Crafting Podcast Cover Art With AI: A Guide To Precision And Impact
7 Dec 2025 · 44 min
Podcast episode (Tom Hazard & Tracy Hazzard) on using AI to create better podcast/episode images. Main points: use unique episode images (not repeated show art), prioritize visual impact to stop scrolling, summarize episode content before generating images, tell the AI desired style/format first, craft prompts broad→specific (style/format → content → fine details), iterate and save prompts/images, and prefer AI partly to avoid stock-image copyright headaches. Tracy reports workflow timings (initially ~30 min, now ~10–15 min) and that their team has purchased many stock images historically (~50,000). They plan to adopt AI in production by year-end.
Chapters
Want your podcast and videocast to be seen, heard, found and rewarded in this noisy digital world, you are not alone. Join the nonprofit community at Podcasters United and podcast industry experts Tom Hazard and Tracy Hazard as they debunk all that outdated and bad advice you've been getting from the podcast gurus and find out what actually works. Today, get ready to finish feed your brand with smart cut tactics proven to cast your original voice to the world.
Hey, everyone. Welcome to feed your brand. I'm Tom Hazard, along with my co host and better half, Tracy Hazard. And today we're going to talk about how to write a better prompt to get better AI images for your episode art. Now, of course, that means and implies you're going to use AI to generate an image. And this is really where the. The entire industry of stock images, this is where things are going, especially with the complexities of copyright laws and all sorts of things. And I'm very excited about it. You know, we've been testing this and you know it. Depending on your level of experience in AI and the kind of images you want, it may be the right time for you to start doing this, or it may not be ready for prime time yet. But it's coming, folks, and no one that I know knows about this better than my co host, Tracy. So, Tracy, let me throw the mic to you here.
Well, I mean, I think the old adage a picture is worth a thousand words is really critically important and it's a sorely underestimated heart of both podcasting and content creation in general. From the blogging side of things, are.
You telling me I need to write a thousand word prompt to get a good image? Tracy?
Absolutely not. You should be able to do this a lot less. But that image, that, that image sends a message. It is the first thing our eye sees before we ever see any of the words in any post is the visual. And so if you're sharing something on social, if you've created a blog, if you've got a podcast with an episode image and are stop using your show image for every single episode and start creating those custom episode images which you.
Can use, which you absolutely should be doing as a podcaster. You really should, because otherwise every episode visually, same, same, same. Seen it, seen it, seen it. Get. Use a unique image for every episode, please.
So here's the thing. There are branding experts out there that Tom and I disagree with and we disagree with this, reminding all of you that Tom and I have art degrees. We have degrees in design, and there is a reason why we Disagree with the over branding, over stylization.
And it's also, Tracy, informed by a couple decades of business experience. It's not just, yes, we have art school backgrounds, no question we understand what makes a beautiful image or an effective image versus not. But it's also really the end of the day. It's for us, it's less about the artistic nature. For me, it's more about the functional nature of an image. Is it motivating the viewer to take an action? That action we want them to take is to play your podcast episode, to.
Play, to click through, to go and do something, to take an action. So we have to stop your eye. And that's its purpose, is to get you to stop. So when we see the same images again and again and again, yes, there is some brand reinforcement, but it is expensive use of space to you to do that. When you're a big brand, you can afford to slap your logo and your image on everything. But when you're small, you need to capture the attention at the moment and get them to take an action, to come listen to you, to come read your stuff, to come find you, to click through. That is its most important role in the visual. It is not to reinforce your brand. Brand reinforcement comes later when you have more money, when you have more business, when you that is its own purpose and its own campaign and you have its own budget for it. In fact, you should have a separate budget for brand building from content building. And that is what we're talking about here, is content building creation in any kind of way, whether it's a blog, a podcast, a video, or. And the images that go along with that is what we're going to talk about today. So when going back to what I said about a picture worth a thousand words, the important part is that no one will get to your thousand words, will listen to your show, will watch your video, will read your blog, unless that image stops them in their tracks and gets them to stop scrolling through their feed and click through it. So a visual impactful image is the most important thing you can do to make your content seen, heard and found. Now, it would be really easy for Tom and I in our business and potatoes for us to make the same image for everything. It would be so much simpler and so much cheaper and so much less headache because designers don't always create. Like some of our clients hate the color orange and others of them love it and want everything to be orange. And so there's preferences in here and visual language that does not exist for most People, most people, people cannot express style and design and visual language and tell you what they want. They tell you what they don't want. And so there's more complaints on the side of that. So it would be so much easier if every, if repetition was the way to go, but we would not stay in business with our clients coming back week after week publishing episodes with us and producing them with us. If we did that, we would be doing them a huge disservice. So anyone who tells you that is trying to make their own life easier and still charging you for it.
