AI for Copywriting by Chelsea Cohen
The Brand Builder Show
The Brand Builder Show
77. Using AI To Gain A Copywriting Edge w/ Chelsea Cohen
/

The rise of AI has been clear for all to see, but most people still don’t know how to use it effectively to grow their business.

The opportunity is huge, but the pitfalls are also equally large.

So, in this week’s podcast episode, we’re joined by copywriting expert Chelsea Cohen to discuss how to effectively use AI to research, write, and grow your business.

We talked about:

  • AI’s copywriting hacks and pitfalls
  • How AI can both strengthen and weaken your competitive edge
  • AI’s SEO challenges
  • Maintaining an authentic brand voice
  • And much more!

Episode Links

Additional Resources

Talking Points

00:00 – Introduction to Podcast Guest: Chelsea Cohen
00:50 – Chelsea’s Background
03:48 – AI Hacks and Pitfalls
06:52 – Can AI weaken your competitive edge?
10:11 – AI-written bullet points and their impact on sales potential
11:48 – Brevity vs. long keyword in copy
14:33 – Copywriting Frameworks
16:30 – Bard vs ChatGPT
17:15 – Brand voice and its implementation with AI
20:44 – Other ways AI can support your business
22:21 – Leveraging AI for Amazon SEO and finding the right keywords
26:04 – Integrating images in the workflow: Writing copy and framing images
27:14 – Future of AI usage
29:16 – Special Giveaway for Listeners

Ben Donovan  00:00
Okay, welcome back to another episode of The Brand Builder Show. And if you’re interested in using AI to fuel your business, keep it growing faster than ever before, then this is the episode for you. That is the topic of our conversation. And to talk through that we have brought back a very popular guest from the show, Chelsea Cohen from SoStocked. Chelsea, welcome back to the show.

Chelsea Cohen  00:21
Yeah, thanks for having me back on.

Ben Donovan  00:23
My pleasure, I think you were episode number three or something like really early on, when we launched this show, 18 months or so ago. This is going to be about episode 80. So, you know, we made it this far to be able to have you back on which is really cool for us. But also, there’s been some exciting things going on in your world. So for anybody that hasn’t heard anything about you give us a bit of a background on who you are, but also, you know, what’s happened in the last year and what you’re up to now?

Chelsea Cohen  00:50
Sure, yeah. So started selling on Amazon 2014. And, you know, really have kind of come a long way in my journey there. Learning, obviously, you know, we have our businesses and we learn different things. And then oftentimes, we like to share those things. So I can give it very, very brief what SoStocked you know about SoStocked, but then also I want to since we are talking about AI and talking about copywriting, there’s a piece of my journey that I don’t often talk about, specifically pertaining to copywriting. 2018, I built a software company called SoStocked inventory management. We ended up selling to a company called Carbon6, which some people have started hearing about. And so currently, my biggest focus is on inventory management profitability through operations. But I have a sweet spot in my heart and a passion for the copywriting side of Amazon as well. And that’s where I really started in terms of sharing things I know with with other sellers started speaking in 2017 on stages, podcasts about copywriting, and really about how to write for Amazon, because a lot of people really don’t know how to write for Amazon and don’t know that they don’t know how to write for Amazon. So that’s, that’s what I love speaking about is, you know, Amazon is a comparative buying platform. And so that needs to be taken into account. And, you know, one of the things that I’d love to bring to light as part of this AI conversation.

Ben Donovan  02:34
Yeah, absolutely. And it is a big conversation that is continuing to move forward. And there was obviously a big spike in discussions around it when GPT, four came out, I imagine the same if they sort of bring out the next iteration of it, but it’s already very, very powerful. And, you know, I think every Amazon seller would be aware that they need to write a listing that converts but maybe don’t know how to. And so, is this whole AI thing, an opportunity to bridge that gap between knowing how to doing something and doing it? Or do you actually do you need to know what you’re doing to use AI?

Chelsea Cohen  03:09
Yeah, I think it’s, I think it can be a slippery slope. I think that we need to know how to use AI. Because, you know, one thing is for sure, you can’t ignore it. The people who ignore it are gonna get steamrollered by the people who find ways to leverage it. 

