Loading summary
Transcript69 lines
- [00:01]
Michael Stelzner
I'm grateful for marketers like you. Not the ones waiting for their boss to tell them what to learn, but marketers who actively plan for their future. Because you listen to this podcast, you're already ahead. You're seeking to understand AI instead of waiting to see how this AI thing turns out. But here's what I've learned after a decade of running conferences. Interest doesn't create results, implementation does. That's why we created AI Business World 2026, where you'll master AI skills that make you indispensable, where you'll get your questions answered by experts, and where you'll connect with over a thousand marketers who are implementing AI right now. Years from now, you'll look back at this moment and remember this is when you got ahead. Head to AIbusinessWorld live and secure your competitive advantage.
- [01:00]
Audrey Chia
Welcome to the AI Explored podcast, helping you put AI to work. And now, here's your host, Michael Stelzner. Hello.
- [01:11]
Michael Stelzner
Hello, Hello. Thank you so much for joining me for the AI Explored podcast brought to you by Social Media Examiner. I'm your host, Michael Stelzner, and this is the podcast for marketers, creators and business owners who want to know how to put AI to work. Getting AI to write high converting copy can be a game changer for your business. Most people simply accept the first result they get from AI. But is that really wise? In today's episode of the AI Explored Podcast, we'll explore how to use AI to create high converting copy. My special guest is an AI strategist and conversion copywriter who helps startup founders and entrepreneurs build brands that convert. She's hosted the AI Marketers Playbook podcast owners. Audrey Cha, welcome to the show. How you doing today?
- [01:55]
Audrey Chia
Good. Thank you for having me, Michael. Excited to be here.
- [01:58]
Michael Stelzner
I'm excited that you're here. By the way, today, as we stated at the top, we're going to get into copywriting. But before we get into that, how in the world did you get into AI? I would love to hear your story.
- [02:08]
Audrey Chia
Definitely. So I did not plan to get into the world of AI. Right. Fun fact. When I first launched my consultancy course with copy three years ago, ChatGPT just launched, and my client actually sent me a text and he said, audrey, you're going to be replaced. And he sent me the link to ChatGPT. So I was extremely horrified, but at the same time, extremely excited about the potential of AI. And at that point in time, I knew that I had two options, right? One, fight the current number two, ride the tide. So I'm super glad I decided to do the latter, because now I fully integrated AI into my business, my clients businesses, and I'm using it to really grow my own business at a scale that I never thought was possible before.
- [02:55]
Michael Stelzner
As you know, I used to be a copywriter as well, and I have a lot of friends in the copywriting world that absolutely freaked out when ChatGPT came out. So, like, share with me a little bit about what was that like for you? Were you a copywriter before ChatGPT came out? You were. Okay, tell us a little bit about, like, what went through your mind, you know, because the struggle was real, right?
- [03:15]
Audrey Chia
So the first thing I got ChatGPT to do, right, I was like, okay, you know what, write a Nike ad. I bet it cannot write a Nike manifesto for me. And you know what, the first version of it was pretty good. So at that time There was only V1, right? Of course it wasn't perfect. At that time, I think ChatGPT was a junior copywriter, right? But over the past three years, it has evolved. And I would say that right now, if prompted, well, it can be a senior creative director. That's how fast it has grown over the past three years. And of course, when I first saw its potential, I already knew that this is incredibly powerful. There was a bit of fear in me, but I also had this feeling that if I'm a copywriter, I would be the best person to leverage it for copy. So why not lean in instead, right? If something is challenging, it could also be an opportunity for growth.
- [04:07]
Michael Stelzner
I love that mindset. So from your view, what do you think is one of the biggest challenges or misconceptions that people have when it comes to AI and copywriting?
- [04:16]
Audrey Chia
I think from what I've noticed, a lot of people tend to use ChatGPT's content as is. So one of the most frustrating things for me is we can very easily tell what is AI content and what isn't, especially if somebody doesn't use that human lens. So if you see words like revolutionize, you know, or you see phrases like it's not this, it's that, it's usually very clearly chatgpt. And if you are not a writer or if you don't put in the effort to review that content, you. You tend to take things as is. Sometimes I see people also copying the system instructions when they are creating content, and that's a really clear indicator that they are not carefully reviewing the work. The biggest mistake there is when somebody realizes that it's AI or it's lacking that human touch, they are more likely to be put off by your content and you're going to be doing a lot of damage to your brain in the short and even long term.