Is exactly that. It making their lives easier. And unfortunately, people's egos play right into this because the sort of rookie error in podcasting that a lot of podcast coaches, far, far, far too many, who either don't care or are unaware that putting a host face on a cover art is not going to help them. And I know a lot of you have your face on your cover art and you're hearing me say this and you're like, what? But that's a whole nother episode. And, you know, reach out to us at Politize, we're happy to talk to you about that. But the reality is that it's easier for a podcast coach to say, oh, we need a cover art for your show. Sure, you're the host, where's your headshot? And then it's easy for them to make the COVID art using you as that image. Of course, then that makes a show all about you and not about what's in it for the listener. And that's just a one clue.
Right? But it's also the episode art. And that's really what we're sort of talking about here because it's content related art that we're talking about creating originals. And there's also the prevailing idea that you stick your face and your guest's face on it. And then it just makes it easy because it's just a template and the faces have changed, but it's not enough. Most of us cannot distinguish the difference in those faces. If those faces don't have the specific, specific energy I'm looking for, if they're not recognizable to me, there is no reason for me to stop. So you are not going to get the impact that you think. Now there are a lot of people out there who go, if you put faces in your social media images, you're going to get more value. And it's true, but it's not, it's not true. In the case of content, click through. So in other words, Getting someone to click through the content.
Plus, if you scroll through someone's LinkedIn feed, Facebook feed, I mean, their individual feed, not just your feed where you're getting stuff from everybody. But if you go to someone's Facebook page, their LinkedIn page, and you scroll through all their own posts, and you have that exact kind of templated episode art, Tracy was just talking about where you have the host face and the guest face, and we've had lots of podcasters that do this. And people just keep scrolling because even though there's a different guest face in there, even though in small print there's a different title, maybe on that image, it all looks the same. And people process images. Their brains process images much faster than they read and interpret titles. So they just keep scrolling because they think, oh, seen it, seen it, seen it. Nothing new here. It doesn't catch their eye. It doesn't make their brain stop and say, ooh, what's that?
So the exception to that is a video still, okay? If you've got, like, in the short and the video's playing and you have an image of a person, it's gonna actually do better than a branded still. So in other words, if you've got your logo and stuff all over the still part of it, that's less likely to get clicked than an image of the person just about to talk. And so because you're like, oh, what are they going to say? They look like they're about mid sentence. So choosing one of the still images, if you're. If you're not going to create a custom one for every YouTube thumbnail or any of those things, or in YouTube shorts, you don't have a thumbnail. It's sort of like picks. It picks a frame. So it's better for you, if not that, if that is not branded and stylized, because that way it will work better than if you did the overly stylized one. So just keep that in mind. But here we're talking about AI. So I want to talk. Get back to that. Like, why are Tom and I recommending AI images when we are designers and we have a team of designers that we pay quite a lot of money for. So why would we be saying use AI?
Long, hard, painful experience in this industry is my answer. What's your answer, Tracy?
Well, two things. One is that our clients are not willing to do the work themselves. So they're not picking their own images, they're not determining their own graphic. And a lot of you aren't. A lot of you have virtual assistants in another country, you don't have a team that truly understands you and your message at a deep level. And because of that, the choices that they're making are not your choices. And sometimes they're not even understanding what it is that they're designing. So if you are hiring someone on Fiverr, for instance, and you put in, say, I need a cover image for this, and you describe it and you leave the link to the video or leave the link to the podcast or whatever it is, they won't listen to it. They're only going by the prompt you have, because they need to spend the least amount of time creating this image, because that's how they make the most money. So they're never going to deeply dive and understand your content. And even in our own company, we have that issue, is that we, our graphic designers, we had to stop the flow of allowing them to create the designs at the beginning. They have to create it after the. The copy team does their work. They title it, they give it a leading paragraph with a big essence of what it's about. They have headings for the sections, and they give it that amount of information.
