← The Copywriter Club Podcast
The Copywriter Club Podcast
TCC Podcast #445: The Proper Place for A.I. Writing Tools with Petter Magnusson
00:54:41
In a world where A.I. can write all of the words for free, what is a copywriter to do? What tools should they adopt and how should they approach artificial intelligence? I invited Petter Magnusson, the creator of PurposeWrite,
Loading summary
Transcript63 lines
- [00:00]
Rob
It's been three years since ChatGPT launched and changed the world. So what does AI mean for copywriting today? This is the Copywriter Club podcast. When AI released the first version of ChatGPT to the public, while what had been a quiet conversation about artificial intelligence happening mostly behind the scenes, suddenly burst into our collective consciousness, writers and creators in particular were suddenly aware that at least at some level, these large language models could do some of the work that we were being hired to do. Many of us dove into these tools just to see what they could do, and we launched a short lived podcast that talked about how AI was impacting so many different ways of working, certainly within marketing, but also in many other industries. And you can find plus episodes of that podcast on the Copywriter Club website. Since that time, the dust has settled a bit. The AI tools have gotten a bit better. Image creation tools in particular are significantly better. Writing tools have also improved. But it remains true today that the best copywriters seem to be able to use them to get the best outputs. If you want good copy copy that captures attention and converts readers into buyers, it helps to have a good copywriter guide the inputs and rework the outputs that you get from the AI model of your choice. Another thing that we've seen over the past couple of years is that while tools like Chat, GPT and CLAUDE get most of the headlines, lots of other tools have added components of artificial intelligence to improve their products, speed up usage and make applications more sticky. That is so that users like you and I won't leave. At the same time, we've seen the launch of job specific AI tools that do one thing like writing emails or writing articles at speeds that humans simply can't match. So when it comes to AI, where are we headed? What tools will we be using to get better results? And how helpful is it to have a user or a prompt engineer or a copywriter who really knows what they're doing versus just playing around to see what we can get a model to do. I asked Petter Manguson, the developer of purposeright, to talk a bit about the AI tool he's built and also the broader environment of artificial intelligence and where he sees sees us going from here. And because whatever happens with AI will have a big impact on copywriters, this is a topic that I may come back to a couple more times in the coming weeks. This whole industry is fascinating and the speed of change is a little bit scary. During our conversation, I had a realization. In the past, copywriters charged for the things that we delivered the the words. Officially, we sold blog posts or sales pages or emails or some other copy, but it was the words that clients expected to get from us. But now the ChatGPT can produce the words pret much for free. We need to move up the value chain and sell things like the process or the strategy, the analysis and the ideas, and bringing that to the AI model you use will make the outputs there so much better. Anyway, I think you're going to like this discussion before we get to my interview with Petter. This episode is brought to you by the Copywriter Underground. You've heard me talk about it on every episode, literally for the last year, maybe even longer. I talk about it every week. The Underground includes more than 70 different workshops and accompanying playbooks to help you gain the skills and strategies that you need to build your business. The playbooks make it really easy for you to find solutions to the challenges that you face in your business. Everything from finding clients to conducting sales calls, to using AI, building authority on LinkedIn or YouTube or Pinterest or dozens of other workshops. You also get dozens of templates, including a legal agreement that you can use with your clients, monthly coaching, regular copy and funnel critiques, and a lot more. And since this episode is on AI, you also get a couple of custom GPTs that I've built and included inside the Copywriter Underground, as well as some AI specific trainings. If you're just trying to figure out how to use this or how to do things like build your own custom GPTs, you can learn more about the underground by visiting TheCopyWriterClub.com TCU and now my interview with Petter Magnussen. Petter, welcome to the podcast. I am really interested in your journey. How did you go from I think photographer, content creator, and now you founded an AI company all about writing? Like how did you get here?
