37: Using AI for Supporting Content Creation Episode Highlights

Welcome to another edition of Using AI at Work. This is a unique episode, kind of a best of edition, focused on a universal need for any business using AI to support content creation. Whether it's social content used to establish thought leadership in your industry, blog posts to engage your website visitors, or external communication about things happening in your company, content has a broad application. So in this episode, we've extracted the best insights from our catalog of episodes to help you start using AI to support support the growth of your business and brand. Let's get started. Welcome to Using AI at Work. I'm your host, Chris Daigle. Each week we'll be learning how today's business owners, entrepreneurs and ambitious professionals are getting more done with smart use of tomorrow's tech. Let's get started. All right, I'm very excited to have Jonathan Green on the podcast today. Jonathan is the host of the artificial intelligence podcast, bestselling author, AI expert, and I had the chance to kind of geek out on a lot of tools the last time we spoke and it was awesome to connect with somebody who's as like tactically familiar with all of the tools and all the moving parts. So we're in for a great treat. So, Jonathan, if you don't mind, take a second and just kind of let the audience know a little bit about yourself and your AI interests.

Sure. I've been a full time entrepreneur since February 2010 when I was fired from what I thought at the time was my dream job. And I was, I was driving home in a blizzard to my new apartment, in my new car that I suddenly couldn't afford. I realized how much power or leverage your boss has over you. And I didn't like that feeling. I didn't want to ever feel that way again. I said, I'm never going to have a boss again. I can't do this.

One of the areas that we talked about this before I launched a podcast. I think, you know, we just released episode 14. You're at episode 300 plus. There is a lot of learning that comes from that many repetitions at something. And that also tells me that you were doing this well before you were AI literate, AI enabled, that you were doing it the old way. And now I would imagine that you've probably found, including ChatGPT, you found some other tools that are supporting the podcast, creation, distribution, all that stuff. I have a personal interest in it. I know that there's listeners that are like, they want to become recognized as an AI subject matter expert in their industry and their domain and a podcast is A fantastic way to do that. So what's the difference between episode one and episode 301 as far as the tools you've been using?

When I started my podcast, it was a solo podcast, not an interview podcast. It was just me. I recorded episodes on my phone with a lavalier microphone. I would walk along the beach and do. And I was doing five episodes a week. It was a huge amount of work, and I wasn't able to maintain it because at the time when I started out, I used to write on my books by hand. I then ran into some problems in my eyes, so I had to dictate everything. I couldn't use a computer as much, which meant I have a certain amount of time I can spend talking every week, and I can't go beyond that. So the podcast got pushed down and it became weekly over. Since then, it's transitioned and it was interview podcast. For a while. I was thinking about shutting it down until I had a conversation with someone. I was like, you know what? This isn't the direction of my business anymore. I'm not talking. I don't like the people I'm talking to.

Like, it's.

I. I'm not excited by that. And I was going to shut it down. He goes, well, what we can do is just rename it and pivot it. So it used to be called the Servo Master Podcast, and it was ranked really high. It's really high in listed notes. It's one of the top 2.5 or 2% of podcasts in the world. So it had a listenership, it had a history, and I've renamed it around 30 or 40 episodes. And now I talk to people that I'm excited by again, and that's really part of it. There's a lot of mistaken beliefs about podcasts. A lot of people think a podcast is a great way to get rich. It's not. It's a real job. It's a lot of work.

Jonathan, if somebody wanted to start a podcast, as crazy as that may sound, what would be? Knowing the tools that we have today, what's the easiest path, the lowest barrier to entry for individuals?