Because the good quotes that can help you get at what are the key elements of the podcast, right? I mean, there's lots of ways you can generate this sort of stuff. So you get a roadmap for maybe what might be a good image.
Right? And so that's what we were starting, we did, was we shifted our model of when it happens in the process so that we could force at least a certain amount of understanding, somebody in the process understanding it, so that it was guided correctly to begin with. So that's. But we learned really fast that if you are not going to do it yourself, it is not going to come across as you. And so AI actually does a better job of reading the entire thing, of understanding the entire thing, so it has a better shot of doing that. So that's the first reason. The second reason we've chosen AI is simply a copyright problem.
And that's the pain that I was referring to, Tracy, which has really. It hasn't always been this way. It's really, in the last two years, become a painful annoyance for podcasters and for us as a company doing done for you production, where we license the images on behalf of our clients from, you know, many different stock photo libraries.
Right. So look, Tom and I have spent over 50,000 images we've bought over the time of our business for our clients, for ourselves. So we've bought a lot of images. So there is no way you can accuse us of not supporting the stock libraries. We were probably one of the heaviest users. The problem with it was, is that even though we legitimately bought them, the photographers, the creators would dispute them anyway and cause all this headache of work. We have a whole episode on that.
Where Tom and I tirade about it. Yeah, we don't want to go off on the deep end here, but basically it's on you, the purchaser of an image or the user of an image properly licensed. Even though. Right. To prove that you properly licensed it. The author of the image doesn't have to prove, hey, you use an image and my research says you don't have a license. No, no, no. They complain first, cause you great headache and then you have to spend the time and go produce your license and satisfy them or another company on their behalf or else you have to pay a huge penalty. And this has become such a big issue in our company. We've had to put people on this full time, not every single month, but some months full time just to deal with it. And that is what's driven us to AI images or made us much more receptive to the idea.
So that's part of it. The other part is AIs ruin the libraries. So because these photographers, these creators are now able to do more searches on what's the most searchable topic title. And so they're overly filling specific keywords and not others. So we can't get enough original artwork in areas that we need it. So it's just not working anymore. And so what we started to do is experiment with all the different AIs. Now there are some, and I'm not going to teach you one here. That is not what this session is about. There are tons of them you can use. Canva has one, Adobe has one, Google has one, ChatGPT has one. Of course. Right. So like they're all there available to you to use. But what I want to talk about today is how we use it to get what we want. Because as I mentioned earlier, most of you do not have the good visual language that you need. It's not a requirement for you to study art history in most colleges. We did. And I might have slept through a lot of those classes, but we still absorb some. Right. So not my favorite seminar class, but. But I do remember enough about it to know what the art periods are. I do remember enough about it. How to describe a style of art or describe style of design. Right. Even though that's not my area, my area of expertise is Textile design, by the way. So when we know we have a better visual language, what I have is an excellent color language. I can describe color to anyone. And so that is my expertise area. So it's one of the places that I've played with a little bit in the AI and we discovered some things that are really important with how you prompt it. And here's what I really discovered, and I think this benefits all of you out there, is that the less information you give it when you don't know exactly what you want it to look like. See, my problem is, is when I go to create a blog header, I already have a visual idea of what I want to reintroduce the image that reinforces that message. I already have an idea in my head because I'm a designer, so I have a visual, but most of you don't. So here's the key. If you take your entire podcast episode and you summarize it, get a summary. And not the ones that say, in today's episode, you want to get a summary that says, describe the essence of this episode based on the transcript you read below and make it compelling for someone to watch it, listen to it, read it, whatever it is that you're going to turn it into, that's what you want to ask the AI to do first. If you throw, and I've seen it where people will throw an entire transcript in, and I've seen the shortcuts, I've seen the tiktoks about it where they tell you to put your entire podcast episode, your video, whatever, just into it and tell it to create an image for it. It's not good enough, because it didn't. It decided what was important in what you said, because I guarantee, you meandered over the course of me, right? They could have decided easily that Tom and I, because we were so energetic about the copyright issue, decides that that's actually what it's about, instead