have that and can use these tools in in that way. Yeah, I think you mentioned something really interesting there too. Not only the money saved, we find ourselves thinking about this a lot as well is the time, right? Is like sure maybe maybe a human could do it for even the same amount of money or something, but the back and forth, the long lead time, the ability. So I don't know if it's like oh AI or AI agents or fleets of agents are going to replace everybody necessarily though they will replace plenty of people I think in this kind of structure but more like you're going to structure these teams that have this like AI native talent that can you need 10 more of you right that are pretty good at all these things and have the vision and the idea of how to orchestrate them you need to go hire those people in spades pay them a good amount of money and then they can go scale up how they're using agents really quickly and flattening the organization right? Or at least flattening the decision-m layers and the lead time on a lot of this stuff. >> Yeah. If if if you're paying people I'm just making math easy. You got five people doing stuff and you're paying them 100 grand a year right now. It's 500,000 payroll. >> Yeah. >> If you get one guy and you pay him 250,000 a year. He's really freaking good. >> Uh and but he's now going to be able to do the thing of 10 people instead of those five. He can do what though, you know, taking 10 people because he knows enough crossover, enough how to guide and produce it. I think that's where the some major opportunity is for for some people that for a lot for for everything. Um that's yeah it's gonna be interesting times to see see where all this goes. There's there's there's no doubt. >> You know what's crazy is I think back to when I was younger much younger in my 20s and >> it's 1800s right? Yeah, it was in the 1800s. Uh I think the uh Spanish American War was going on. That's right. Yeah. Yeah, I remember. I remember. So, uh Teddy and the team and uh I had a promotional company. So, I just want to show you the events that I had to go through compared to nowadays. So, if I had a client, which you know, let's say I already have the client. So I would pack up all these promotional products, t-shirts, jackets, trinkets, like all all these little chachkis, put in a hockey bag, go over to the person, show them, and you know, if they wanted gray or if they wanted white, I have to go back to the office, come up with panone colors, send that over to them. Now, that could either be by courier, which I'd have to send. A lot of these clients were outside of the city, so I'd be going cities over, so it would cost extremely like a ton of money. Then coming up with the logo, you'd use Electraet, which you have to do one letter at a time and scrape it on. And if the font was wrong, you have to go back and do it again. But you get all this information, send it over to them for correct. If they needed, I remember Mercedes-Benz, I was doing something with them, and they wanted something a 16th of an inch difference. had to go redo it, send it back over to them. It was a nightmare and it went on forever. Now, you could use fax machines, but a lot of the hard copy signoffs, they were used, they wanted to see the actual uh mockup itself. And nowadays, if you think of what you can do with all that time, it's just it's it's not even night and day. Like, >> that's a lot of cigars. That's a lot of cigar smoking time. It's a lot of cigar smoking time. It is. >> That's a lot of cigar smoking time. >> But you know what? I got to the the you know, if you look on the positive side, every time I had, you know, a couple hour drive, I was smoking a cigar in the car. >> There is an advantage. >> I I actually I envy you guys, your cigars so much because like I have now so much time where I'll give AI instructions, go turn it loose on something, then switch gears to something else. I'm like, man, I would just like to sit back for a minute or two, puff away, Claude gets done with something, jump back in. You know, they can be be healthier than me just like checking email compulsively. >> What what I do is I turn So, if Claude is doing that for me, I go out to my balcony, play with my dog, smoke a cigar, and during that cigar smoking time, my mind is racing. >> Yeah. Right. >> On ideas. And it's actually it's it's those those ideas that come to you in the shower when you're not distracted and you're just >> and then I'm talking to uh using Whisper or another AI to talk to my phone notes. And I just did this the other night uh Sunday night. And then yesterday I was like, "Nor, I need to talk to you." And I hadn't organized it all yet. So I was like >> I had like 15 different little I I I emailed them to myself to a special email address. That's how I do it. Just my little process. So, I was like, I need to copy and paste these in, summarize it, because I even read them back myself. And but that that's the power of of these tools is it makes you just so much more productive and so much more efficient on on stuff. And >> but where where are you seeing the biggest misuse by marketers? I mean, where where are you seeing not another one of these? And they think and they think it's the best thing since sliced bread. Yeah, I think the misuse I would say is one what we talked about at the top of the episode where people just think, well, isn't it great that I can spin up 20 variations of this email or blog post or whatever? And you're like, well, this is all slop. Like, did you take any time to like just because it can do it doesn't mean you should you should spin up 20 different versions of the email if you've taken the time to train AI on how to actually do that effectively. So I think that's one. I think two is like just broadly it's not misuse but people broadly I think your average marketer is getting kind of stuck where it's like everyone now knows AI is a thing that it's important. They're often at work being mandated to start using it. And now they're starting to experiment to their credit with like lots of different use