Ben Donovan  03:26
Absolutely. 

Chelsea Cohen  03:26
However, at the same time, you know, if a person goes too far into AI without understanding its weaknesses, you can lean too heavily into technology and take the human element out. And then you’ll just kind of be in a wash of sounding like everyone else.

Ben Donovan  03:48
Yeah, yeah, definitely. Let’s talk about some of the potential hacks that you can use for AI. And maybe some of the pitfalls as well then to sort of balance out because I know from using Chat GPT myself, there are days where I feel like oh, my gosh, this is gonna revolutionize my workflow. And then there are days I want to just like smash my head against the screen because it’s not doing what I want. Right? Like, yeah, normal frustration. Are you finding other people are finding that out to you? Someone?

Chelsea Cohen  04:15
Yeah, definitely. I think, you know, some of the benefits of AI are the ability to take a lot of data and condense it. The ability to fight out of blank page syndrome, especially if you’re not a natural, natural writer. How do I start or how can I have an outline you know, it’s always great to start with an outline and then edit from there. brainstorming ideas, social media posts, email campaigns, you know, you can say hey, give me a for you know, a force it for email sequence for someone who just signed up for my email list or something like that. So there are a lot of benefits to that the other one that is really great is using a bar because there’s a lot of focus on chat GPT. But chat GPT doesn’t scan the Internet, it has the limitation of being only as old as, or has data only as as old as September 2021. So there’s a huge gap, you know in its knowledge base. So Bard comes in. And that’s something that you can use to do things like taking your competitor’s listing, and plugging it in and saying, you know, give me a summary of all of the reviews, what does it do? Well, what does it not do well and why would I want to buy this product? Or comparing your listing to another listing? Say, tell me, what’s the difference between these two products? Why would you buy one versus the other?

Ben Donovan  05:44
Yeah, definitely. I think the example he gave about the email sequences is fine as well, I actually, funnily enough, a couple of days ago, used it to create, like a three-email sequence for a product launch I’m doing for pet product, and one of the lines dedicated, like, sometimes it can be so creative, like I gave it quite a detailed prompt, prompt, you know, I told it, that you’re expert copywriter and talked about all of the different things we wanted in it. So it took me a good five to 10 minutes to create the prompt, but then it literally outputs the three emails and included lines, like, you know, Hurry, this offer is going faster than a kitten chases a laser do just stuff. Just nice. It’s so much more creative than I sometimes give it credit for. So I think that there’s, you know, there are lots more ways of using it, maybe then we are doing so far.

Chelsea Cohen  06:31
Yeah, yeah, exactly. So there’s yeah, there’s things like that, you know, that can really help to cut down a lot of those times and also save money on, you know, paying someone else to do certain things. And then there are other things that, you know, you just shouldn’t give that completely over to AI.

Ben Donovan  06:52
Yeah. On that note, are there ways that using AI can weaken your competitive edge?

Chelsea Cohen  06:59
Yeah, so by definition, AI is derivative, right? It looks at other sources, and it uses those sources to come up with something new so you can start sounding like everyone else. And a lot of the space already does. There’s a specific format that everyone uses on Amazon, which is interesting because I’ve talked since 2017, I’ve talked against this format. But this the format of doing a subheading, right, let’s say, strong material or something like that. And then you go into the bullet point. And there’s this formula of, we’re gonna do you know, all caps tell you what the bullet points about, and then, and then do a semicolon or a dash, and we’re going to write about it. And so you so listings already have a standard template, that is actually not the most effective way to write your listings. And we can talk about some of the things that we use the techniques that are used, but we’ve already got where we sound like everyone else. So the thing that’s coming now is that we used to have a competitive edge, say on foreign language sellers. Now foreign language sellers can start sounding like us so that is a differentiator. If you sound if you sound good, and so does everyone else, well, then now we have to start sounding great, just like our images, just like our EBC. Why does someone click on yours versus someone else’s? Is it because your image looks better? Right? So so why don’t we look at that, in terms of we have to make our bullet points look better? 

