The E-commerce Leader

[00:00:00] JM: And so I literally recreated the book project, which was a summation of 67 AI tools had perplexity do all of the descriptions. Plus chapter and section descriptions
[00:00:14]
[00:00:51]
[00:01:13] MV: So Jason, we’re back on with another podcast about AI. It’s the gift that keeps on giving. You’ve written a new book for Shopify site owners called AI Marketing Power Tools. This is like a fun combination of topics, very obvious combination to do. You’re a writer and you’re also a Shopify advisor and store owner, and you’re giving it away for free as part of a traffic stacking challenge.
[00:01:34] So there’s lots to unpack there. Tell us more about what’s going on.
[00:01:38] JM: Yeah, yeah, yeah. So, um, really excited about this new book that we’ve written. I wrote it in a week. Uh, using, um, me as sort of the conductor and director of the idea, a writing assistant, uh, named Cassie, who was my writing assistant. And then we used a tool called Perplexity.
[00:01:55] ai that wrote a lot of the descriptions. And the basic premise of the book is very simple. We, um, created a list of, um, AI marketing related tools. That are literally game changing. And there are 67, I believe, uh, at last count in the book. And this is a version one of this book. We might end up with it being a much bigger book in the future, but the 67 that we list are remarkable and amazing and mind blowing.
[00:02:22] And, um, the fact that we wrote the book using one of the AI tools, I’m happy to talk about and explain the story of, and all that, but, um, it came together so quickly and nice, nicely. That we’re really excited about it. And yeah, we’re giving it away. We’re also giving away another book that is similar in nature.
[00:02:40] It’s called the, um, the, uh, Shopify apps, uh, book. And that book is basically a summary of the Shopify apps that our CEO small group really rave about. And so this book is AI focused. That book is app focused. All of it’s really intended to help Shopify store owners. Uh, improve their traffic, uh, to their sites and, uh, the challenge is all geared towards that.
[00:03:04] MV: Wow. Um, so lots going on indeed. All right. So, um, let’s dig into a little bit of this then. I mean, first of all, uh, why did you decide to write this book? I mean, I know you. You’re an author, you’re a writer. So when you see something, you write about it, but why this book and why the apps book, and why is it coming across in the challenge format?
[00:03:22] It’s free gifts. I was saying to my wife, um, wow, there’s a lot of bonuses here. I’d be charging for this, but you’ll never get away for free. So what’s, what’s going on here?
[00:03:30] JM: Well, I’ll answer the first question, which was why I decided to write the book. Um, it was probably, well, it was December when chat GPT 3.
[00:03:37] 5 came out. Right. And, uh, at that time. Um, I started basically just in my on my iPad, making a list of AI tools that I would hear about. So, you know, all these articles are just the Internet flooded with AI related articles and listicles and, you know, this kind of thing. And so I just started making a list of them, uh, in my iPad.
[00:03:58] And, uh, it just continued to grow and I explored them more and more just individually, and then I thought, of course, this thing could really make a great book and we had this traffic challenge coming up and I thought, you know what, I’m going to write this book and I’m going to give it away as a free, a bonus, and I’m going to use, um, in my mind at the time I said, I’m going to use chat GPT to help me write the book.
[00:04:21] And so I, we announced it, we put it out there. We started telling people, Hey, there’s an AI marketing power tools, but coming out, you know, and, um, this was all like 10 days ago or 14 days ago. And, uh, so I thought, well, I’m just going to use chat GPT to write the bulk of the content. And on day one, that proved incredibly disastrous.
[00:04:42] And did not work and I was in the valley of the shadow of death for two or three days and then I found perplexity and um, so this is a very interesting little story here. The reason, uh, chat GPT did not work to write, help me write this book was because, um, it lies all the time. And, um, it just makes stuff up.
[00:05:05] It fibs, I like to say. And so, uh, and, and I found, I didn’t know it was fibbing at first because I don’t know these tools very well, right? Like these new, these tools are new. So if you ask chat GPT, Hey, what does this tool do? And why is it beneficial for Shopify store owners? It tells you something that sounds plausible.
[00:05:24] And so, but then here’s the, here’s the little story about how I. Deciphered the, the fibs. There’s a tool that’s very cool for podcasters, um, and it’s called, uh, auto, uh, pod fm. And so I asked chat g PT to summarize what auto podd.fm is and why it’s useful in some of the key, uh, you know, attributes or benefits.