cases, but it's often still patchy and in silos and still pretty surface level like we discussed to your point before like 99.9% of people would not do the like first principles thinking assistant. And there's such value in sitting down and really building out your own personal knowledge base and way of working and training AI and how to do that as a true thought and strategy partner and advisor that I think you could have been doing years ago but people still haven't done and the models are so much breathtakingly better now. It really should be we really should be evolving much more quickly how we're using these tools beyond just the basics that we've all started to learn. I get it like people are at different areas of their journey, but I think really taking the time to get more sophisticated with how you use AI would benefit a lot of marketers. Hey, Kevin King and Norm Ferrar here. If you've been enjoying this episode of Marketing Misfits, thanks for listening this far. Continue listening. We got some more valuable stuff coming up. Be sure to hit that subscribe button if you're listening to this on your favorite podcast player or if you're watching this on YouTube or Spotify. Make sure you subscribe to our channel because you don't want to miss a single episode of the Marketing Misfits. Have you subscribed yet, Norm? >> Well, this is an old guy alert. Should I subscribe to my own podcast? >> Yeah, but what if you forget to show up one time? It's just me on here. You're not going to know what I say. >> I'll I'll buy you a beard and you can sit in my chair, too. We'll just You can go back and forth with one another. Yikes. But that being said, don't forget to subscribe, share it. Oh, and if you really like this content, somewhere up there there's a banner. Click on it and you'll go to another episode of the Marketing Misfits. >> Make sure you don't miss a single episode because you don't want to be like Norm. So, kind of a followup to that, if you had a single piece of advice for marketers who are just feeling overwhelmed, what could you give them? >> This is going to sound really basic, but not enough people, most people don't do it very well is like you have to have some type of documentation of your use cases. Like this is how you avoid this like, oh, I use it for email sometimes or I do this or I chat with it for that. Great. But like you need to sit down and actually even if it's uncomfortable for you try to map out what work you are doing in a week or what what parts of your business what makes that machine run whether it's your job your function your business doesn't have to be perfect but you need like a map to follow because if you don't have that you're just going to endlessly be chasing stuff and I I do it too like I understand the temptation and there's so much going on but like you have to have something to come back to. We we teach often people how to build out even just basic use case workbooks that are prioritized based on the use cases that have the highest value to you and you have the highest ability to execute on. Even something like that which is super simple whether it's that or some other process SOP set of processes checklist what what have you some type of map they keep coming back to and say oh okay last week I used AI to solve that thing this week it's time to use AI to solve this thing. systematically approaching your work like that, I think would benefit a huge amount of people. >> All right. >> What about mixing systems real quick? Uh, Norm, what about mixing systems? So, if you're a client, >> uh, if if I mean, you're an agency and you're you using a claude and you're mixing clients in there. So, let's go back to this first principles thing. And so, you got these files and one time you're saying, "Okay, use this MD file because we're writing for client X. Now use this MD file because we're writing for client Y. Is it going to get confused or should those all be on obviously they should be in separate chats. Uh but is it going to then kind of pollute the entire system when you have a hundred different ones that have been kind of run through the same model? You know, it's hard to tell because especially with the memory functions these tools are building in where it's kind of uh it's kind of passively remembering things about you and your preferences. That absolutely could mix up a lot of things. If Claude is remembering preferences here for client A, preferences here for client B. I think a big way probably around that is you would be very diligent about structuring projects within quad one for each client or or client specific projects would be my guess that the best way to do that um without you know kind of having been in the agency world for a while you have to give some thought to document knowledge and context management with these tools because you probably if you were running multiple clients I probably wouldn't be just doing it all willy-nilly in the same account without some type of partitioning using projects or something like that. >> Yeah. You you'd have a project, individual instructions, and individual skills even. >> Yeah. >> Yeah. Yeah. >> Yeah. Okay. Now, Kevin, don't interrupt me. Hey, Mike. >> Norm. What uh >> when are you going to send me those those new pictures? Are you going to send me >> two snaps in a Z? >> All right. So, Mike, we're at the top of the hour. And at the top of the hour, we always ask our misfit if they know a misfit. Oh, I do. I do know a misfit in terms of a marketing misfit. Um, I've got a guy who everyone should go check out for sure. >> Awesome. >> Yeah. So, I could share details about that if you want. >> Yeah, let's do that. And then we're going to have Mary reach out and get that information from you. So, who are you recommending? >> So, my good friend uh Jeffrey Klene, his website is gkli n.com. Uh Jeffrey is an awesome educator, speaker, marketer. He helps people use AI across marketing, gives some great talks on it. He's really smart guy, really funny, really personable. I've become really good friends with him over the years. Um, you know, first