of deciding that it's about image and creation and artistic and visual impact. Right? Which I really hope that the visual for this is visual impact. So I don't know what it's going to be yet, because we'll create it next. But the idea of it is, is that if you give it the summary of what's important and you agree with it, because now it's only a paragraph for you to read. It's not the whole transcript, but it's like a little paragraph and you agree with it, edit it. If you don't, and you drop that in and ask it to Create an image and be very clear about two things. At the beginning, you're going to ask it for the stylistic type. So you're going to say, I want a landscape image or I want a realistic image, a photographic style image. You're going to tell it what type of image you want and perhaps its format. You want a square, you want a rectangle. What are you looking for? Right? And then say based on this, and you're going to insert the paragraph, let it come up with ideas. Because you don't have that visual language. This is of great benefit to you because a lot of times when we use the stock library, it's keyword crammed. So they know that real estate and money are going to go hand in hand. And they cram that into images that have nothing to do with real estate and money. And so it means you have to search through, sort through. You get frustrated and you pick one in the beginning. And I guarantee you 20 other blogs already have that today. So it's, you know, probably thousands of blogs have that same image, right?
Oh my gosh, I remember there's an image we used many, many years ago in our business, a stock image on some promotional material that we gave away at one of the early trade shows and such. And I see that image used over and over again by other podcasting related companies or other companies that talk about speaking from the stage or something. Some of these images really get overused and this is now an advantage, I think, for AI.
Yeah, absolutely. And especially when you're talking about in print media and other areas like that. But for this purpose here, when you don't have the ideas, when you don't have the visual language, letting it generate something is really good and important. So I'm a big fan of the Adobe image creator. I should really like how it works. But a lot of you don't have. Adobe Canvas works fairly similarly if you want a low cost one. I actually like shop Shutterstocks, because Shutterstock, you can use a image to start from and then AI edit it. So you can have a hybrid between supporting creators and making it custom. So that's kind of nice as well. So like, you can remove people from the image, you can flip it, you can change the colors, you can do a lot in it. And Shutterstock is a fairly simple, easy way to do that. So those are some of the ones that I just, I support and they're easy to use. But in can, in Canva and in Adobe, the. One of the major things that happens is when you start to refine from one of the images. So if you let it generate these ideas and you go, ooh, this one's kind of going in the right way. It needs some tweaking, it needs some adjustments. I need to correct it from here. But you go into that one, it's now learning what you like, and it's now learning what you're interested in, and it's now going to take it down another path without you having to revise the prompt to keep it going. So, like, it's now like, gone off on a branch, if you want to think of it like that. And so that's what I really like now in that section of doing the branch. So now you're branching off. You're coming off of something you like. You're going to build off of it. Now, here's where we get more specific. And this is the most important thing to remember. The order of details that we think are important, the order we put them in. So if we say, I want a brunette woman in my image and you put it first, then the AI is treating that as the most important thing. Instead, you might want to say, hey, I want something in the pointillism style that is a landscape image for my blog. Those are really the more important thing that you want. So, you know, keeping it up in those things. And then start to say that feature a brunette woman lying in a field of daisies, bright yellow daisies, like, make them unreal, or bright pink daisies, like, make them totally different, you know? And because it's the pointillism style that I requested at the beginning, I'm abstracting it. So get into the details of what you want in the frame last, otherwise it focuses on that.
And so what I'm hearing is don't let your own biases or preferences influence the order of the prompts, because you're probably going to end up with a lot of images that are all the same or not quite hitting the mark as well as you would want it to. You kind of really get yourself out of it and make sure you are crafting prompts that are truly more meaningful in alignment with the outcome you want, not your own biases. Would you agree with that, Tracy?
Yeah, exactly. And. And what we have to remember is that sometimes we just have to rearrange the words in our prompt. It's not that the AI doesn't get it. It's not that it's wrong. If you just reorganize the order of things, all of a sudden, it will clarify it and it will understand and that has to do with that order of importance, getting a weight in terms of what it can accomplish. So it's like we're asking it to do too much in an image that it has to decide how to prioritize it. We think we're being really clear about it, but we're really not.