- [04:43]
Petter Magnussen
First of all, thanks for having me. I've been like to be honest, I have discovered the POD fairly recently, but I have listened in and I really like what you do. So I'm gonna be stuck in here for a long time listening. There's a bunch of episodes, so yeah, it's so many I want to listen to. So I'm surely going to do that. Yeah, so I have a bit of a weird mixed history. So I started out as a programmer actually a long time ago, and then I started studying physics and I studied at like engineer physics engineering. Then I went to Japan for my work in advanced laser physics kind of thing. And then I stumbled into sales for that laser company doing sales. And that led me into marketing, where I became marketing manager for an IT company in Sweden. And we did extremely well. We happened to sell modems at the time that Internet exploded. So that was my boss. He still thinks that I had magic hands or something because he became super rich from that and he still thinks I had a part of it. And maybe I did, but. Yeah. So anyway, so I did that. And then I started getting really interested in photography. And I was having reached some of my life goals, to be honest. At the early 30s, I thought, I want to change direction. So then I just bailed out. I applied for art school in Norway and I got in. I took a bachelor in art photography and then I went to do master's degree also in Sweden, in photo and film. So then I was like a photo artist, actually exhibiting in galleries and stuff like that. Classic art, you know, career. Until I got a little bit angry with that world in a way, because it's not as. It's not as free as you may think. You know, I used to think that art is free, but in the art world, to survive, you have to be fairly political. You have to know certain people, you have to network, and you need to make pieces about the right topics if you want to get the scholarships and the exhibitions and, you know, that might be all fine. But then I saw how people are. Were adapting to that. And that, like, goes against everything that art is for me. So that kind of got me. Yeah, like. And then I thought, well, I might as well go commercial. So. So. So I did that. So I went into commercial photography and filmmaking. And now I have a small production company in Stockholm and we're doing corporate stuff mostly, so a lot of B2B topics. And that's when I started to see what. What kind of led into purpose. Right. I have a lot of. I see exactly the same. I saw the same happening in visual content as in copywriting. I think that people come to me and they were like, hey, video is hyped. We want to make a film or something. And I'm like, okay, great. Why do you want to make a film? And they're like, it's hip or something, you know, it's like. And I'm like, okay, who's going to see this film? I don't know, everyone. And I'm like, okay, right. You know, So I had to start, you know, the journey with them to like, okay, let's find out if you actually should make a film and who should watch. What is your target audience? And, you know, pain points and interests and stuff like that. So that was kind of a struggle sometimes to make people understand that, yeah, of course I can just make a film for you, but that will make. Not make you happy or the viewers or anyone. So. And then we started producing some text content too, and especially for ourselves. And I, I think the trigger point came because I was trying to hire a guy that was not very good at writing, to be honest. Oh, maybe you should edit this out in case he listens today. But anyway, yeah, I came across a guy that was gonna do some, some writing for us and then I saw the same pattern in text and I think that's what triggered me, like, because he would produce content pieces that had no direction, no purpose, no, you know, not thinking about who's going to read this and why are we writing it and why are they going to read this. Everything like that. And that's got me started that, okay, this is, this is exactly the same problem. But, and, and then at that time, AI came along, you know, chat GPT and everything. So I started playing with that and put that tool to work by kind of turning the process around. Because when you normally, when you use ChatGPT, you write a prompt and that will do something for you. Yeah, and I could see the same problem there, like, you know, generic content just exploding on LinkedIn and platforms like that. I'm mainly talking content now because that's the area that I am familiar with. But yeah, you know, people are mindlessly prompting stuff and they get something that looks pretty good. I mean, AI writes pretty good, layout is nice with nice headlines and stuff. So, yeah, on the surface all is fine, but it's really horrible. It's got no, no value, you know.
- [10:33]
Rob
Yeah, it lacks so much. I mean, just from an emotional connection standpoint.
- [10:39]
Petter Magnussen
Yeah, yeah, exactly that. And there's so many problems with it, you know, like the emotional standpoint, the story, the style, and again, having a purpose with it. Like, who are you writing for? Everything like that. And the thing is, I discovered that you can actually, I have AI, make pretty good text. But again, and this leads me back now I'm like floating around in my thinking a bit, but this leads me back to the visual area because I saw the same. Exactly. Thing happening there because I'm coming from a visual side of things and I saw the same thing happening there first before I came to the tech side, like imaging, you know, AI imaging first you saw it and you thought, wow, this is so cool, it looks amazing. And, and, and then I'M trying to use it for professional purposes and then I see, oh, it's not so easy to direct this thing. I, I can make something cool, but when I do have a brief and I do have a brand and an image and a style and a mood board, and it's extremely hard to make it do what I want because I have a vision, you know, for what I want.
- [11:54]
Rob
Right?
- [11:54]
Petter Magnussen
Yeah, a vision for the result. And that's when it's like tricky and I started realizing that. Huh. All right. Yeah, you need to actually feed this pretty much information to get what you want out of. Is possible though, you know, if you give enough information, you can get pretty good results. And that's what I discovered with text 2 that if you have a copywriter mindset. So, you know, I mean, first of all, you need to know what kind of information you should put in your prompt. If you ask it to just like generate a LinkedIn article about like in my case then selecting a corporate photographer, which could be a topic on our blog, you know, like one of those value bringing articles that are not selling us as a company, but like seemingly neutral guide kind of thing, then that will be totally generic. Not helping the customer and not helping us in any way. But yeah, if you start putting in like the good stuff, like who are we writing for? What is actually important to think about? But then you are in a situation where you're almost writing at your. So then the question is, what's the, what's the purpose here? Yeah, how did you go from, how.
- [13:13]
Rob
Did you go from there to wanting to build your own tool that's going to do this? Because I mean, it's one thing for a copywriter to say, okay, I'm going to, you know, do the back and forth with Claude or Chat GPT and I'm going to tweak it or maybe even set up a custom GPT at this point in order to feed it some, some preliminary data so that it's giving me, you know, a particular voice or whatever. But going from that to, well, I'm going to actually build a different tool that's going to do all of this stuff in a different way. That's a whole other jump.