- [05:14]
Michael Stelzner
Yeah, I like that a lot. I recently saw, I don't know if it was real, but there was something going on on social media that showed a newspaper and I think, I don't know if it was from England or whatever, and they accidentally left the prompt in the article on the front page of the newspaper.
- [05:28]
Audrey Chia
I wonder if the editor is still going to be there the next day.
- [05:31]
Michael Stelzner
No, it's pretty crazy. But I agree with you that because for especially people that are not writers, professional writers, it is such an upgrade over the kinds of writing that your average marketer or entrepreneur is capable of creating that it just seems so good that this desire to want to just copy and paste it is there. But people are pattern recognizers, right? Humans are. And we can tell sometimes when it's written by AI because there's been a voice change or there are certain kinds of frameworks that it tends to use repetitively over and over again and it feels somewhat repetitive. What I would love to ask you when, when AI and copywriting are coupled together in a good way, in a beneficial way, what is the upside? What's been your experience, the benefits of this?
- [06:20]
Audrey Chia
I would love to say that ChatGPT is my bae or my best friend. Now I've learned to turn it into my intern, my coworker, my partner, a consultant who can operate alongside me. And I say that because personally, as an entrepreneur, it has allowed me to enter at least 3x the amount of work I can do or amount of clients I can work with at any one point in time. Whereas in the past I would have been limited because of my ability to produce quality content, right? So being able to grow my business is number one for clients. The benefits are amazing because now they can rapidly test content at scale. So here's an example, right? I have a client that is a crowdfunding charity in Singapore and they have never run ads before. So prior to me stepping in, they have never done ads. They don't know anything about putting messaging out there that gets donations. So of course the first few iterations we couldn't really figure out what is the right message to test, right? But what ChatGPT was able to give us were very interesting messaging angles that were tied to clear audience insights. So with AI, we could quickly develop a whole series of ads, messages, angles and put it out there and within less than three months we got from $1 return on ad spend to a $7 return on ad spend. So for every $1 we invested, we got $7 in donations. This process might have taken a lot longer without AI. This process might have been a lot more rigorous and intensive without AI, but with the help of it, we could very quickly drive results. So I think the time that it takes to really get to your end results is. Is a lot shorter now with AI compared to before.
- [08:05]
Michael Stelzner
I love that. And I will just share a little story myself as a copywriter. Many people that listen to this podcast may know that I formerly had a career as a copywriter. Just like Audrey. I believe I'm a good writer, and I believe that AI has made me a superior writer. So I call it like, you know, I'm like, human enhanced because I'm already really good as a writer and because Audrey is really good as a writer also, I believe that one of the big upsides is this can make us exceptionally better as writers. So for those that are listening, that already have the gift of copywriting or writing in general, this can make you a much better writer. And one of the things that we've been trying to do inside of Social Media examiner is increase the segmentation of the emails that we sent. So in the past, we would send one email to everyone. Now we've got all these little niche audiences that we can send to where I did not have time to create the copy in the past. For example, we have people that are interested in AI because they've submitted inquiry for one of our reports or they read some of our AI content. There are our AI interest list. Then we've got people that belong to our AI business society. So now I'm writing like five times as much copy. Do you find this to be true as well, Audrey? But I'm producing it in almost the same amount of time that it took me to write one. Are you finding something like that to be true as well?
- [09:21]
Audrey Chia
Yes. And I love that you mentioned hyper personalization. Right. I think this is something that AI has given us that we couldn't do before. So being able to even tailor your newsletters to your audience, being able to tailor ads to your audience, even WhatsApp reminders, right now you know who has booked X number of calls and you know how to nurture them. I think these are insights that AI can give us as well as we can use AI to then create content that's tailored and specific to each audience to increase the effectiveness of any message.
- [09:56]
Michael Stelzner
Love it. Okay, well, You've got a process for leveraging AI to create high converting copies. So where do we want to start? Where do we need to begin if we want to do this?
- [10:06]
Audrey Chia
Anyone who's listening, That's a great question, right? And I think maybe I can give an example so there's a context to have this conversation around. One thing I love as a copywriter is creating high converting landing pages. And for many people they might just ask ChatGPT, Hey Chad, write me a landing page and then they will get something that's super generic, vanilla sounding, probably fall flat for me. I actually have a seven step process in which I go through very systematically to push AI from one step to the other. And I can even walk us through maybe a couple of those steps so we can see how that looks like in practice.
- [10:45]
Michael Stelzner
Perfect, let's do it. Where do we want to start?