[00:05:52] And it said to me, and I, I was looking at auto podd FM’s website, like I could see what it does. It’s for podcasters. It’s an automation tool that does editing, splicing, you know, all the stuff that you’d be familiar with, Michael and chat GPT’s description was autopod. fm, a revolutionary podcast platform that caters specifically to auto enthusiasts.
[00:06:18] Uh, discover the various facets of the automotive world from racing and restoration to the latest innovations and industry news. Now that description has nothing to do with what autopod. fm actually does. And I thought to myself, Oh, I just caught chat GPT and a huge lie. It made a plausible sounding lie.
[00:06:44] And of course, I’m trying to write a book with 67 descriptions of AI tools. And I quickly realized. I’m screwed. Like, this thing can’t be my writing partner, you know.
[00:06:56] MV: I’ve got my own ChatGPT lying story because I was basically kind of trying to train it by browbeating it the other day when my teacher instincts kicked in and basically I, I wanted to, and I’ve also been out writing some books using ChatGPT and then kind of editing it heavily and You know, so I’ve got two or three books that I’m quite keen to, to get writing, you know, I’ve been selling the scratch page for ages, as you know, but, um, one of them was going to be based on a couple of books.
[00:07:20] I didn’t want to plagiarize them in any way, but I wanted to weave in some of the content to an initial outline that I would then edit heavily. But one of them was based on 80, 20 sales marketing by Perry Marshall. And I said, I said, actually, instead of writing a summary of it, could you please write the actual chapter outliner?
[00:07:36] You just give me the chapter, you know, content. And it came up with something which is completely made up. And I said, no, that’s simply made up. It’s right here in front of me. This is just wrong. Please. Can you tell me the exact names of the titles? And then it just made something up. And I said to it, I gave it firm instructions to never lie to me and make stuff up unless I gave it specific permission to you, but I don’t think it’s
[00:07:55] JM: going to work built to lie built to create.
[00:08:03] Narrative. That’s the right answer. It is not built to answer a truthful set of questions and it’s, and it can answer narrative questions beautifully strung together, but could be completely and totally fabricated. And so that’s what I discovered. And then when I discovered perplexity. ai, what I quickly learned was they positioned themselves as a more accurate and authoritative.
[00:08:31] Uh, you know, uh, answer engine because they have multiple vectors of data that confirm for it, that what it’s saying to you is the truth. And that’s very interesting. So here’s, here’s perplexities description of autopod. fm autopod. fm is an AI powered podcast. editing tool that uses machine learning to automatically edit and enhance audio recordings.
[00:08:56] It can remove background noise, adjust volume levels, and even add music and sound effects to create a polished final product. This tool is designed to save time and effort for podcasters who want to focus on creating content rather than [00:09:10] spending hours editing their recordings. Autopod. fm uses natural language processing and other AI techniques to understand the content of the podcast and make intelligent editing decisions.
[00:09:22] It can also transcribe the audio and generate captions, making the podcast more accessible to a wider audience. Overall, autopod. fm is a powerful tool that streamlines the podcast editing process and helps podcasters create high quality content with minimal effort. And when I read that description from perplexity, I was like, If I’m a manager interviewing two writing assistants and one gives me the chat GPT response and one gives me the perplexity response, I’m hiring perplexity every day all day long.
[00:09:58] And so I literally recreated the book project, which was a summation of 67 AI tools had perplexity do all of the descriptions. Plus chapter and section descriptions and then Cassie of course is tuning it all up putting it through Grammarly Giving me feedback. She’s a real person And and I was sort of the maestro, you know, the conductor I wrote the introduction and the conclusion and I read some of the the apps that just you know The AI tool descriptions and thought now that’s sort of confusing her That one doesn’t merit being included in the book.
[00:10:36] So I was sort of the executive producer, I guess you could say, uh, of this book and it’s a good book. I mean, this is well written, um, and I’m blown away. I’m giving all credit to perplexity. I, you know, I’m, I’m acknowledging them. I will also just say this, um, I instantly took what perplexity gave me and put it through another, what I consider AI tool, Grammarly, and ask Grammarly to check it for plagiarism.
[00:11:03] And it is not a plagiarizer. In other words, it’s creating original written content based on a summation of the wisdom it finds on the internet. So it is a, a, a creator of this content, if you will. Therefore we don’t cite our sources in this book. It just is, it’s matters of statement and fact, um, that, uh, perplexity is writing.