Yeah, the AI can't read your mind and filter that out. You've got to really. And this is why we're talking about this, because there are different ways to do this. And while there may not be an absolute right and wrong way, I do think there is a good and a better way.
Right. So think about it. Get going from the broad to the small. So go to the finer details last. Now, some people will tell you to go to the style last. I say no, I say if you're really hot on having a pop art style or a pointillistic style, an abstract style, just start there because you're going to get better, a better visual to start from. And you're also going to make sure it understands when you say abstract, what you mean. Right. And so I think that that's. I personally found that for those without a deep understanding of art history and the art styles in general, that's an actually easier way to go, is to start from one that you think you like or you think you would like to like to address or like to utilize, and then go into that type, making sure you're describing what it's for, because it is clear to the AI in the most case. The Gemini one, especially in Google, understands when you say this is for a blog, it gets what you mean by that. It understands what good blog images are. Not every One of those AIs are good at that, but Gemini is in particular. And then go into your more descriptive language about what you're looking for about, you know, the. Like that paragraph about what your content's about. Right. And then in last thing, as you get to it and you start to get to something you want, here's where you might want to play with it, with color and style. So say, take this image and show it to me in an abstract style. Show it to me in a photorealistic style. Show it to me in. This is where you can start to explore and play with some of those options and ideas and see what happens to it. Make it all. Make it tonal blue, you know, to give it a feeling and a sense of color and space. Or I was wrong about the yellow daisies. Definitely make them all pink, you know what I mean? And so change the color tone of something, because color has one of the most big impactful things you could do to it. So creating a color impact difference. And this is something that we've always done even when we used more templated versions with our clients, because they were more comfortable in that template world, we still varied the color and style from week to week so that it wouldn't be exactly the same. There was like three different ones that they would rotate through so it would have a different sense. And so often the clients were uncomfortable with the different colors, like they couldn't stand it, but it's really what made it stood out.
And that's because of different accent colors may not be in their brand book or their brand guidelines from the people that created their brand identity. And that's where you gotta strike some harmony between reinforcing your brand identity and being recognized and standing out, grabbing people's attention. And that is a tricky, you know, harmony to achieve sometimes. So, Tracy, a question came to mind as you're describing, about this process of the images and then changing the colors or the styles and such. What if you going through this process, and I'm talking more from a practical perspective, because you know this better than anyone, you know that I know is that when you've gone three or four steps down a path and that ends up getting off track, off the rails, not what you wanted, and then you want to go back to that image that was three or four ago, can you get there easily or do you have to be saving out images as you go to make sure you preserve an earlier generation?
It depends on the AI, unfortunately. So, like, if you were using Adobe, you get about 30 images or so before it requires you to actually start again. So what I typically do is do three passes. So I'll end up having generated 90 images by the time I'm done with it. But what I do is I weed out and delete the definite nose out of the IT so I can keep the image count down and hit the 30 images. So I'm cultivating the ones I want to go back to, and I will get a set of 30. In some of the other ones, it overwrites them. And you.
That's what my concern is, that it might be overwritten. You lose, oh, I had it and I lost it. And you have to sort of reinvent the wheel of all those prompts, right?