- [10:47]
Audrey Chia
Awesome. So before anything, before creating any piece of copy, we always have to go into the audience insights, right? Any good piece of content that converts actually taps into somebody's pain points, fears, motivations and desires. So for many people when they create content, they're just writing something. But when you write something, it's not effective if it's not tied to what your audience actually feels. So my goal for any piece of work is to figure out what exactly are my audience insights. So of course there are, you know, the manual way of doing things and then there is also the AI supercharged way of doing things. In the past you would have to speak to different customers, write down or transcribe their notes, go through this very slow, painful process. But right now you can easily use an AI note taker, for example, Fireflies or Sybil, right? These are tools that you can have on calls or on your phone so that when you have client conversations, it's already all transcribed. And the best thing is you then pull all these transcripts, feed it back into AI or ChatGPT and ask it what are the common insights that you have noticed in these 20 over compensations. Now this process would have taken a lot of time if it's done manually and you probably miss out certain core insights. But with AI you get a super clear and comprehensive understanding of your target audience. So that's one way to do that.
- [12:19]
Michael Stelzner
I have a question on this. So if I understood you correctly, at this point, you're having conversations with customers or prospects, I would imagine, and what are you asking them? So that when you do ultimately use a note taker, it could be Google Gemini, it could be Firefly, whatever. What kind of questions are you asking so that you get the kind of insights that you're seeking?
- [12:39]
Audrey Chia
And I think one thing I like to do is to ask questions that tap into their emotional state. So what keeps you up awake, for example? Right. Why would someone want copywriting? Is it just because they want to make their brand sound a lot more professional? There could be a reason, but the innermost reason or the thing that keeps them up at night is no revenue for my business, my landing page isn't working, I'm not getting customers, I'm spending 50k on ads, and it's not giving me a decent ROI. Right, so you want to dig deeper into their pain points. So if someone responds with a superficial answer, you want to dig deeper and say, okay, so what's actually keeping you awake? What are you actually most afraid of? What's stopping you from taking action? Why would you say yes to something? Why would you say no? Are there certain hesitations you might have? I think all of these questions that dig into the emotional state really help. And if you need help, you can always ask ChatGPT to come up with a list of questions for you first that digs into their pain points, fears, motivations and desires and use that as your baseline.
- [13:45]
Michael Stelzner
Just side question, do you think surveys could be helpful here? Open ended questions on surveys as well for people that maybe don't have the time to do all these interviews. What's your thoughts on that?
- [13:55]
Audrey Chia
If you have a big enough pool of audience to survey, I think that's good because quantitative data is always good. But I think it would be more beneficial if you had both quantitative and qualitative data. I found that the best insights actually for me usually come from clients words. Right. So the thing that they say actually defines the message that goes out. So here's an interesting example. I used to work with a telehealth company for men. So we actually supported men who were going through sexual dysfunctions and needed medication for that. And we provided a discreet way for them to get those treatments. Again, it's a topic I'm not familiar with. It's a new audience and I had to figure out the audience insight. But there was this particular line that someone said that really stood out. And he said this in his own words. He said, if I cannot stand or function as a man, I am not a man. So if I cannot stand, I'm not a man. It's so interesting, right, because that really ties back to a lot of things. Identity, dignity, pride, sense of self. So it's so much more than just the physical act, but it goes internal. So my goal for client interviews is to figure out what are these core audience insights that we can pull so that we can distill it into messages that work.
- [15:22]
Michael Stelzner
Awesome. So what do we do next once we've got all these interviews? You hinted that you're asking AI to analyze this. Tell us a little bit more about what we're asking AI to do with all this data.
- [15:32]
Audrey Chia
Yes. So now you have all this data. Right. It's usually the copywriter's job to read everything and then to distill the insights. But it can be a very rigorous process, and not everyone might have that time or expertise to do so. So instead you tell AI, hey, you are an expert content strategist and copywriter. I want you to pull out key insights for my business, and you explain what your business is about. So giving AI the context so that it knows what insights to pull is very important. Otherwise it doesn't know what you're looking for. So the more specific you are with your results, the prompt itself, the better the results. So in this case, we will wrap up the audience insights section by really asking AI what are the common insights and golden threads that they can see through the different data points and interviews that we have. Only then we move on to stage two, which I would say is the competitive research phase. So once you figure out what your audience actually are like or are looking for, now we need to figure out who else is in the market who is pitching something similar to them. So we need to figure out who are your competitors, what are they saying, what are the key benefits they are claiming, and how do you stand out? A lot of founders think that they exist alone, but they don't. Right. So when founders come to me, they like to say, oh, we can do 100 different things. And then when I ask them, do you know what your competitors are saying? Most people would be like, I haven't looked into that. Right. So you don't exist in a vacuum. Remember, you're always in a crowded marketplace. You need to know what the other vendor is saying, so you can stand out and make sure that your wares are the most interesting ones. So what you can do here in this case is use, for example, ChatGPT's deep research or a tool like Perplexity, which I find very handy for research. These two tools, use them to really understand who are the different players in the marketplace and ask for key insights. For example, what are their core value propositions? What are they known for? What are their Key benefits. What are some client case studies they have? So you want to know where the market is at, then you can know how you brand yourself and position yourself differently and distinctly from everyone else.