Ben Donovan  08:47
Yeah. You mentioned that structure would be good to hear that because that’s the structure that a lot of people would know and use. And what are the downfalls of that structure?

Chelsea Cohen  08:56
Yeah, so you know, a lot of I remember at an event. I think it was 2018, someone got up on stage in front of 3000 people and said, people don’t read bullet points. And he was talking about images. So 3000 people write down, people don’t read bullet points. And then you have a whole bunch of sellers who are not looking at copywriting as a tool. And at the same event, I have people coming up to me saying, you know, I loved your talk, or I you know, use your service and I increased my sales by 10% by 15%, whatever it is, you know, and you have this ripple effect as you start to convert, your traffic increases, and your ranking increases. And so it can really have a huge impact. But if we only look at it as a place to put our keywords and to give kind of generic features and not really sell and sell in a way that is compelling, then you know, we’re basically we’re switching good for good. And that’s not good enough. So of course, I’m going to say people don’t read bullet points because I’ve never written a bullet point that someone wanted to read.

Ben Donovan  10:11
Do you find that AI using AI to write those bullet points to write your descriptions that can diminish the sales potential of that copy? Unless you really heavily edited?

Chelsea Cohen  10:21
Yeah, it’s, you can get, like I said, the foundation, but the techniques, some of the techniques that we use are things that helped to combat that. Because there is a truth and people don’t like to read things. Blogs, I’m guilty of myself, I open up a blog, you know, says oh, the top 10 things that you need to know about whatever. And so what do we do we go when we skim for what those 10 things are, we don’t read through the whole thing, we just look at, you know, these the bolded sections. So this is important, that’s important. So how do we take that natural tendency of a consumer on the internet, and give them the ability to skim something? Right now, we do subheadings. And that basically says, um, this is all that this bullet point is about. And since you, you know, you can read the subheading and know exactly what I’m going to talk about. And you’re basically taking that tool away, you’re saying you don’t have to read this. So of course, they’re not going to read it, they’re just going to read the subheading, and move on. But what we tend to do is, first and foremost, we have to get them to understand that we’re going to tell them something that’s going to matter. I mentioned comparative buying platform, people already want to buy the product. So your job is why yours versus your competitors. Right.

Ben Donovan  11:48
On the topic of structure, without wanting to get too much off-topic. Are you along, would you then go for brevity, you know, short readable, rather than long keyword fields, trying to fit as much information in there as possible.

Chelsea Cohen  12:04
Right, so not too short. Because you’re when you’re so it’s so short, all you leave room for is? Is the logic, the features. You don’t give them the emotion people buy on, you know, emotion justify with logic. So what is that product going to do for them? What are they going to feel like when they read that, you know, that listing, and when they get that product, so we tend to like to keep them you know, not going to five, the 500 character per bullet, but around 400 is kind of our goal. We want to make sure that we have longtail keyword words in there. But the way that we guide the AI is we don’t, we don’t create subheadings. Right, we’ll use all caps at the beginning. But we’ll stop it say in the middle of a sentence. And the first, the first few words of the first bullet point, are the most important piece of your copywriting. It’s what there’s a term called hook in a story, you know, it hooks you and makes you want to read more. So we always take the negative reviews of our competitor. And we say things like, you know, we used to say things like unlike other products that been breaking rust, you know, of course, a keyword in there, ours does not because we have this type of material. Now you’re not allowed to compare your products to others. So we’ve had to get creative and say, you know, you want to avoid this situation, or you want to avoid products, or stay away from products that this we don’t necessarily full on. There are ways to change the language without you know, making the Amazon algorithm, man. But that’s the hook and you will take the all caps and you’ll carry it through into a sentence that kind of guides the eye into reading the sentence. And then we’ll take specific portions specific features that we want to have to stand out let’s say you know the material you know, someone is selling plastic and you’re selling silicone, you might want to point out, you know, we you know, we’re not using plastic, we’re using silicone and that might be in all caps. So throughout the bullet points, we’ll have all caps for specific features and benefits that we want to have standout so that if someone just read the all caps, they would get an idea of the value that you’re bringing.