- [13:44]
Petter Magnussen
Exactly. Yeah. And I started out exactly like what you say now. I started building larger prompts, longer prompts, more complicated, and then into custom GPTs. But then I started running into problems because when the prompts got longer, like when I started getting to the really good ones, when I actually specify the tone and the target audience audience in detail and Everything like that. I started having problems with the AI. Cannot, can then not any longer keep track of what it's supposed to do. It's losing stuff when it, when the prompt gets too long, it's hard to. It doesn't know where to focus, kind of. And that's when I started thinking, wow, I would actually need. And then I know that a lot of people make then several custom GPTs, like one for this, one to find the target audience and one to do this and that, that. And then I thought, huh, I started doing the same. And then I thought, what if I would have a tool that can tie these things together so I can actually do these things step by step. And that is kind of countering the limitations of the, of the prompting. So. And then I came up with the idea of making like a traditional. Since I'm coming from a programming background, I started thinking a traditional program would actually be pretty good here. Step by step, first do this, then ask the user about this information, then go to the LLM and ask that to provide something back. So that's how it started. I started making a simple scripting language that can call ChatGPT and Claude. And that is actually all that purposerite is, in a way, it's a scripting language and it's. Yeah, it's actually even open. So users can also make their own custom scripts if they want. So it's kind of custom GPT on steroids, I say. And the thing, what happens there is that then you can kind of decide now we're getting. I don't know how nerdy you want to get here.
- [15:53]
Rob
Well, I'm curious. You know, obviously there are a hundred writing tools out there now for, you know, I mean, so even stepping beyond ChatGPT, which will write or Claude, which will. Right. Or Lechat or hugging face or Grok or, you know, they'll, they'll all do sort of the same thing. But your tool, you've dialed it in for some, well, I mean, purposes and it's even called purpose, Right. So I'm not trying to be clever by saying that, but, but what have you done in your tool to make sure that what's coming out of it is actually meeting the needs of the user.
- [16:35]
Petter Magnussen
Right? Yeah, so. And that's when we get a little bit nerdy now because what, what I do is to, to not confuse the AI is that I call the, the AI the LLM, ChatGPT or Claude in this case actually, which is what I'm calling, I call them in small chunks and I can decide. Okay, I'm gonna do this techie part. I don't know if you. You can edit this out if you don't want it, but anyway, there's something called Context window. So every time you write a new prompt, you will also send along everything, the. The whole old conversation. And sometimes it even warns you, like, hey, should you really continue this conversation? Because it's starting to get very expensive. Because you see, even if you write short prompts, it's going to send away all the old conversation with it. And that can sometimes confuse it. So in this scripting language, you can choose if you want to send along the context or not. So you can make little simple questions so that. So this is not really about how writing the writing style or anything, but it's just making it easier for the AI to solve a certain task. So I could, like, I could scrape a web page, which is why I do in one of these guides or like mini apps inside of purposerite we call guides and one is called Article Writer. And then you're given a URL, it will scrape that page and then you can say, please find suitable target audiences. Who do you think I'm trying to talk to on this website? Then it can give you suggestions. And then I will only do that. So that is a small, simple task. And then I can be pretty explicit about that task and give it a pretty advanced prompt for that task. Because it's only one task. If I would mix that into a longer prompt, that would do 10 things, it would be too much information. So that is one way that you can like kind of give it one simple task, do this, and then will perform much better. So. And the other thing is that. So I'm doing that in steps. So first, if you look at Article Writer, you can of course yourself say, if you know your target audience, you just write it in in there. Or you can say, please help me analyze what do you think is the target audience? And then it will look at the target audience and compare that to the. I mean, I'm talking mostly LinkedIn content here. So you are probably trying to sell a service or a product.
- [19:10]
Rob
So, well, I mean, copywriters do everything, right? So it can be everything from, you know, content for LinkedIn, whether that's for our own businesses, or whether we're helping a client write that kind of an article. But obviously LinkedIn will have a character limit where maybe a blog post might be able to go longer or a sales page which could go for say, 18 or 19 pages, right? Where you know, it can get very extensive depending on what the product is. So obviously there's lots of different contexts as to the kind of copy that we're writing or to the audience that we're writing to. And your tool, as I understand it, is, Is maybe better for some of those shorter length ones. But does. Would it also help write a sales page at, you know, considerable length?
- [19:51]
Petter Magnussen
Yeah, it can do longer stuff, absolutely. The thing is, I have chosen to focus on one area and this is purely actually a marketing decision, because being a scripting language, I could actually make those scripts do anything. You know, like I could probably make a script that helps write a chapter in a book or something. Maybe not the whole book, but at least, you know, so. But I'm, I have decided that my target audience is people trying to build a personal brand or a company brand on LinkedIn with content. So just to, you know, narrow it down, make it easier to place this product, but there's nothing stopping it to do other things. And that's why I talk about content so much. And this is, you know, because this is the target we have now.