- [17:59]
Michael Stelzner
So that is phase two question about this deep research. If someone does not know who their competitors are, how ought they prompt ChatGPT or Google Gemini or, you know, everybody seems to have a deep research tool. How ought they specifically ask this to be able to get what is quality research? Do you understand what I'm asking?
- [18:17]
Audrey Chia
Yes. So if you are struggling and you like, I don't know who else is out there. So usually what I would do is I'll tell chatgpt what's its role. So for example, as an expert content strategist working on a competitive landscape analysis, I need you to help me to pull out the top 10 competitors in my region that is specific to this niche and then I'll explain what I'm doing. Then I would say, please only pull out competitors that are. And then I could give it a checklist. Right. So sometimes what you want to do is to tell AI what you want and what you don't want and be as explicit as possible and don't just say, come up with competitors, give it a number of competitors. I also find that if you let it run well, sometimes it gets a bit lazier at the end or it can give you 20 competitors, but not in depth. So I want to quality control it and say 5 to 10 usually is a sweet spot and I can see what kind of information AI pulls out for me and see if it's suitable or not. If AI doesn't get it right the first time completely, fine, tweak your prompt. Or if you would like, use another AI tool and do the same research to see what you get.
- [19:34]
Michael Stelzner
Love it. Okay, so up to this point we've talked about the importance of gathering insights, ideally through interviews, one on one with people asking them questions that kind of reveal their deep desires, motivations, struggles, fears, all that kind of stuff. And then the second thing is to do some research on your competitors. Even if you know who your competitors are, you may not know what their unique positioning is. Right. What their value proposition is. I can help you gather all that information. So now we've got all this insights, what's next?
- [20:07]
Audrey Chia
So if you look at this, it's really a step by step process. So when I'm guiding AI, I'm giving it steps to get to where I need it to be. So the third big piece of the puzzle is figuring out what you as a company have to offer, right? Because if you tell me you have 100 different features, my question is, does your audience care that you have 100 different features or do they only care about the one thing that you solve? So we had to figure out of this hundred different things, what is a key vitamin and painkiller. So I always like to tell my clients that you could have a product that could be a vitamin and a painkiller. A vitamin is something that is nice to have. A painkiller is something that really solves a painful problem and you want to identify the painkillers. So we have to figure out all the different things you can do and deliver as a company. When what is that one thing your audience is going to care about that's going to make them say, aha, I need to try it, right? And that also requires a bit of testing. You may not get it right the first time, but what we want to do is to set you up for success as much as possible. Given the knowledge you know about your target audience, given the knowledge you know by your competitors, given the knowledge you know about your company, then we come up with that core value proposition. So over here, I like to explain your core value proposition as that one liner that differentiates you from someone else. In this case, it could be, I help target audience solve problem through solution, right? Or I help target audience achieve benefit through the solution. This gives you a very simple framework to get started so you can start thinking about who you're actually helping, what you're actually delivering, and how you're actually delivering it.
- [22:01]
Michael Stelzner
There's a moment happening in companies right now. Leadership is discussing 2026 priorities. AI implementation is at the top of the list. And someone asks who on our team can actually lead this? That's when it becomes clear some marketers are prepared. Most are not. The ones who invest in real AI training. The hands on, implementation focused kind suddenly become invaluable. Here's what happens when you attend AI Business World 2026. You walk in uncertain. You walk out confident. You arrive overwhelmed by AI tools. You leave knowing exactly which ones to use and how. You come feeling irreplaceable. You return indispensable. You multiply your output. You end guesswork on which AI tools actually work. You create content faster while keeping your authentic voice. And you automate hours of daily work. And you teach others inside the company how to do it. When your company asks who can lead our AI implementation, you'll raise your hand with complete confidence. Head to AIbusinessWorld. Live and grab your tickets today so you can become your company's AI expert. Okay. So I would love you to share with me what your value proposition is and how this value proposition kind of sits as a lens or a focal point for what you do with your copy.