Ben Donovan  14:33
Yeah. You mentioned hooks there. Is that something that you are feeding and like I’ve tested with some, you know, copywriting frameworks like problem agitation, solution, that kind of thing. Are those things that you’re testing feed into chat GPT with any success.

Chelsea Cohen  14:51
So I mean, to some degree, you always have to, yeah. chatting with you. You’re always surprised by what you get. You know, it’s it’s sometimes it can In can get it and sometimes it can’t, we found that oftentimes it’s easier you can feed in a good process would probably be to use barn for example and get the, let’s say the top three complaints or say top five complaints or whatever. And that gives you the kind of the tools to work with. And you say, Okay, well, here’s the top complaints about this, how can I use this to my advantage? And you might be able to ask, you know, chat GPT? How would you write? You know, how would you write a bullet point to evoke this, you know, to compare my product to this product? And you’re always going to have to tweak I think, what, what is being put there? And so, you’ve always got to have that human element, but you can get out of the way there.

Ben Donovan  15:49
Yeah. You mentioned processing data and stuff early on. Is that kind of thing? You mean processing reviews? And that kind of?

Chelsea Cohen  15:57
Yeah, yeah, you can, you can have Bard look at reviews and do a breakdown of reviews, I’ll even put, you know, the percentages of reviews, what are the things that people liked, this would even be really good for product research, take the top, you know, 5,10 competitors that are doing well, and get them to break down, you know, each listing, and then maybe combine that data to find areas of weakness, where you can then source a product that addresses that, you know, it’s really fast, fast research that way.

Ben Donovan  16:30
Yeah. You mentioned Bard quite a lot, are you using it more than ChatGPT?

Chelsea Cohen  16:34
For listing analysis? Because Bard can look, it can actually read the listing, you don’t have to take the listing and copy it. You can copy the link to Amazon. 

Ben Donovan  16:47
Go to the page and read it. 

Chelsea Cohen  16:48
Yeah. And so you can do that with reviews. You can also do that with the listing, you know, break down, you know, the listing, you know, break down the brand voice, what is this brand voice? You know, some tips there,

Ben Donovan  17:04
They just removed that recently, right with, there was a browsing feature on GPT 4, but I think they’ve read it.

Chelsea Cohen  17:09
Yeah, yeah. Recently, and yeah, so Bard has been, you know, been pretty good at that.

Ben Donovan  17:15
Yeah. Cool. brand voice, you mentioned that it’s something that’s key to a lot of people that are building brands on Amazon and want to retain that brand voice. How do you do that with AI?

Chelsea Cohen  17:27
Sure. Ah, yeah. So if you’ve got a rough idea of who your audience is, you can plug in and say, Hey, I’ve got, you know, I sell these types of products, you want to add some adjectives to kind of give an idea, you know, are your products, you know, fun? Are they you know, party-oriented? Are they sophisticated, or luxurious, give a couple of descriptors, and then give a little bit about who you think your product is for? And then say, you know, develop a, an avatar around my product. It can spit out an avatar, you can even ask, you know, what media sources do they consume, you know, and they can give you some ideas of that. And then, of course, from there, you can ask it to build, you know, a brand voice, you know, marketing personality of your brand. And then even ask, if you find specific companies that you really like. Let’s say you’re in the food niche, and you say, Okay, I really liked Bon Appetit Magazine, or I really admire, you know, Apple and what Apple stands for, you can get them to break down, especially if it is an iconic brand. Why does the brand work? What is this brand all about? And then you can go through and it’ll spit out the data so that you can then take that and say, incorporate some of this into what I’m doing and what my brand represents.

Ben Donovan  18:56
Yeah, that’s really good idea actually using real prominent brands that you aspire to and, you know, get an insight from the AI they’re very good tip there. You talking a lot about the you know, add in these words, add in those emotions, adding those modifiers does a lot of information, you know, each time you do a new prompt, do you have some kind of system whereby you’re collating all this stuff somewhere?