- [20:42]
Rob
Yeah, that is. And so this is, this is interesting. Obviously, a lot of, A lot of copywriters are on LinkedIn are using the tool to, you know, create interest in their services and their thinking. One of the criticisms I hear about LinkedIn is that it is full of AI generated content and that it's not reflective of true thought leadership. You know, people aren't actually writing what they're putting up under their own names. Now, I, I'm sure that this is happening with copywriters too, although that boggles my mind that writers don't actually write. But, you know, maybe it's happening. I hope that most of this content that I see, you know, on LinkedIn is actually real. But how does your tool help pull out actual thought leadership, actual real writing, and not. Not become the writer itself?
- [21:37]
Petter Magnussen
Yeah, that is a fine line, actually. And I'm trying to avoid. Because I could of course, make a guide that says, press button, get content.
- [21:49]
Rob
Exactly. Especially if you can. If you can just read another website or another post or whatever, rewrite it in this voice like that. I mean, I can do that all day long.
- [21:58]
Petter Magnussen
Yeah, exactly. So, I mean, I could do that, but I hate that stuff myself, so I'm trying to, to not do that. And these guides that I do, and I think that that is actually potentially even a weakness of the tool. I've heard from some people that, you know, what I'm doing is I'm asking you to provide a lot of information, to write something, you know, okay, I can look at the webpage to find the target audience and the pain points and stuff and background information about that product or services or something. But you still have to choose, you know, what pain points you want to address here out of suggestions that you can add yourself. And then it will also, after that, it will come back with areas that it thinks you should provide additional information. So it's. It will not base the whole piece on just that background information. It will say, okay, I found these 10 areas that you could potentially say something about. And then you can write that in. Of course, you can be lazy and say, nah, not gonna do that. But if you take your time and do that, the piece is gonna be so much better. And that, as I said, that is the same with style and voice and stuff like that. You can specify so much. And that is also, like I was saying, part of the weakness, perhaps, because I'm looking at spying how people are using the tool, and I see that quite a lot of people are actually bailing out in the process. So that's something I need to address, I think. But I'm not sure if I should go all the way there either to meet those requirements, because I think a lot of those people come with the expectation that, oh, it's a content tool. It's one of those. We scraped a thousand LinkedIn posts, and now you just put in your company name and press this button and you have copies of that content. And if you expect that and then you get a million questions about tone of voice and target audience and stuff like that, it's not what you expect. But that is also what makes it different in the end. So I think it's a matter of finding our target audience. You know, who is, who is this for? And.
- [24:15]
Rob
Yeah, well, I think one of the things that I've struggled with with AI is, you know, that lack of humor, humanness, and the fact, you know, when people find out something is AI. You know, if we were having a chat, you and I were having a chat, and I found out that you were a bot AI, Suddenly the value of what we're talking about is reduced immensely. In fact, I might even feel cheated that. That I got the Petterbot instead of the actual thing. Right. And so I think one of the challenges that creators like you, who are, who are coming up with these tools is, is how do we preserve humanness if we're using tools that are very unhuman or inhuman in. In the Creation. And, you know, there's a whole spectrum of criticisms that happen here, you know, where, hey, if I can, if I can take an artistic style and an AI can create an image in that style, and I no longer need the artists to do that anymore, what's the value of style anymore? Right. Or, and, and the same thing applies, obviously, with writing, with copy. And so these are some really big questions around AI that I know people are asking them. Clearly they're being asked, but I don't know that there's a lot of discussion or a lot of people that are really thinking through, like, okay, this is actually something important we ought to be preserving.
- [25:40]
Petter Magnussen
It is, yeah, it is super important. And yeah, it brings another thought to my mind there that I think there will always be a reaction to things like this. You know, people are starting to be really, really good at spotting the AI generated content. And you're just like, you scroll past it, because as soon as you feel that this is AI, so that is a problem. But I think there will always be a reaction. I'm seeing already now a reaction to this. I, you know, I am connected. I'm hanging around, talking a lot about LinkedIn, but that's my hangout. You know, I'm spending a lot of time on LinkedIn and I'm connected with a bunch of professional copywriters of different types. And I can see the reaction happening. A lot of people are starting to write in a, like, consciously anti AI way, which actually is a bit dirty, sloppy or even, or something like that, you know, because AI is writing so good in a way, you know, formally. Yeah, yeah, it's nice. Sentence lengths and no grammar problems and no spell errors, spelling errors, you know, all of that. And that's where I see a lot of my copywriter friends are going ballistic with trying to prove that they're human. So that's an interesting trend that I kind of like that.
- [27:10]
Rob
Yeah. From somebody who is a bit of a stickler for grammar, it's really painful to see it. But on the flip side, again, we as humans value humanness. Now, I am not anti AI at all. I use AI every day in my business, and I think it's an immensely helpful tool. But these are some of the things that I've been thinking about too, is, okay, when I write an email to my list, and, and I've yet to use AI to write an email. One exception, when I was saying, hey, look, what, what I can do in an email a couple of years ago, but should I be putting a tag at the bottom, you know, it's like, hey, this is written only by me. I did not use AI right. So that people see that or you know, like, maybe there's an opportunity at some point for somebody who's going to develop, you know, a validator where it's like, hey, yeah, this is 100% human written. Or you know, this is AI assisted. Or I know, I know there are tools, you know, GPT0 and those kinds of things that try to establish that, but they're, they're not great. They're not great yet.