Yeah. And so saving some or downloading some, using them as the next starting point. That's why I don't like some of the other tools. They become more difficult to do. It's not like you can upload the image and start from there. Not all the tools will let you do that. So you can't go back if you didn't save it, and that's problematic. So that you're going to get into a place where you do do that if you need to. What you realize is where it started to go off. That's where you want to have your prompts. So here's another prompt tip. Have your prompt in a separate document. So every time you make an adjustment to the prompt, make an adjustment to it. Sometimes I'll use the AI and say, how can I get better results? Ask Gemini how can I get better results in creating an image from this prompt. And it'll actually revise the prompt. Now, some of the tools actually will make suggestions to your prompt. It will already make the suggestion to you as like, refine this prompt to get a better result. Right. So it's already doing some of the work for you. That's great. But now you have to save the refinement back into your document. So you want to kind of keep going. So you learn. And so this is where it gets cumbersome is I usually have a file on my a file folder on my Google Drive or in my on my desktop that is synced to my Google Drive and I have a Google Doc and I have both things open at the same time. So that any work that I'm doing, I'm saving to the drive and I'm saving to the dock. And so I can go back and forth and I can jump back to where I need to to start again. But if you get really off tangent, you clean go fresh. So if you've ever done ChatGPT, you just start a fresh new communication. You stay out of the tab, don't go back into the tab for it. What I have found though is if you're working in an entire tab and you're able to. And Gemini is one of those things because then they've now enabled the image creation side of things, right. It's not available to everybody. Just want to say that out there. That's one of the reasons I'm not showing it to you, because it's probably going to change like 10 times from this video, right? But if you are doing it where you are creating the copy, the summary, you've done that all in one place for your episode going in there, it will have learned from what it knows about that episode. It's actually a lot less work to Ask it for an image at that point, because it's got all the learnings about how you summarized it, how you got to that, what you thought was important, what you edited out, if you were editing the document, it learns more about that. So creating the image from within that space can actually be a really good thing as well, rather than going. Separating it out. But for the most part, most of you have to use the image generator in an another program. And so this integrated way of using it is fairly new.
So, Tracy, it really occurs to me that to get good at this, I mean, it's not, you know, maybe certain AIs have a little different process, little different way you do things, and some may be easier to use than others. But it seems no matter what you choose, there's really some skills that need to be honed in getting the AI to generate the kind of images you want. And you're obviously revealing some of your recommendations and those skills here. You really. It is a new skill that people are going to learn how to, for a lack of a better term, because I actually have. I didn't create, obviously, I didn't invent this term. I've heard it spoken about in a way that I didn't appreciate in the media, but is, you know, prompt engineering, I mean, really there, or in our case maybe it'd be prompt designing. Right. You're seeing as we're art students that and we're talking about generating images at the same time, but there really is an importance and a skill to be developed in how to do it. And I bet a lot of podcast listeners may not want to spend that much time learning how to do this for themselves. Do you think?
Yeah, I mean, I think it is time consuming in the beginning, but I can say, like I. Like I said, I could generate 90 images in Adobe, and by that last one, it's exactly where I want it to go. That Last set of 30 is doing like final refinements, like getting the right details, making sure, like, I want her eyes closed and lying on the field of daisies. Right. Like, things like that would be detail refinement to it.
So how much time would it take you to go through generating 90 images to get to one that you're happy with? That seems like a lot to me.
Initially it took me about 30 minutes till I got to this system. Now I can do it in 10 to 15, depending on how picky I am. I think most people would stop way before I do, but I'm a. The design details I'll notice when Something's not right. You know how I am. Like, I was like, oh, why is that one incongruous with the rest of the photo? Like, you know, why is that different? Because I can see it and most people can't. And so I'm a little.
Do you go with stock photography where you're going to just search and search and search and weed through all sorts of stuff until you find something that you like and that's going to take a certain amount of time, or do you learn how to do this, maybe spend some more time initially and get at the point where in 5, 10 or 15 minutes you can get the image you want? I mean, 15 minutes to me still seems like a lot of time, but maybe compared to searching through a bunch of stock photo libraries when you're not finding something you're happy with and just accepting something that's below the standard of what you were hoping to find, maybe it does make sense to spend the time prompting the AI and then you get an image that's copyright free that you can use and you don't have any of the potential future headaches either. That investment of time actually, given our experience, would be desirable.
Absolutely. Yeah. So for me right now, I think. I think I can mostly do it in about 10 to 15 minutes maximum, depending on how picky I am about it. I think for your average person who might not be as detail oriented, I think under 10 minutes and. But that's worth it because if I stop traffic, if I stop the feed and get them to click on it, because it's an image they've never seen before, it caught their eye, it was impactful, it gave them the message, wow, this is about whatever the subject is that I was trying to reinforce. Then I'm going to get more results.
On social media, on my blog, right?
And get more people to watch our YouTube videos, listen to our podcast. This is the whole point of this, right? And look, I mean, I guess if you are not interested, don't have the time or, you know, don't care or don't believe that a unique image for every episode is going to get you more people to click and try your episodes than you stopped listening 30 minutes ago, probably to this episode. And so, okay, that if you want your episode to really have more impact and attract more people to click it, at the end of the day, that's really what we're talking about.