- [23:36]
Audrey Chia
Definitely. So for me, I help founders and business owners build brands that convert through a combination of content strategy, AI and conversion copywriting. Now why does this work for me? Right. There are many other copywriters out there and there are many other marketers out there. But my unique experience comes from this blend of AI, content and strategy that very few people have. So I'm using these three things, which are traits about me, or in my case, my company, to make me unique. Then I'm figuring out what is the pain point of my target audience. They don't want their brands to sound good. Of course that's a good thing. But what do they want to do? They want their brands to make them money. So we need to build brands that convert. Right. And I'm very focused on startup founders or small business owners. Why? Because they don't have the resources to hire a full on agency, which would cause them tens of thousands of dollars. Instead, working with someone like me as a fractional lead makes it so much more convenient because it's flexible, you get the results you want, it's a lot faster, and it doesn't cost you an arm and leg to get there, which is something they cannot afford at this point in time. So I've nailed my icp, I've defined my offer, I figured out what my audience wants, and I've found out that sweet spot between the three things that I do that deliver the maximum value to them.
- [25:04]
Michael Stelzner
Okay, so what was the problem that you solved Again, just so can you say that one more time?
- [25:08]
Audrey Chia
So for the problem that I solve is a lot of people put copy out there that doesn't convert. So a lot of brands, for example, if you're on LinkedIn, you can be posting for two to three years and yet it doesn't convert. You have a landing page, you put it out there, you pump thousands of dollars and ads into your landing page, it doesn't convert. So my goal is to really build brands that convert and that can happen in many different forms.
- [25:31]
Michael Stelzner
Okay, perfect. So I love this and there's people that are very fascinated by you, but let's bring it back to them for a second. So when you say I help target audience solve problem through solution, what do you do with that now? Because everybody's going to obviously hopefully follow this model. And how do they use this now in collaboration with AI on all the future content that they create? Because this somehow some way is going to get fed back into AI, is it not?
- [26:05]
Audrey Chia
Yes. So one tip that I have in this stage of your journey is if you are struggling to form this value proposition, you can get a little help from AI. So if you tell ChatGPT, hey, you're an expert content strategist, I need you to put together a positioning statement for me in this format. Then very importantly, feed it all the information that you have found from your audience, Insights, competitive research company, Deep Dive. Give it these three buckets of information and then get it to form that one liner. When you do this process, remember to use your human brain to actually review the copy that ChatGPT gives you. Sometimes it sounds very confident, but it doesn't make sense. So use that critical human mind to look through the options, Pick one that you think works best for you and test it with your market. So once you've decided on that value proposition, now you can start building out your landing page. I won't go into details today, but right now you have all that you need to build a good landing page. So you have to tell AI, hey, I want you to use all this information plus this value proposition, and I want you to build this landing page in this specific tone of voice, this structure. I want you to include XYZ components, these information, and then it's going to give you a very strong V1. Right. So I'm not asking AI in a single prompt to build this landing page, but it's a very step by step, rigorous process where I'm guiding it through my train of thought to get to the end result that I'm looking for.
- [27:45]
Michael Stelzner
As a fellow copywriter, I'll just add a couple thoughts in here. I love this strategic positioning concept. I find that even when I'm out in public, like if I'm at a meetup or something and there's a big circle of people, the people that say, I help blank. You know, with blank, it's just like so much easier to understand what they do because you know how sometimes people ramble on and on, you don't really know what they do. So this is kind of the kind of thing that you could use over and over again in all sorts of different ways. But specifically with AI, what I like to do, I use Claude, some people use Chat GPT, this is not going to matter. Or I'll say, like, create multiple variations based on this information and Then what it will do is oftentimes create five to seven different variations. I'll read them all, and then I'll ask it which is best. And what I want to do is I want AI's unbiased opinion on which one it thinks is best. Do you ever do this?
- [28:38]
Audrey Chia
Yes. It's interesting you said that. There is someone that I know of who created a custom GPT just to critique work. So that is a workflow that you can also introduce.