Chelsea Cohen  19:20
Yeah, so should have and one of the things that we put together is some kind of shorthand prompts. AI will remember things so you can to a certain extent, layer on in the same chat so you know, my customer avatar, you know, based on this avatar, give me brand voice based on that brand voice give me 10 social, you know, 10 ideas for social posts. And so you can kind of layer that, at a certain point AI starts to get confused and forget. So sometimes you’ll have to then carry some of that over and start a new chat. But if you can take your brand voice that is created, save it, look, look through it, see if you want to streamline it, and then use that when you’re always going in and plugging in, you know what, what emails should I be putting out? Even Customer Success emails or customer success customer service emails, you want them to sound like your brand voice. So you can even create templates for customer success.

Ben Donovan  20:24
Yeah, that’s really good, really good. You’ve mentioned obviously listings, emails, social posts, there are any sort of Amazon-specific or more Amazon-centric ways you can use ChatGPT, Bard or any AI to really improve and speed up your workflow.

Chelsea Cohen  20:40
In terms of, you know, on the listings itself, you mean?

Ben Donovan  20:44
Just any other ways that AI can support your, you know, your Amazon business or your E-commerce business.

Chelsea Cohen  20:49
Yeah. So one of the ways we also have found success and you know, always comes with a caveat, especially when you’re looking at trying to generate fact-based or math based. Math usually serves some other purpose, which there are some dangers there. But things like research, like if you’re trying to find we do a lot of you know, I have an agency for copywriting, we do a lot of different listings. So we’ll get a supplement product, for example. And we want to know about these ingredients. So that generally can take a long time or added time to research, you can go in and research and say, Hey, what are the benefits of, you know, this type of ingredient? Or maybe you don’t have an ingredient that a competitor has? And you said, can say, what are the downsides, oh, it hurts, you know, your stomach, so you can put up, you know, we don’t have this ingredient, because it hurts your stomach. So that gives the idea for the reader, the potential customer, maybe I should watch out for that ingredient. And then you’re in a roundabout way, dissuading them from any of the competitor products that use that ingredient. And it’s because you’re able to get, you know, you always want to verify, because sometimes, sometimes AI will start making things up, we’ve found occasionally, so it’s a good idea to, after they give you, you know, those things to try to search in Google to find some other, you know, sources for that.

Ben Donovan  22:21
Yeah. In terms of keywords, specifically in the listing, Amazon SEO, what are the challenges with that? We talked about sales copy and stuff that converts, but the keyword side of it, obviously is very important there. You know, how do you find AI is working with, you know, finding the right keywords, doing the right SEO stuff?

Chelsea Cohen  22:40
Yeah, that one AI is not that great at understanding what you’re trying to ask it. Especially when it comes to longtail keywords, we’ve done you know, some tests where you’ve asked, Hey, take these bullet points and put some keywords in, and it’ll spit, you know, we’ve had it spit out, you know, a rewrite of the bullet points and then say, here are some keywords and it just gives you a list of keywords at the end that didn’t incorporate, or it will repeat the same main keyword over and over again, you know, we’ve had it, put it, it put it at the end of each bullet in parentheses, not doing you know, so it has a lot of trouble with finding places to work those in. And especially in an organic way. I think that that and we’ve even seen trouble, you know, other people actual people have trouble with how do I take this keyword? I mean, we’ve had to train our writers on how to how do you do that. So you know, I would review and say, Hey, you can make this more organic if you say it in a little bit of a different way. Because longtail keywords generally make the copywriting sound like it’s trying to stuff keywords and sell buyers are more and more getting to be smarter. So anything that makes them feel like they’re being sold, you want to shy away from you want to have them organically read through the listing without realizing that, hey, I’m trying to get your money.

Ben Donovan  24:15
Yeah, and slightly moving away from me that I am without wanting to ask too much of a general question. But obviously, you’re in the weeds with this stuff all the time you’re writing good copy that converts, what would be your kind of overall strategies and tactics that you would share, you know, if you needed to give someone the 8020 on writing copy that drives traffic converts, what would be some of the kind of key tips that you’d finish up with?