Using every tool available to us. That's what we're about here at Potatoes, using every single tool available within the podcast ecosystem. Within YouTube's system, within the social media environments that none of us invented, but we are needing to work within. How do we make the biggest impact? How do we stand out? How do we get people to listen to our show and not that other show? This is one of those things right here.
Yeah. Now, this is really early on. What gave me the idea for using individual episode art and all of those things even before the podcasts were. Because we were doing blog posts, right? Way before anybody else was doing it. And the RSS feed didn't allow episode images back when we started podcasting even.
Well, or also a lot of the listening apps, even if you did have a unique image when it was available in the RSS feed, the listening apps ignored it. In fact, only till relatively recently in the history of podcasting, which started, of course, it's getting close to 20 years ago. Right. And I think 2025 wanted to officially be 20 years that, you know, in. In the history of podcasting, it's been only relatively recently that Apple did it. You know, they have now. And a listener was happy to point that out to me when I still wasn't aware because it. It had been a long time. They didn't. And other apps also did not. But now I think the majority, and I, I. It's hard for me to think of an app that does not display the individual image.
I point to what I was saying about it being a unique image is because we were coming out of this blogging and writing side of things as we were approaching it. I was being trained by Inc. Magazine because I ended up writing a column for them. And there were two studies that they used to point out, and the image selection. Now they had a serious license with a very large stock library, and you had to use the licensed images. That was all. And you searched within where you submitted your episodes to search within to find them. And it would take sometimes an hour for me to publish an episode just to get an image that I thought would work that hadn't been used by hundreds of other columnists before. Right. Because that's also the problem is the minute you type in some of those keywords, everybody gets the same one, and it would flag it until you can't use it. And so those were issues. But they would also cite this study that BuzzFeed did, and this BuzzFeed would publish a listicle of like, it was something like the 10 dirty fruits or something like that that, you know, you shouldn't be eating. And so it'd be like, you know, strawberries and like all these fruits and stuff that you should buy organic is basically what the article is about or the listicle was about. And it would have like the fruits and so they would rotate and try every single fruit and they would run all of these test of it of which fruit of the list was at the top of image. And it would nine times out of 10 be the most colorful choice, but not the best choice. And so it showed you that the most visually impactful image was the one that worked and got most people to click into the article, not the one that was actually related to the content the most, which is interesting. So thinking about that, maybe the AI is going to do something that surprises you because we're a little too much caught up in our content and in our brands and in what we're all about and our colors. So we started out the sort of pre call part that is not in the recording here where we were talking about I had a cheer meeting with a cheer moms meeting the the other day or the other night and we were talking about the sign that we create to promote our big fundraiser. And it's in school colors. But when you put a navy blue and silver or gray because that's what it looks like when you print it sign outside on the grass lawn, it doesn't stand out. And we had a whole conversation about should we deviate from school colors? Like it was a controversial conversation. Should we deviate from school color to get people's attention?
I would say yes, because the whole point, Right. Is to attract attention and to raise awareness of whatever it is. Right.
Well, hopefully by the time this blog post publishes for this episode, I'll have my sample ideas.
Yeah.
Yeah. Because I, I'm. We're gonna not print them for a couple months, but I will have the sample idea because I need to generate that for the next meeting because I told them I would show them what it looked like because of course they don't have a visual language.
Most of them, a lot of people if they, if they there. I mean not a lot. I don't know how many, but there are certainly some people and we've encountered a lot of them in our careers who all they can do you is tell you they like it or they don't like it. They can't tell you why and they can't give you a vision for what they would like to see. So you've got to create something and just show them.
Right. So I'll give you the before and after and first off it was like take the flyer and put it on a sign so it has too much words and too much other stuff. So my second version will be very minimalist but very visually impactful and colorful. And so I'll show you the before and after in the blog post. I'll eventually add that to it. It may not make it when it first goes live, but it will within the next 30 days. So it'll be there because I have to generate that image. So I will share that with you. So make sure you go to the blog post for this episode@podcasters united.org and check out some of the visuals that we use now. Keep in mind we are not yet using our AI images in our actual production flow, so we are not using them on our episodes yet. They are coming and it will happen by the end of this year. But this, what you're getting is a sneak peek into my training on how to do it.