- [28:50]
Michael Stelzner
Yeah. And I think the idea here of asking it, which is because you're going to have your own biases, because you're not necessarily your target audience. Right. And AI is going to follow the instructions pretty clearly. And it should make a case. Number three is best because it most strongly addresses the motivations and fears of your target audience. Right. So then you can say, okay, well, I don't like this one as much as the others, so can you model number seven and make it sound a little bit more like number three? And then all of a sudden it'll come up with more variations. And this is the part that copywriters kind of instinctively understand you have to work with the AI and don't accept its first iteration. And I know we're about to transition into this right now, so let's go ahead and move into the next part of the equation here. But folks, you know, like, challenging the AI, right? Or providing that. Okay, I really like this combination of words over here with this one over here. It's kind of like you just have to get in the creative sandbox a little bit. I found.
- [29:43]
Audrey Chia
Yes. I think, remember that the human part of things is as important as the AI part of things. So you need to give it very specific feedback so that it can get you the output that you want. Just by saying something like, I don't like this doesn't get you to your end result. You have to tell it, I don't like this, because there you go. And then you'll get better results.
- [30:04]
Michael Stelzner
So let's talk about how we can actually make it sound better. You know, now that we've got this strategic positioning kind of thing figured out, how does this work into the creation of our future copy?
- [30:15]
Audrey Chia
Yeah. I think one thing I can talk about is also the tone of voice. Right. Because this. Now that I have all the base information and materials, I think one thing that people often struggle with is Audrey. It doesn't sound like me. How do you even get it to sound closer to my brain? It sounds too generic. And that's something we need to fix with AI. Because. Because if you don't prompt it well, most AI copy sounds the same, especially ChatGPT's Claude is slightly better. But if you use ChatGPT, you've got to take note. It likes to use like what Michael said, same sentence structure, repeated phrases and patterns, and you got to watch out for it. So there are two ways in which you can prompt for a better tone of voice. Number one, you can give AI certain keywords. And what do I mean by that? AI is pretty good at registering certain keywords and pegging it to certain voices. For example, professional and corporate. If you use this keyword, you would see that AI tends to come up with copy. That's really good for a corporate brochure. And the tone of voice pretty consistent. Or if you use something like warm and personal, it would also sound a certain way. So it registers certain keywords really well. However, if you use something like witty, you would find that ChatGPT struggles with it because witty is very hard to define for AI. Everybody has a different sense of wit and humor. So you wouldn't want to use words like witty or humorous because it's very hard for AI to gauge what type of wit are you looking for or what type of humor you're looking for? So that is one way. So if your voice follows a more generic sounding category, this is perfectly fine. But if you want to go one step further and get AI to write in a voice that is a bit more nuanced, you want to beat it. Examples. So give it three to five examples of one type of writing. If you are looking at LinkedIn posts only, give it LinkedIn content. Do not mix LinkedIn content with website content with ads, because every single piece of writing is slightly different. Right? So give ChatGPT one type of content, three to five pieces, and ask it to analyze the tone of voice, style of writing, sentence structure, any specific quirks that the author has, then explain what's this voice, then replicate it for something else. So two different ways. Try both and see what works best.
- [32:54]
Michael Stelzner
Yeah, let's unpack both of these a little bit. Okay. First of all, you said give keywords like warm and personal, witty, which is it struggles with professional and corporate. In my case, it might be casual. Where are we giving these keywords? Are we giving these in the prompt? So, for example, create a email with a tone of voice that is, you know, warm and professional. Is that kind of where we're including that keyword?
- [33:21]
Audrey Chia
Yes, I always like to include it in the prompt, especially if I already know the content I'm going to write. So as an expert copywriter, can you please craft a newsletter of X number of characters in a tone of voice that's personal and warm? And I could say things like, avoid words like, so I can also be very explicit with when crafting my instructions so that it knows what I do not like, so I can have a list of banned words that I add into my prompt or add into my custom instructions so it knows to avoid these words, and then that actually gets you slightly better results. But I highly recommend the second approach in which you give AI examples, because then AI is able to learn from those examples and give you a more nuanced form of writing.
- [34:11]
Michael Stelzner
Yeah, I like this a lot. Now, to the examples thing, there's a couple things I wanted to explore with you. Obviously, you could take some of your best, highest converting work and you could ask AI to analyze it and to kind of come up with a description of your tone of writing that you could apply everywhere. The challenge, obviously, you know, as a copywriter, is that we write copy differently depending on the medium in which we write it. For example, a sales page is much more lengthy. An ad is much more simple.
- [34:41]
Audrey Chia
Yes.
- [34:42]
Michael Stelzner
And LinkedIn posts are a little bit longer and they all have different objectives. So do you recommend coming up with a tone of voice instruction set that is reverse engineered from examples, or do you simply recommend attaching examples every time? I'm curious what your thoughts are on this.