[00:11:29] And, um, so if you want a copy of this to see how this book came together, Michael, you’ve got a copy in front of you. You should see the table of contents. This is a robust book. Um, and so anyway, that’s the way it came together.
[00:11:39] MV: Wow. That’s very cool. Um, I have to say the description as a podcaster, you know, working hard and assistant to do all the things that as soon as you described the podcast app, I’m like, wow, that’s amazing.
[00:11:49] And it just goes to show what a total disaster chat GPT had made of it. Cause I, I’m not. That interested in auto parts or whatever it was trying to sell me on. It’s just, wow. But that’s interesting, isn’t it? I mean, so what you’re implying about the role of the, there’s so many things to unpick it, but one simple thing is the role of the quotes writer is, I guess you’re transitioning more into a publisher, traditional publisher slash editor, not in the sense of just, you know, correcting grammar so much, that’s more of an AI tool doing that for you, but the sense of having a concept, this would make a good book.
[00:12:19] That, that instinct that you have, um, is partly what you’re bringing, isn’t it? And then, you know, Deciding on the focus and what’s helpful to human beings to use and then project managing it and then at this editing and for sense and for usefulness, uh, which is quite interesting. So you’re still bringing value just in a different way.
[00:12:36] Yeah. Yeah.
[00:12:37] JM: That’s the way to look at it. I think it’s almost like a movie producer, um, or, you know, like a director. The, the thing that’s interesting is if you go to some of these sites, like if you personally went to autopod. fm and I ask you to write a 200 word description of what that thing does. It is not easy to do like it, you know, go to any of these tools.
[00:12:57] I mean, I mentioned 67 and I’m looking at the list right now and you, you look at their, their homepage because, you know, if you look at a homepage of a website, what they’ll use is very, very concise, punchy, little bits of copy that they think express a narrative or touch on a felt need. And then they try to explain who they are, what they do, but they do that on a different page.
[00:13:23] It is not easy to do a 200 word summary of a site like ChatFuel, AgentGPT, AutoGPT, uh, Precept. ai, on and on. I mean, I could list them all. Um, to do that would be, be literally hours and hours of scratching and pecking around and looking at Wikipedia and looking at about us pages and, and all of that. And Perplexity literally does it in real time, real accurate.
[00:13:50] And, uh, yeah, I just, it’s, it’s mind blowing. It really is mind blowing. Wow. Sounds
[00:13:54] MV: good. Uh, and I guess it’s horses for courses, isn’t it? I mean, chat GPT had a certain kind of. Charisma is not the right word because it’s got a rather clunky, repetitive style, even chatty for, but it’s, it’s got a fluency and a plausibility that is incredibly high, which actually like a, a, a bad hire could be really dangerous actually, because you might realize we can deliver a Jaguar or, you know, or a Rolls
[00:14:21] Royce, you know, engine. And it’s utterly made up. It just sounds like it sounds like a character from a movie, doesn’t it? So there you go. It’s, uh, so having, having discussed the limitations, obviously you got 67 tools. I would guess everyone should just get the book. I mean, just simple question. Is the book available separately from the traffic stacking event?
[00:14:39] Cause it does sound really worth having. Okay.
[00:14:42]
[00:15:30] JM: You gotta go through my, you gotta go through my three day challenge and get it as a bonus. Now I might put it out on Kindle. I might take another whack at it and do a second edition. That’ll take us like, you know, three days or something, but, um, but for now, no, you just get it at traffic stacking.
[00:15:45] com. It’s a challenge and it’s a free bonus. The, and I think the phrase you use the horses for courses phrase is right. I like chat GPT a lot. It was sort of, and it sounds. It’s so weird and creepy to say, but sort of my first love of AI tools, right? In December, we were all like, this thing’s amazing. You weren’t
[00:16:05] MV: blown away with it.
[00:16:06] I’ve got to say, yeah. So maybe the, the love affair has turned slightly. Yeah.
[00:16:11] JM: But what you realize is the, the. technical competency of the thing they made has a specific use case. And what everyone’s doing is what you just described trying to do yourself this morning, make it do stuff it wasn’t built to do.
[00:16:27] And you don’t really quite realize it wasn’t built to do it because it’s so darn verbally fluent. And so if you’re, if you’re hyper verbal. You can kind of fib your way through a lot of answers to a lot of stuff. And I do fear to be completely blunt, that marketers are using chat GPT wrongly to accomplish projects and activities that they think it can do.