- [28:14]
Petter Magnussen
No, no, it's, it's super easy to trick them. And here I, I do feel a little bit dubious about myself because I have, I have a guide in purpose, right, that is called Read Write and that is pretty good at mimicking a lot of these things so it can trick these AI detection tools. And I'm sometimes wondering like, should I really do this?
- [28:41]
Rob
No, these are, these are good questions. And I mean, at some level we know this stuff is going on, it's going to happen and we've got to embrace AI, you know, in order to do some things. But also it would be helpful if, I hate to think the government needs to do this, but you know, if the industry could self regulate a little bit and just say, hey, this is where the line really is and you should really shouldn't be, you know, getting too close to this line, you know, for some of this stuff. But yeah, again, who knows? It moves so fast, right?
- [29:14]
Petter Magnussen
And it does. And I think another conclusion that could be reached is that it doesn't really matter, you know, as long as the writing is good and I can get some value from it. Does it matter if I helped do it? You know, that's view of it.
- [29:30]
Rob
So yeah, yeah, that's, that's a great question too. And, and how far, you know, does, does that go? Obviously, like I said, I use AI in my writing for helping with headlines, for identifying benefits and it speeds up the process for identifying needs and problems. There's so many good things that it can do. And so you know, where you're, you've created this tool too that also writes really well or can help. If I'm writing an article or whatever and I get stuck, I'm not sure where to go or if I need ideas. Like it's super helpful at prompting those kinds of things. And so the right mix of machine and brain, human brain, can actually elevate where we are if it's done right.
- [30:14]
Petter Magnussen
Exactly. Yeah, I think so. And I have actually become Less afraid of the creatives losing their job. I mean, yes, creatives are losing jobs at this moment. And especially, you know, I'm in the photo video business. A lot of photographers are probably going out of business because of this. But that has happened before in that business. We're more used to it than on the text side. Actually for text, this is very new, but it's happened on the photos. I can give you an example. When digital cameras came, you know, it, the small local newspapers, if they would like go to a farmer that had a funny carrot, they would bring a photographer specifically to take a picture of the carrot.
- [30:58]
Rob
Right.
- [30:59]
Petter Magnussen
And then digital cameras came and then that photographer was out of the job because the, the writing journalist could take that picture of the carrot. So we have seen this before and yeah, those low, yeah, excuse the word, but low level work that goes. And we've seen it happen and that's maybe what's happening here too. That, yeah, some of the low level work is gone and that we just have to live with that.
- [31:24]
Rob
It happened with stock photography as well. I mean, I, you know, I was in the ad agency world when stock photos started coming online. And I remember specifically photographers we worked with that were, you know, worried about, you know, they were, they were losing a lot of the typical, we'll take a photo of this typewriter or you know, if it, if it wasn't original, if it wasn't specific to an assignment, suddenly there were a thousand images out there that we could choose from. And so you're right in some ways there is a stock copy thing that is going on. It's a little bit different from stock because it's not always going to be the same. But you know, AI creates that stock. And so the challenge becomes how do we as humans or humans using a tool like PurpleRight, create the stuff that's original, new, different and unique and you know, shows off that creativity, that humanness in some way that Stop Copy no longer does.
- [32:27]
Petter Magnussen
Exactly. And that's where I'm not so afraid. I mean, if you're operating on a slightly higher level that you're a bit more strategic and creative, I think you're still needed. Because if, if we go back to images again, it's really easy to prompt a cool image. But what does that image say? What does it do? Same with the text. You need to have a vision first of all, for where you are going. Even if you're using AI, you need to know which direction you should prompt the AI. And you need to be able to judge the Result also something comes out, is this good? Is this on brand? Is it, you know, and so we, we are needed I think still. So I'm not that worried, not yet at least.
- [33:12]
Rob
Yeah, I think what you're saying is the opportunity is, you know, an analysis and strategy and figuring out direction and as far as like the, the actual operational tactics, you know, we might need to figure that stuff out, but the exact wording or whatever, you know, there may be a place for AI to step in and do like you said, that lower level thinking that can be easily replaced.