through work and now just good friends and he's always got something interesting to say. He's at the forefront of all this stuff. >> Perfect. That's great. Now, Kev, you're gonna say because I can read his mind. >> How about your contact information? Is that what you're going to say? >> Uh, very good, Norm. Very good. >> In terms of getting in touch with me, a couple ways that are super useful, too. So, I think the easiest, best way people can just find me, super easy to remember, is just I'm pretty active on LinkedIn. So, Mike Kaput. Um, you'll just see me pop up right away. It's not a very common last name. you shouldn't have to dig for it too much. Um, that's a great way. Feel free in any context like I accept connections or whatever. Don't worry about it. Like feel free to reach out, connect with me, shoot me a message if you want to chat about anything. Happy to see however I can help if you have any questions about AI. Also, our website smarterx.ai. So, the word smarter, the letter x.ai, that is where you can find out everything we do in terms of AI education, both for marketers and non-marketers. Uh we help all sorts of companies and functions now learn and adopt AI and you can check out what we're doing these days over there. >> Okay. And what about your podcast? >> Well yes, our podcast is the artificial intelligence show. So myself and our CEO Paul Ritzer every week break down what's going on in the world of AI, not just for marketing and business, though that's a fair amount of it, but just in general if you're having trouble drinking from the fire hose, we do it for you every week a little bit. So yeah, you can go check that out. You can see that uh on any of your podcasting platforms of choice. We'll pop right up. >> Awesome. >> Very good, >> Mike. I really appreciate you coming on. This has been fun uh geeking out on on a few things here. Uh this has been great. >> Likewise. Thank you guys for having me. Really appreciate it. >> Hey, you're welcome. It's been awesome. See you later, Mike. >> And Norm, you can't go to Smarter X because that's only for smart people. So just that he didn't say that, but I'm saying that. So just >> Hey, Kevin. scratch out on your notepad where you just wrote smarterx.ai. Just scratch that out because it's always smart people. >> I'll do that. And I just want to know one thing before we go. Does your voice get on your nerves, too? >> All right. >> Thanks, Mark. >> See you later, Mike. >> Thanks, guys. >> Uh, this world of AI is changing so much, man. Um, >> really, it's it's crazy to stay on top of it. I think his his suggestion of uh you know make a make a plan of what what are you doing? What are you trying to achieve and how does AI what can AI do to make that better for you rather than just going chasing shiny objects and going oh that's cool that's cool that doesn't >> you're not going to get anywhere >> you're not going to get anywhere. Uh, and that's kind of I haven't done written it down like he said, but that's kind of the approach I take is like I see stuff all the time. People, you know, in our space like, oh, look at this cool this this design tool. Oh, you got to go play with uh the new uh open AI uh uh version of cloud code, what they call it. I'm drawing a blank right now. Um whatever uh codeex codeex um and I was like, yeah, but when I have a use for it. Um, so that's and I don't know how some of these people that are playing with all these things like when do you get anything done? It's like all you're doing is playing. Uh, but I think that that was a good point. Um, but you know, Dragonfish, we're using a lot of AI and going to be building a lot of stuff in uh, as well. So, uh, it's exciting times for sure. >> Yeah. And, uh, just uh, wanted to ask you something. What are you going to do when Kelsey and I are in Miami this weekend? uh collecting all those cigars in having all those cigars in these different lounges. We'll be thinking of you. I just wanted to let you know that. >> Yeah. Well, I'm going to I'm going to be like I tried like I told you I tried to change my flight, but uh I'm on the way to England. So, Friday night. So, your events is on a Friday. We're recording this on a on a Tuesday. >> So, your event is on a Friday night. I'll be ahead of you actually uh in a lounge in London, England. and I'm going to go to one of the little fufu little gentleman's lounges or probably a couple of them and check them out uh Friday night in London. So, I'll be having a smoke. So, we'll have to exchange photos. >> Okay, we'll we'll do that. >> We'll exchange photos just >> this time. Keep your clothes on. >> Nothing below the weight. Nothing below the waist. Okay, >> that's not a cigar. >> All right. >> All right. So, how do people get a hold of us? If you want to if you want to hear more of the the us uh braiding each other um and uh sharing with some guests, you can uh check us out at marketingmisfits.co.c not.com but.co or you can check out the newsletter misfits.news. And then uh if you want to look at our our my pretty face and Norm's beard, um how do you do that, Norm? >> All you have to do is go over to YouTube and check out uh Marketing Misfits podcast. Or if you want the golden nuggets, these are the three minutes and under clips, you can go again over to YouTube and check out marketingmisfits clips. That's right. So remember misfits.news. Get on the newsletter cuz we take these podcasts. Sometimes you you can't listen to all of them. You know, they they run typically about an hour each, but we we actually summarize them. We we mix and match different episodes together. Uh it's a really good uh really good newsletter. It's totally free. Comes out every Wednesday. So misfits.news. We just started it uh about 8 weeks ago. Uh so hopefully you'll go and sign up there and look for an announcement coming soon on something else that uh we're going to be doing. Really fun uh fun thing. But we'll tell you about that in the next episode. So uh Norm, I guess I'll see you next week, right? >> All right. We'll see you later. >> All right. >> Have a good day. Bye.