[00:16:55] And it’s so verbally fluent that it’s, it’s, it’s like fibbing to them and they don’t even know. And then they’re passing that on to customers or, you know, information products or whatever. So, um, it is horses for courses. Chat. GBT is super cool for, I would call it creative narratives. You know, if you ask it to make a sales letter for you on a specific topic, it’ll do that flawlessly.
[00:17:20] If you ask it to make a, uh, a, uh, an email inviting someone to an event, it’ll do that flawlessly. If you ask it to describe something that is a real thing in the real world, you gotta look for the fibbing. You know, like in, in your example, you just ask it to describe a real, um, a real book from Perry Marshall.
[00:17:39] It, the other thing that’s really well known now is it can’t even do math. Like, you know, if you ask it to some, like a column of numbers or whatever, it’s like really bad at math. Um, so you can’t, you know, you got to factor that in. Um, so these are limitations, limitations. We
[00:17:56] MV: all got seduced by, um, chat TPT because it’s in the Example of, you know, as you said yourself, the phrase Cambrian explosion brings to mind that suddenly there’s such a, uh, incredible spread of tools suddenly, and that are of a different, a step order, a change order, what’s the word step change order of magnitude better than anything that’s come before.
[00:18:16] And so we can lump it all in with chat, GPT, because they’re, they’re really great job [00:18:20] of somehow becoming the buzz thing. And now I guess. You know, instead of one tool that’s amazing at certain things and very flawed at many, we’ve got 67. So talking of which, um, obviously 67 would be insane, but can you give us a few of the tools that really stand out to you?
[00:18:34] Perplexity clearly is one of them. You sound like you’ve got a new love now. Um, what are some other ones that you would recommend?
[00:18:41] JM: Well, yeah, perplexity is one to check out for sure and to understand why it’s useful, um, in the context in which we’ve described it. So that’s that’s an obvious one. We both have really begun using, I think, read dot a I, which is a wonderful, absolutely incredible note taking tool for zoom meetings and Kyle, my business partner found it, plugged it into our meetings and You know, if you’re a professional consultant or coach, and I’ll just describe like how Kyle and I work, generally speaking, when we’re working with a client on, uh, their traffic and conversion issues for Shopify or whatever it might be, Amazon.
[00:19:18] Um, uh, because I guess, because I did it for a year or so, and then I asked him to start doing it. He is the recorder of the notes and action items and the minute taker, you know, not minutes, but you know what I mean? Like any key links or topics or. You know, to do all the action items. He would do nice notes after every coaching call and send those to the client with a link to the zoom, uh, you know, meeting itself for replay.
[00:19:42] And then one day I logged in and read. ai popped out and he said, Hey, this is a new AI note taking tool we’re going to try out. And at the end of the meeting, that thing sent out a set of notes that you would think. You were like, you know, the president of the United States or like you were at the UN or something like that.
[00:20:01] Like this thing takes notes. Like, you know, you’re, I don’t know what the, like the, like the admin assistant for, I don’t know who, like Elon Musk or something like that, although he doesn’t like meetings, but, um, I’ve got
[00:20:13] MV: one little story about read. Oh,
[00:20:14] JM: come on. It’s amazing.
[00:20:16] MV: Yeah, it’s good. It’s good though. I do like it.
[00:20:18] I’ve used it pretty heavily since Kyle recommended it. Cause obviously, you know, you guys are sharp people and make great reference recommendations. I did have an entertaining. Meeting that the other day. So, um, so it doesn’t always get everything and I said something to some guy that I’m working with to get some LinkedIn leads generated for amazing FBA to, you know, help more people in the 10k collective mastermind.
[00:20:38] And, um, I said, he said something like he said, yeah, read that. I’m not sure it works in my accent because he’s from the North of England. And, um, I said, yeah, your accent reminds me of a coach I used to have. Uh, he lived in Sheffield, which is just, you know, in South Yorkshire, in the north of England. And the notes came back and said something like, uh, Michael agreed that Danny sounded like a coach driver from Sheffield.
[00:20:57] And the coach driver’s like, basically like a bus driver. I thought this is a massive misrepresentation. So it can get it wrong. I mean, don’t get me wrong, but obviously an assistant for the Philippines could be getting it more wrong. So, you know, but it’s not always, I would say like anything else, I think we need to not turn our brains off.