- [33:38]
Petter Magnussen
Yeah, yeah. And again, referring back to the visual side of things, we have seen a shift also there that users want to do stuff themselves. And I think again that's something that we just have to adapt to. So, so what we're doing is, you know, we would normally go and film everything for the customers. Now we're having a new service where we can actually work with the customer to think about what is what, what should be filmed, what should be said. Then they can actually film it with their iPhone and send the files for us to editing. So our role becomes different. And the most important part there is the strategic part, like what should we, why should we film, what should be said, what should. And then who holds that camera? And if that camera is a professional camera, it's not that important, to be honest. And I think it's a bit the same with text, a lot of simple text I think the customers can write themselves. Now we just have to realize that we're not going to do that so they're going to come to us for the more strategic and creative stuff and that's where we can shine and yeah, and also on a purely strategic level, help them point, you know, the stuff that they make themselves. We can probably help them with that. Like you know, take good, good hourly rate to. Because that's what we do. We only, we have this so like social media package and then we spend quite a lot of time thinking, okay, you want to build your personal brand. Who are you? Well, as a copywriter then in this case, you know, who are you on LinkedIn? Why, why are you different from the other copywriters? What Persona are you going to have when you are on camera? Are you going to wear a T shirt, a suit or what is, you know, to be on brand? And that I think is super valuable. And we need to realize that for, for text too, that that is a job that needs to be done and that we can charge pretty high for to analyze things and have a strategy.
- [35:39]
Rob
The Irony here is from a copywriting standpoint, and this probably is true from a visual standpoint as well, is for the last, you know, 100 years or so, we've been selling the outcome and giving away the strategy for free in order to get the outcome. So if I sold a content strategy or a blog post or a sales page, that's what I'm selling. But all of the thinking, the strategy that goes into it was happening, you know, kind of, kind of for free, you know, on the back end. And what AI is forcing us to do is saying, hey, the words are free, but the strategy, the figuring it out, the approach, the voice and putting your mind power into that, that's where the real value was all along, and charging for that. And that's definitely a shift in the way that we see the creative business.
- [36:32]
Petter Magnussen
Yeah. Wow. That's exactly my thought, but you twisted it in a great way and I agree totally with that. And now I don't want to be here promoting purpose, right. All along, but I have been thinking about exactly this problem. So I have made in purpose, right? I have something called profit sharing. And that is because I've been thinking exactly in these lines that how do we handle the situation where the customers want to write some stuff themselves? They know I have chatgpt, I can write. And as a copywriter, you just going to be. They're just going to call you when there's something really important to write, you know?
- [37:16]
Rob
Yeah.
- [37:17]
Petter Magnussen
So. And the rest they're going to make with ChatGPT and it becomes horrible. So, and that's what I'm thinking, that why don't we as copywriters sell analysis? Like, again, who are you? Tone of voice, brand like and different brand for different services, maybe even because we have different target audience for different parts of what we're trying to communicate. Why don't we sell that and in the case of purpose, right, make custom guides for the customer. So instead of the customer going off to ChatGPT and prompting really badly, we can say, I've done the analysis for you and here are a few custom guides that you can use. You know, I know you're writing it. You're writing this simple newsletter every Friday. You can actually do that. And I made a custom guide for you for that which is on brand and everything is prepared for you. And the cool thing then is that you will, as a copywriter, you get a kickback. So when you make a guide and share with someone and they use it, you get a kickback from that. So that's Kind of a way of. Yeah. I'm also thinking that staying on top of mind with the customer. So I have on purpose made a lot of the guide, have a lot of description lines. And I'm thinking that people can actually put a little tagline there, like, remember me, Rob, when you want to write the important stuff.
- [38:41]
Rob
Yeah, yeah. Interesting. Yeah, that makes sense. So that kind of opens up a broader question. How do you see copywriters using AI in general moving into the future? Obviously, it's going to be doing a lot of writing or writing assisting, but how else can it help us improve the product that we're. We're providing for our clients?
- [39:02]
Petter Magnussen
Yeah, I think that. But most of it we're already doing. It's just that it will be better at it. I think the. One of the most important things is just analyzing text and massive context, you know, that is so useful to just pour stuff in there and be able to talk to that massive data chunk. So that, I think is a super important thing. Save a lot of money. Because that's. That's not the fun part of the job either, is it?
- [39:33]
Rob
Yeah, analysis, I mean, well, I mean, it can be fun, you know, when you stumble across those things that it's like, oh, wait, I, you know, I just discovered something new. But I don't see that being that different from AI helping me uncover something new, you know, and saving 10 or 12 hours of going through, you know, spreadsheets worth of data or whatever and having it say, hey, have you thought of this? You know, the, the. I guess my question will always be, do I trust that the AI is uncovering the stuff that I would uncover? Right. Like, how do. How do we make sure that that analysis bot is better or as good as Rob when he's, you know, analyzing a spreadsheet. Now, I'm not necessarily saying I'm all that great at analysis. And so it may. It's probably already better, but obviously, you know, in using a tool, you. You have to trust the tool.
- [40:27]
Petter Magnussen
Yeah, exactly. Yeah. And I, I think there, of course, we are at the point where it is probably, as you say, probably better. As long as soon as you have a lot of least. I mean, if you have a couple of pages, you're probably better, but as soon as you start getting hundreds of pages or something like that, I don't think it's very unusual that a human can keep up with that in any way.
- [40:52]
Rob
Yeah.