- [35:00]
Audrey Chia
Interesting. I think that AI finds it very hard to understand tone of voice when it's just a description. So if you were to ask AI to describe your tone of voice, it would describe my voice as confident, energetic, friendly. If I were to use this copy and paste description into a new chat, you would realize that it struggles to turn that description into actual copy that sounds like me. Whereas if I gave it an example of my newsletters, now it has a structure and a system to follow. So I like to say, think of AI as an intern. Right. If you tell your intern, write in a tone of voice that's confident, friendly, witty, and warm at the same time, the intern is going to go, I will try. But if you give the intern three examples, hey, these three are great newsletters, just right. In the same manner, the intern will know how to write in a way that's a lot more similar to what you're looking for. And it's less abstract. So attaching examples typically works a lot better than just giving a description.
- [36:05]
Michael Stelzner
Okay, so there's some people here that are saying, mike, Audrey, that's a lot of work and it is. But what's your thoughts about using projects inside of ChatGPT or Claude to kind of have it there so you don't have to do it every single time? Versus custom GPTs, just because obviously you can upload this data and have it there? Like, what's your thoughts on that?
- [36:29]
Audrey Chia
So I like to tell my clients, right? What is your objective? Like, don't use AI for the sake of it. So if you are thinking, should I use custom GPTs or projects? I always ask, what is your objective? Right. For every client, I usually start with a project because think of it like a folder, right? It's my knowledge base. I can store a lot more information compared to a custom GPT and I can always ask AI to pull certain information out. Whereas for a custom GPT, think of it as an intern, highly skilled in one task. So if you want to build a GPT for LinkedIn, post writing, for ad post writing, for newsletter writing, there will be three different GPTs. But the fun part is you can build your custom GPTs and give it specific knowledge. So for my LinkedIn GPT, it knows my LinkedIn post examples, my newsletter GPT knows my newsletter examples, and then what I can do is to have a project folder with a lot more of my knowledge with regards to AI copywriting content strategy. Then the magic is I click into a project folder and then I call one of the GPTs in the folder. So now I get dual purpose, right? Because this GPT knows how I write for LinkedIn, but it also also can tap onto knowledge in my folder. So it pulls new information and crafts content with it.
- [37:50]
Michael Stelzner
Yeah. Tell people how they can call a custom GPT in their project. They might not know how to do that.
- [37:54]
Audrey Chia
Yes. So many people may not know this because this is like a super secret but underrated hack that you should try. So if you go to your projects and you just start a new chat, if you use the function, meaning the one you use for emails, you just use the add function and you type in the name of your GPT, you should see it pop up. So you can actually call on any of the GPTs that you have built or any external GPTs that you are currently using, and you can call it in your project folder. And this is a very handy tool. What I love doing is stacking the different GPTs. So imagine I have a project folder with a campaign brief. I can call my Facebook ad GPT to write a Facebook post. I can call my Google Ads GPT to write a Google post and then I can call my LinkedIn ad GPT to write a LinkedIn ad all in the same project folder, tapping on the same sources and information.
- [38:52]
Michael Stelzner
So just so we're clear, for people that do not have clients, what's the advantage to the project folder and the linking of the custom GPTs? Is it because you can have special knowledge that runs across all the custom GPTs and somehow it'll share the knowledge? Is that kind of what I'm hearing you say?
- [39:09]
Audrey Chia
I think from most people, if you don't work with clients, then you may or may not need projects in the first place. So if you don't work with clients, then you can even have one folder with many things. But I have seen my own clients, although they don't have external clients, they can have 10 different folders for different things like products. So they can have. Yes, because there are so many use cases. Right. Even in your Google Drive you would have 10 different Google Drive folders. It's the same way with chat and how you organize your thoughts.
- [39:41]
Michael Stelzner
Got it. Okay, so up to this point we started with strategic positioning, right? Which was this insights and data. Then we've got this tone of voice, right? Which is how do you want the AI to sound? What's next? Because I feel like there's something else we need to talk about here.