Chelsea Cohen  24:41
Sure. So I would say, you know, as we talked about the hook, taking, you know, negative reviews and using them against your competitors in a way that doesn’t make the Amazon algorithm mad. Also, remove subheadings from your listings and you Using using the all caps to guide them into reading the rest of the bullet point, using all caps. And you know, as a caveat, Amazon doesn’t like all caps, they don’t want it. But we haven’t found me, you know, probably a year and a half, two years ago, some people did get a warning email, but we haven’t had people get shut down for using all caps. And we found it converts a lot better. So we’ll give, we’ll give the choice we’ll give, here’s the all caps version, here’s the you know, standard version, and the seller can make the decision. But it is very effective to have your best features and your best benefits in all caps, and don’t go too heavy on it, because then it just looks like a block of text, you want it to read comfortably, but to upon skim selling the point those would be the hook, the removal of subheadings, and then the all caps used to focus on the features and the benefits, or the kind of things,

Ben Donovan  26:04
How much of your workflow is integrating with the images? Are you just writing the copy? Or are you kind of then using that to frame the images? How does that process went for you?

Chelsea Cohen  26:12
Yeah, so that’s what we recommend. You know, we don’t do photography. We have, we do EBC a bit of BBC as well. And the work that is done with the reviews and with figuring out how to, to frame and sell on a motion and to you know, sometimes you have a gifting bullet point, we always end with a brand bullet point at the end to give give people an idea that this is an actual company, you know, especially if you can say small business, you know, I’m a mom who’s trying to solve my own problems, or whatever it is, and taking those bullet points and relaying them as the same story. You’ve done all this work to find the best ways to present your product, and you need to as much as possible, repeat that message to people no matter how they consume content, whether they read bullet points, or you know, simply look at images, it should be very cohesive and say the same thing.

Ben Donovan  27:14
Yeah, definitely. Good. Okay. And then just to finish up in terms of the future of AI, obviously, you don’t have a crystal ball, and you can’t predict all of the future. Where do you see this going? Are we just going to have a just a smoother steady growth of AI usage? Are we going to see some more massive spikes? Or what kind of do you feel for the industry?

Chelsea Cohen  27:35
Yeah, I think that it will, you know, there’ll be ebbs and flows, I think that it is only going to improve, I don’t think I think that there is a danger in some people, like I said, going too far with it. And that, you know, there are people who are going to become AI innovative. And then there are people who become AI lazy. And so that’s kind of the the risk there. But I think that if you don’t learn about it, and you don’t figure out ways to and maybe the creative side of it, there are certain things you don’t want to let go of right, the things that really truly matter, your images, your you know, your copywriting, your messaging, product selection, it can help you to shortcut some of those things. But if you give it wholly over to AI, that’s where someone who’s using it in intelligent ways and maintaining hold on those important, creative, and messaging and brand-driven aspects. You’re gonna be, you’re gonna be overtaken?

Ben Donovan  28:41
Yeah. Yeah, the hybrid approach, I think is so key, isn’t it? And it’s going to be what creates the best content because it helps you to create more, but it also helps you to create it in a conversion sort of focused way. So yeah, no, I completely agree with that. Yeah. Good stuff. Well, this has been super helpful, Chelsea in terms of I think a lot of us get our heads around AI for the first time and how it can really help us in our business. Thank you so much for sharing your thoughts on it. I think you’ve got a bit of a giveaway for our listeners, something they can pick up to help them with this whole process. Where can they find that?

Chelsea Cohen  29:16
Sure. So that’s a catapultconversions.com forward slash AI prompts. We’ve put together some AI prompts for different aspects of your business, and you can get it there. You can also find us on catapultconversions.com. We do a free listing analysis for anyone who’s interested in that as well.

Ben Donovan  29:35
Amazing. Well, we’ll leave the link to that obviously in the show notes and the description so people can grab a hold of that. Thank you again, Chelsea, for coming on. We really do appreciate your time.

Chelsea Cohen  29:45
Yeah, thanks.

Ben Donovan  29:47
Awesome, folks. Well, I hope you got as much out of that episode as I did. It’s given me loads of ideas to test and play with the different AI definitely need to get into Bard and test that out as well. So get into it. Have a little play with it. Chat GPT and Bard are free to begin with so you can have a play use them in your business and keep growing quicker than ever I will see you in the next episode same time next week take care.

Leave a Reply

Your email address will not be published. Required fields are marked *