- [40:53]
Petter Magnussen
So I think that that's an important one. And then I actually think it can help with creativity and this might sound a bit weird, but I do think that it triggers a lot of thoughts, you know, and I actually made a guide for that in purpose I called Idea Expander. And it's not doing anything advanced at all. It's just a very simple prompt. Basically you just give it a basic concept and it will then twist it around and say, or you can think this way on this topic and then you can take one of those and branch out from there. And that helps a lot, I think, with just coming up with ideas. So you can, like I put in simple things like, you know, it's important to think of the target audience when you write content and then it twists that around and comes up with something entirely different. And so that's how I have done some of the best content pieces, actually using that kind of, you know, to, to trigger my creativity.
- [41:54]
Rob
Yeah, yeah, that's when I'm using AI, that's basically one of the main purposes that I do too, is that I'm, I'm asking the engine to, you know, ask me back questions or to help me think about it in different ways or unconventional ways. And you know, that back and forth between me and, and the AI seems to, at least to me, it feels like it's helping me get to something slightly different or something newer than my own thinking.
- [42:23]
Petter Magnussen
Yeah, no, I, I think so too. I think we're, we are as it's super interesting with AI, how it operates. I think it gives a clue about ourselves. I'm of the opinion that we are just AI and the algorithms are very similar. Because if I look at going back to images, image generation again, how that actually works inside of an AI. So you take a pure noise image, like total random pixels, and then you, from there the AI sees something vague like, oh, it's a little bit dark up there and the prompt was dog, you know, and then it's going to change those, like assuming that could be the head kind of. This is a very simplified way of describing it, but I, I see how that is close to how at least my creativity works. You know, it's really, really hard to sit there with a blank page. But as soon as you write something, that could be the start, you know.
- [43:28]
Rob
Yeah.
- [43:29]
Petter Magnussen
With an empty page and you're going to write, oh, I'm going to write an essay. Or you just write sun and then, ah, sun. Right. And then you can start. It's the same as this little blob in the noise in a way. So I, I think it's basically the same how we operate so it will be interesting to see how, how quickly we get to AGI. You know, this real well.
- [43:52]
Rob
That was going to be my next question. As somebody who, you know, is, is operating in the space, working with these models, do you see AGI as a realistic next step and how quickly do you think we get there? If it is realistic?
- [44:11]
Petter Magnussen
Yeah, the question is about the definition. There's been some weird definitions, definitions floating around. Like, some definitions have been financial even, like when, when AI can make profits bigger than certain numbers, it's AGI, which is really weird in my opinion. But yeah, I'm not sure. I mean, we are already at a level where AI can do many tasks better than most of us. You know, if you look at a lawyer or doctor even, you know, AI can, can do mo better than most of us on that. But on the other hand, it's also completely wrong most of the time. Like there's always some hallucination in there and I'm wondering if you cannot. Yeah, that, that seems like a big threshold. Like it needs to somehow understand what is realistic. And that is not. I don't see that happening. It's just getting better at what it does, but it's still stuck in this hallucination world. So. Yeah, I'm not a machine learning expert.
- [45:27]
Rob
But yeah, yeah, neither am I. On the other hand, humans also glitch and hallucinate and, you know, break down. So maybe, maybe part of being an actual intelligence is sometimes we're also actually dumb.
- [45:43]
Petter Magnussen
Yeah, exactly. And very stubborn about being right.
- [45:47]
Rob
Yeah, there could be something to that. Okay, so what's next for you as owner of purposerite, where do you see this tool going and doing in the future?
- [46:02]
Petter Magnussen
Yeah, I think what, what is happening now is that now the tool is working and I need to start making a lot more guides. And I'm thinking that I should actually cooperate with copywriters on that because I have limited capacity and I'm not the best copywriter, to be honest. So I, I need to make more guides that is doing more stuff because right now, to be honest, it's a bit limited. The guides we have and they take a lot of time to make these guides. It's a lot of work to. You know, it's really tricky to make them because, yeah, to get technical a bit, but as the prompts again get longer and so on, it's hard to balance them. Like if you have a lot of choices, then you can have situations where one choice is canceling the other. So. And right now I'm the only One that can do this. I need to find some people who can actually help me make, make guides.
- [47:06]
Rob
And those you're probably talking to the right audience. I have a feeling that you may get a few emails from listeners who are eager, eager to find out more about that kind of an opportunity. But it, it does seem to me, and I've been saying this for a long time, but for copywriters who have resisted using AI for a variety of reasons, you know, whether it's not human enough or it doesn't feel right, or, you know, worries about plagiarism, that kind of stuff, it still feels useful to me to be experimenting with these tools. You don't necessarily need to publish what you're doing or, or use it for your clients, although ultimately you may choose to do that. But understanding how these engines work, the differences between the different kinds of applications, it's a little bit like refusing to use AI, feels a little bit like refusing to pick up a smartphone. Yeah, you can continue on with the flip phone, but the rest of the world is doing some pretty amazing things that you no longer understand, you're no longer a part of. And yeah, you could, you can set down your smartphone, you know, there, but if you haven't picked it up in the first place, you don't know what's going on. And I think there's maybe a metaphor there for how we should be engaging. And purposerite may be a tool that people start to play around with a little bit on their own and there is a free option so people can, you know, play around and see how the tool works without worrying about having to pay for it at this point.