[00:21:15] There are two dangers that strikes me that reminds me of earlier conversations, which is when you hire somebody and you just go get it done, and then you assume they’re doing every single thing you’ve ever said ever, they understand everything about what you just said in terms of the instructions, they understand English flawlessly, they understand your industry flawlessly, and the vagaries of your business, these are all pretty silly assumptions.
[00:21:36] And I think it’s very easy to fall into that with AI. You assume, Oh, it’s a machine, therefore it doesn’t make mistakes. But of course there are. Many, many, many levels of decisions it’s made by algorithm rather than by brain. But so I guess it comes to the same thing, which is you’ve got to manage the tools just as you have to manage people, right?
[00:21:53] JM: That’s right. Absolutely right. Let me, let me mention a few others that might be particularly interesting to you. Um, there’s a, a tool called crystal nose dot com and it is a power tool and, um, it. basically allows you to plug in anyone’s LinkedIn or Facebook profile, like their online profile, however they’re on the internet.
[00:22:19] Um, and it will apply personality profile to them before you’re meeting with them and give you talking points. And like it uses like the disc, uh, you know, profile tool and it’ll give you customized things to say and things to not say That would trigger that type of person’s personality. Um, it’s literally like having a prep session before, you know, you meet with someone, um, and it is.
[00:22:47] Basically giving you an unfair advantage, uh, in, uh, preparatory work for meetings. And, you know, if you’re in a really, really high stakes meeting, like if you’re, you know, if you’re negotiating, you know, a big deal, that is the kind of preparatory stuff that you do, you know, who you’re talking to ahead of time.
[00:23:04] And as a person who sells consulting or sells services, you generally try to look at people’s. And, you know, like how we work, we have an application form. People fill out a lot of information before we meet with them. And the goal of that information is so that we understand how best to talk to them and how to customize the conversation to fit their needs and, you know, the situation.
[00:23:28] And this tool is, uh, interesting in how it helps you do that. So that’s just one, uh, example. I thought it was interesting. Um, and we’re going to use it more and more in our In our practice going forward, uh, i’ll mention another one and let you chime in on these but I don’t know if you’ve ever used headline studio before Yes, i’ve used it quite a lot.
[00:23:46] Yeah. Okay. So there you go. So, you know this one So I had not used it before and it was on the list and the uh, the use for it is Really really fascinating. It’s basically a tool that allows you to tune up your proposed headlines to an optimal As much, you know, optimization as possible, uh, against, um, what it, what, what the AI behind it knows is engaging, uh, uh, you know, powerfully constructed, uh, you know, headline content with the word styles that are known to be powerful, the emotional, uh, you know, phrasing, and it will score your headline and you can rewrite, rewrite, rewrite, rewrite, rewrite, until you get something that is.
[00:24:31] Really, really dialed in. Um, I liked that one a lot. I’d never used it before. So, yeah,
[00:24:36] MV: it’s really good. Well, you know, if, if anyone’s listening, they want to see an example of it, where they just need to go to the e commerce leader. com because as long as, um, our editors, uh, and writers, so, um, junior and, and, uh, following our, uh, systems or the SOP, um, we using chat GPT, maybe we need to change the perplexity, but basically say, you know, um, do some keyword research, we pick a keyword and then.
[00:24:57] And this is a classic use of a I with a human to do content marketing. But neither of us are doing it right. So it’s an assistant plus a I cooking something up between them and then using chat to generate a bunch of different headlines and then pop a few of them into the headline of the mate optimizer.
[00:25:12] It’s not the absolutely optimized process, but it’s an example of. How you can use tools together to, you know, optimize the output of one. And indeed, I think that people make a funny thing about AI replacing humans, but humans have been using tools for physical manufacturing since, well, you know, it started in the UK earlier than anywhere else.
[00:25:30] If you’re using a lathe to, you know, to turn a piece of metal and then another tool to smooth it off, it doesn’t mean human ingenuity isn’t involved or judgment isn’t involved. I think I’m constructing a process and having a clear judgment of whether it’s ready or not. And good. I think of the things that are still left for us.
[00:25:47] Anyway, it was just struck me describing that process. Anyway, you’re going to give me some more tools. We’ll give the listeners more tools. So I don’t want to. Sure.
[00:25:54] [00:27:30]
[00:28:02]
[00:28:51]