I think that there's three pieces to it. The first is the technology for recording your episodes. So I've been using Riverside for a long time. As I told you yesterday, I just discovered I have an account. I have an account with a competitor that I use descript for editing my episodes. You. You need descript in your workflow. It's simply the best AI tool there is. It's the best for video editing, it's the best for background noise removal, it's the best for audio enhancing, it's the best green screen removal tool. They've added all these features, has the best eye fixing when you look down and it fixes. They've become this really powerful tool and they don't advertise as an AI tool. They have an amazing replacement. So I say the wrong word all the time. And if I say the word chapter, but I mean book, I can just type in the word book and it will fix and it will sound like me. That saves me so much time when I'm editing. I do that all of the time. Every single video I publish, I do it fix at least one word, usually three to five. All of those features are amazing. They also. They own squadcast. So I found out because I have a big account with them. They go, oh, you have a free SquadCast account at the highest tier. I was like, oh my gosh, I didn't know that. So I'll probably switch to squadcast because I end up editing it in descript anyways. And the process of moving it to descript is so much faster. So actually I look at it go, this will be a faster process for me. It has a feature. The feature that I really look for in recording is one that records locally and then uploads. So when I'm with a guest, if their power goes out, the episode's not lost. That's what I look for. That's. And I used to have. I have a lot of power outages, so that's actually something I specifically looked for. That's why I started with Riverside. The second thing after that, once you think about that, you want to think about the tools you're using, what type of headphones, using what type of speakers, using what type of microphone. You don't have to spend a lot of money, but be prepared to spend at least $200 on your microphone that at that price point, you're going to get something really solid. It's going to make a huge difference. Most important thing is to really develop a process. So I was actually working on this yesterday. So I have this on my desk, was writing out what happens after I do an episode. All of the steps to go. I need to. Some people call this a macro or process or a series of steps or order of operations or system. But I was working on it going, well, what are the order steps? Because there's a bunch of small steps in between. So first I take the episode. There's a problem I've been having with Riverside, which is that the audio is coming out corrupted. So I have to download the two wav files. So I have to merge those two wav files with the video and render that. That takes time. That creates the first thing. Then I put that into script. I use the script's audio enhancer and remove the filler words, the ums, the ahs, the essos, those all get removed. I export that. That's the version that goes into video to make the clips. I know that if I do the final edited version, which has a music and stuff, those AI tools will pick the music sections because they're the loudest or the highest energy. So if you use Opus, it will same thing. They all do it. They all. I don't know if the new version of video will do it, but in the past they always did that. So that's my process. Then I edit. So that's the thing that's gonna make clips. Then I add in the intro music, the outro music, that and export in 4K, that's the one that's going on YouTube. And then I'll upload a buzzsprout. Then I have to then do a version that's at 1080. Because you're not allowed to stream or put content on LinkedIn above 1080. I live stream the episodes on LinkedIn. That's where my biggest audience is. And you can't, if you have a 4k version, LinkedIn just won't show it because they don't rewrite, they don't know how to handle a re render. So you have to be at 1080. So there's a whole bunch of issues with streaming to LinkedIn. It's the worst platform for it. But when you get it right, it really gets a lot of viewers. And I get viewers every single week, new viewers. I get more viewers that are not from my existing audience from those than anything else that I do. Because the person usually tags themselves. I can tag the person, say, oh, here's the episode you're on. Just doing that means it will show up in their followers feeds. A lot of them, most of them reshare it. Most of them do all the other things to generate buzz there because they don't have to do something hard like post something new. They can just click repost. That process means I have to generate four versions of the video. So that's a process. Then I make the clips, I have to put them in one folder and then I have to have the link to the clips folder I have to have the link to the Buzzsprout episode, the audio episode, and link to the YouTube, so I can send all that to the guest in one email. So that's a lot of steps. But it takes me about two hours per episode. So today I was working on it because I had six episodes to do in a bunch I hadn't done in a couple of weeks. And you just work through the process. And then I have a process for creating my cover art for each show. Each show has its own cover art. Because it's on YouTube, it needs a different picture. So I take a picture of the guest, I upload it to midjourney. I generate a cartoon based on them, and that way it catches the eye. It's not original picture, and that allows it to be a little bit different because most guests just want you to use the same picture they use everywhere. I'm like, that's boring. That won't let me stand out. And I do get some people who complain about it. I'm like, well, let me ask you a question. How many of the shows have you, in any look, have they shown you the thumbnail besides mine? Oh, you never look, so you don't actually care. And it's a certain type of guest that will push back on me. I go, listen. And they go, I want you to use this picture of me. I go, yeah, I did. I have a specific aesthetic, a specific cartoon style. It's not supposed to actually look like you. It's supposed to be fun. It's supposed to be close. It's supposed to just be a little bit different. And I design it in a way that gets the grabs the eye, which is more important to me than it looks exactly like the person. It doesn't need to look exactly like me. And you'll see when you get. When you see your image, it looks close to you, but it's not perfect. Yours for something. I had to do a lot of iterations for your image today to get one that I thought was looking pretty cool and enough like a cartoon that you could tell it wasn't just a fake photograph. Yeah. So that's all part of the process. And knowing that 1 out of 10 guests or 1 out of 20, will push back on the image, and it's like, well, you know, it is what it is.

All right, we've got Edward Rar on the podcast today. Welcome. And, Edward, why don't you tell everybody a little bit about yourself, but also about your cool company.

Hey, Chris. Pleasure to be here. So, yeah, I'M the CEO of Podcast AI and we automate podcasting. So we, right now we automate everything that you need in terms of post production, distribution and promotion for your podcast.

So as I shared with you before we started recording on Monday, I spent. We have a. We've. We've kind of hobbled together our own version of it through automations and Zapier and APIs and very complicated stuff. And it took us about 90 days. And to explain it to the audience, it took me about 45 minutes. Now, how long does it take for somebody to do almost the same thing we're doing with your tool?

We'll be back after a quick break.

Do you want to start your own unsexy business offering a local service or a trade to your local community? Come check out the Owner Ops podcast. We bring on people like Sam, who has built a $22 million roofing company in five years. I've also brought on people like Brock.