- [39:55]
Audrey Chia
Yeah. So I think the final step, whether it's for a landing page or any other content, is your structure and output. So what I found is that a lot of people don't know what is the end result they're looking for. They don't know what a good landing page look like. They don't know what a good ad looks like. So when AI produces something, they take the information as is, which is the first problem that we mentioned. So if you were to string everything together in an ideal world, you would tell AI what is the structure of the output you are looking for. In the case of a landing page, there could be so many different landing pages. Like what? Michael, you said there are the sales landing pages, which are long form, copy heavy, no use of visuals. Then you also have the brand pages like Nike. You know, they might have a lot of visuals, short punchy lines, not a lot of copy, more motivational quotes. And then you could have, you know, the in between the conversion copy says landing pages. Right. So identify which type of output you're looking for. Then tell AI, hey, this is the structure of the output I want. Or tell AI this is an example of an output that I want but now can you take all the knowledge from what we have done so far and then condense all that knowledge and recraft this landing page according to the output that I am looking for? So the more specific you are, the better the results. So don't let AI imagine what a landing page could look like. Tell it what you are looking for and then it can get you to the end result so much quicker.
- [41:34]
Michael Stelzner
For people that are not copywriters like you and me, they might not even know what a good example of a high performing landing page is. Can AI help them find examples? What's your thoughts on this? Because not everybody really knows by looking at a sales page whether this is high converting or not. Right. Or whether this is moving people in the direction that is the desired outcome.
- [41:55]
Audrey Chia
Yeah. I think that is also why copywriters still need to exist.
- [41:59]
Michael Stelzner
There you go.
- [42:00]
Audrey Chia
And that's also why you need to reach out to folks like the both of us. Right? Because at the end of the day, AI can give you 100 headlines. Do you know which one of the hundred is going to work best? Of course you can ask AI, which one do you think works best? But there is a high likelihood that as a trained copywriter or someone who has done this and put so much time into honing their craft, we have an eye for certain things. In the same way, a designer will always be the best at prompting for design. And when they look at something, they know this doesn't look right or this needs to be shifted somewhere else. Whereas I could ask AI, hey, you know, what's wrong with my design? And it would say, oh, this looks fantastic. And they may not give you the right input that you need. So I think that human touch is where it comes in. So if you don't know what good copy looks like, I strongly suggest that you also consult some of your human friends or experienced writers or just do a bit of learning yourself. Read up research and don't just take AI's answers as is. They may give you some good results, but you need to be also discerning about the quality of content it's producing or the quality of work that it's telling you, hey, this is good. Because good to AI may not be good to humans.
- [43:19]
Michael Stelzner
Yeah, I love that. I also think that it's very possible that you who are listening have purchased a product that you felt compelled by the copy on a product that's in a totally different category, for example, and you could say, hey AI, go take a look at this page and reverse engineer the structure of what they're doing here and model the output after it. What's your thoughts on something like that?
- [43:46]
Audrey Chia
You can try that, but I would say always use that human lens. So when I have done reverse engineering like this, even as a trained copywriter, I realized that AI can get you 50% of the way, but it's missing a lot of things like certain psychological tactics, specific pain points, sometimes it exaggerates things. So all of these are content nuances that actually make or break the page. So yes, you can publish something. Is it good? I'm not sure it's the same way AI can create a logo or a brand video. Is it good? That's something for us to really look into after.
- [44:22]
Michael Stelzner
I love this framework. It's been so amazing. Audrey, if people want to connect with you on the socials, where do you want to send them? And if they want to potentially work with you, where should they go?
- [44:32]
Audrey Chia
Yeah, absolutely. So I typically work with startup founders or small business owners. If you are struggling with with building a brand that converts if your landing pages ads LinkedIn isn't working, reach out. We can get it fixed and sorted. So you can find me Audrey Chia on LinkedIn. You should see a little mini rocket beside my name and you should see my signature red color so you can find me there. Or if you would like, you can visit www.clotheswithcopy.co Audrey, thank you so much.
- [45:03]
Michael Stelzner
For sharing your insights with us today.
- [45:04]
Audrey Chia
Awesome. Thank you for having me.
- [45:06]
Michael Stelzner
Hey, if you missed anything, we took all the notes for you over@socialmediaexaminer.com a84 Be sure to follow this show on your favorite podcasting app and if you've been a listener for a while, we would love a review. Let your friends know about this show. You can tag me on Facebook, LinkedIn and or X and do check out our other shows, the Social Media Marketing Podcast and the Social Media Marketing talk show. This brings us to the end the of of the AI Explored Podcast. I am your host Michael Stelzer will be back with you next week. I hope you make the best out of your day and may AI help you become more successful.
- [45:39]
Audrey Chia
The AI Explored Podcast is a production of Social Media Examiner.
- [45:47]
Michael Stelzner
Get your tickets to AI business world right now by visiting AIbusinessworld live.