I think that training will be very important and very valuable. I think you've already convinced me. Wow. There's a lot more to it and there's a certain level of experience that you need to gain in trial and error to learn what's going to work and what's not. I mean, you're given a good roadmap though, Tracy I think that certainly our listeners can go and and gives a try and have greater success than they would have before. So I think that's great. But I think at the end of the day it's going to be one of those things that follows that old cliche. If you want something done right, you got to do it yourself. Fortunately, you don't have to create your own image from scratch. With Illustrator or Photoshop you have AI tools and you can generate something now through the words that you prompt it with.
So anyway, I hope to see some really cool AI generated images out there. I don't believe in any way, shape or form, even if they end up marked on social media as AI, that it's going to be a deterrent to anyone.
I agree.
So go test it out, try it, check it out and see what you can do with this and share your results with us. Let us know what you've done.
That'd be fun to do a follow up if we get a bunch of images in from people.
Yep, you can reach us anywhere on social media at United we pod.
All right, let's let that be the end of the episode and we're going to sign off on Live Stream. I hope some of you out there enjoyed this. And then next week we'll let you know. We're going to be talking about optimizing blog and podcast metadata for search and discovery. I know a little bit of a technical thing, but boy, makes a difference. And those of you out there in Live Stream, come back and check it out. You're going to really like that. All right, signing off. We'll be here for our clients in Q and A now.
We are, because I already turned it. I already hit stop. It's now up in the top corner. I cut you off. Yeah, but I cut you off before you said that, so.
Oh, okay, fine. That's fine.
It's easier to find the stop and stop now. So even though, like, Zoom keeps moving everything around, that's fine. Yeah.
Okay. So. All right. I know Jacqueline's not feeling well because she's got Covid, so I'm impressed you're still here. Hope you enjoyed this. We've lost a couple other people.
Tom and I have both had it a couple of times, so.
We've had it a couple times. Yeah. No fun.
It takes a lot out of you. So rest up. Don't worry about your podcast. It'll be here when you get back.
And I see Jacqueline, you emailed me while we're on here. I'll check that out and respond. If you have any questions, please enter them. And if not, we'll let you go and get some rest. So my camera didn't crap out, Tracy. Although I just jinxed it. It's probably.
Yeah, now you're losing your voice instead.
I don't know what's going on with my voice. Yeah.
Okay, thanks, everyone. We'll see you next week.
Okay, thank you.
Heard something valuable today, but didn't get a chance to write it down. No worries. We've got you covered. We partnered up with podcasters united.org where podcast hosts like you can get free resources, free masterclasses, the largest podcast tip library assembled, and free promotional opportunities. While you're there, don't forget to follow us on YouTube and on social @unitedwipod so you can get feed your brand live. Until next time, keep podcasting.
More from Tracy


High-Converting Podcast Outros: How to Drive Listener Action, Simplify CTAs, and Boost Retention


Podcast Email List Strategy: How to Capture Listeners, Avoid Social Media Traps, and Build Owned Reach


YouTube Channel Branding: How to Structure Playlists, Hook Audiences, and Build Your Podcast Brand With Fran Asaro


Podcast Website Optimization: How to Avoid Critical Errors, Boost Organic SEO, and Drive Downloads


Inclusive Podcasting and Leadership: How to Humanize Interviews, Build Resilience, and Expand Reach With Kevin McShan


Top Podcast Growth Strategies: How to Vet Guests, Elevate Content Quality, and Leverage Video With Kelly Smith


Podcast Review Authenticity: How to Build Organic Ratings, Avoid Bot Scams, and Protect Audience Trust


Podcast Guesting Alchemy: How to Pitch Hosts, Build Authority, and Drive Long-Term ROI From Mark Herschberg


Intentional Optimism in Podcasting: How to Plan Topics, Engage Audiences, and Build Authority With Andrea Johnson


Multi-Channel Podcast Marketing: Scalable Growth Strategies with Christian Bacasa