- [48:37]
Petter Magnussen
Exactly. Yeah. And you actually have ChatGPT and Claude in there, so. And that is a, a fun thing also to just try the same prompt, same prompt on ChatGPT and Claude and see the difference. They do have a personality, I would say. Yeah, it's, it's funny because in, in, in the article writer guide, for example, I'm using ChatGPT for research things like find the target audience from looking at this webpage. But then for the actual writing, I think Claude is better.
- [49:07]
Rob
I agree.
- [49:08]
Petter Magnussen
Yeah, but they are, they are doing a different thing. Like Chat GPT is way too formal and I'm struggling to like loosen up while, while Claude is sometimes the reverse, it's being, I think it is trained much more on social media posts and stuff because it easily becomes way too like tick tock style basically. And then I have to tell it to calm down, you know, like become a bit More corporate or, or strict or something. But I think that is easier than having ChatGPT loosen up, is more difficult than having Claude cool down a bit and be a bit more strict.
- [49:46]
Rob
So yeah, it's an interesting challenge. And again, one, if, if we're just prompting and using the tools that it becomes pretty obvious the differences especially, I mean the two big ones are obviously Claude and chatgpt. But there are, you know, I mean Grok is incredibly sarcastic because it's trained on that Twitter data, which is a place where so many people are, are unkind and very sarcastic. You know, if you're using meta or the chat, you're just going to get a different outcome with each. And so yeah, my takeaway here I think is, hey, well there's really two One, we need to be playing with these tools and two, we really need to be upskilling when it comes to strategy, thinking, conceptualizing and those, those higher level marketing skills because like I said, you know, where the words used to be the thing, we sell the words now free. And it's the thought that needs to bring the value to the table.
- [50:47]
Petter Magnussen
I think that sums it up very well. You know, the words are kind of free now so that we need to focus on what our actual knowledge and that is to realize that that is actually not putting the words down.
- [51:00]
Rob
Yeah. Peter, if somebody wants to try purpose, right. Or to connect with you, where should.
- [51:06]
Petter Magnussen
They go if they want to try purpose, right? It's purposeright.com and purpose. Write as in writing then and not write as in being right.
- [51:16]
Rob
W R I T. Most copywriters should be able to figure that one out.
- [51:20]
Petter Magnussen
Yeah, exactly. Right. Yeah. So that's the one. And you can write to me on inforposerite.com and I will reply and hopefully I can. Yeah, depends on what you ask. Hopefully I can answer something reasonable.
- [51:35]
Rob
That's been interesting. I love diving into these tools and seeing, you know, where, where we are going and asking some of these questions. So thank you for your time. I appreciate it.
- [51:44]
Petter Magnussen
Thank you so much. It's been really fun and yeah, I'm amazed. Time just flew by here.
- [51:52]
Rob
That happens. Thanks to Petter for sharing his thoughts about AI and purposeright. I've linked to Petter's contact information and to purposerite in the show notes. Even though I've been thinking, thinking about and talking about using AI for the past few years, I still have really mixed feelings about the tools and their impact on those who use them. Both our clients and copywriters, content writers who are using them in their own businesses. I teach a marketing class at a local college in my city, and I've seen the use of AI skyrocket over the last couple of years. My students love it as it makes writing easier for them, particularly about topics that they may not know a lot about. But in asking a model to write a paper on a particular topic, a student who uses AI is not doing the same level of deep thinking on a topic as a student who's not using those tools might do. I tell them I'm interested in seeing their thinking, not the way that they use AI tools to do thinking for them. And the same is true for marketers and copywriters. When the starting point is artificial intelligence, the end result may sound okay, but it will lack the human input that adds credibility, trustworthiness, and reality. But more importantly, copywriters who start with AI are outsourcing the strategic thinking that makes marketing different and great, and if they don't learn how to do that at some point, it's going to impact their career prospects. On the other hand, others are doing it, so I'm going to fall behind if I don't. It's a bit of a catch 22 and goes back to what I said at the introduction. When the words are free, we need to learn how to do the activities that make the words more valuable, things like strategy and analysis. I highly encourage you to try these tools out, but not as a starting point, as a partner to elevate your thinking to help with research and brainstorming for improving, not replacing your thinking. And like I said, we've got some great resources to help you do that. Inside the Copywriter Underground, which you can find at the copywriterclub.com forward/tcu that's the end of this episode of the Copywriter Club Podcast. If you like what you've heard, please share with someone you know. If you don't know another writer or freelancer that you can send this to via email with a little note that says, hey, I think you'll find this interesting. You can visit Apple Podcasts or Spotify or wherever it is that you listen to your favorite podcasts and leave a review. If you haven't left a review before, now is the time. I promise when you share the Copywriter Club Podcast, your friends will thank you. Don't forget to check out the copywriter underground at thecopywriterclub.com forward/tcu. I'll see you next time, freak.