Who generated $1.3 million of revenue in.

His first seven months installing septic systems. If you are looking to build your own local service or trade business, come check out the Owner Ops podcast. That's ownerops.com o w n r-ops.com.

Well, we can explain podcast AI in 45 seconds and we can set it up in less than 90 minutes.

That's amazing.

So this is the promise of AI, generative AI, right. It's like, oh, it's supposed to make things easier and faster and all that kind of stuff. And yes, we've done it internally, but what you're doing here is off the shelf. Somebody can. So where does it start?

We talk about.

Podcasting is a popular topic on using AI at work, because AI contributes to it a lot.

Right.

So for somebody who. Let's start with somebody who's maybe hasn't started one yet because of the concerns about technological impediments or the lift required to get all these things done. Can you kind of walk us through what that process looks like for a podcaster using podcast podcastai.com yeah, you just.

Open an account on podcastai.com, pick a plan, create a show, create yourself as an AI host. That basically means you're creating a profile of yourself. You're uploading a three minute recording of your voice, and it now knows your voice, so it can actually produce audio in your voice, which is pretty amazing. And then you assign that host to your show, you set the metadata for your show, you give a description, a title, you set some cover art for your show. So you Just draw something in Canva or Photoshop and set that as the show cover. And now you're ready to upload your first episode. And your first episode is really just, you know, record a zoom call, record in quicktime, a monologue, maybe introduction, do a Google Meet, record that. Tons of ways to actually record your first podcast. You don't need to start, you don't need the production quality to be super high. Joe Rogan is probably the top podcaster. Well, guess what? If you look at his first hundred episodes, the production quality is not super, super high, right? So people will forgive you if you start without the highest production quality. So you just really just do a zoom recording. You and your friends, you and a lead, you and somebody, somebody interesting in your market, just record that. And we have something called an episode studio. You can upload your episode. You can, if you want to create an intro and outro, you can do that. That's totally optional. The episode studio will affix that to the, to the recording on both sides. It'll automatically find something that's a good cold open and sort of put that in front of your intro, and then it'll give you a fully rendered video. But that's an optional step. You can really just go, go straight with your recording, upload your recording or your finished episode to Podcast AI. That's your first episode. You hit generate. It'll do your description, it'll do your chapters, your key points, your title. It'll do all of that stuff. We call that post production. That's all the information on your feed that goes out to Apple and Spotify. At that point, Podcast AI will generate an entire website for your podcast with all the transcripts and everything, just completely automatically. And so, you know, we're already. We prototyped blog to podcast, so we found people that do blogs pretty consistently. We made it so it just scans their blog as part of the docket every Monday, generates an episode, ad libs, an intro, ad libs, a conclusion, ad libs, all the posts they've done into content and chapters. And it has intro music, outro music. It's fantastic.

So when is that going to be released?

So we have that prototype with customers. If anybody's interested in using it for their, for their blog, we can do that right now. Just get in touch and. Yeah, our first customer who started deploying it once we finished setting it up, he asked, okay, so what do I have to do now? Right? And I said, nothing, just keep logging.

Right?

This thing's on autopilot.

I want to Introduce our guest. Chris, why don't you tell everybody a little bit about yourself and what you're up to in the AI space?

Yes. So I spent my career as a software engineer, worked at places like General Instrument, Motorola in the C band satellite industry, Hewlett Packard with their printer projects, Sharp Healthcare, which is a large healthcare organization here in San Diego, Mitchell International and Transcorp. And now I'm in the casino gaming industry as an IT manager. And about a year ago when all this IT, this AI stuff started happening, I put my nose to the grindstone and started basically just spending time in the morning every day learning this stuff.

Let's talk about that learning curve for you. You got a computer science background, you've been involved in technology for your entire career. There may be people listening that would be of say, of course it was, you know, of course you're the person on the other end of this. Of course this was easy for you.

But.

Did your computer science background come into compressing that learning curve at all when it came to generative AI?

Not initially. I, you know, I was hiring ghostwriters to write blog posts and I'm an Amazon seller on the side. So I basically was hiring people to write content for me. And when the generative AI stuff I started with Jasper and when the generative AI stuff came along, I started using it for that and realized that I could basically do that instead of hiring people. I was paying quite a bit of money, like $75 per blog post just to get traffic to my E comm site. And that's basically how I started using it. And then once I discovered automations and the powerful combination with the generative AI and automations, I started playing with some of the no code solutions to do some prompt chaining and some automatic report generation with those solutions.

Yeah, that was one of the things that I was very excited about talking about. Here is what initially got me really interested was most people who will be using generative AI on a day to day basis don't have a technology background. So most of the time when we're doing these interviews, we're not really exploring what happens if you get really sophisticated with generative AI, with APIs and the assistance and stuff like that that we talked about on our other call that wasn't recorded. So I definitely want to pull on those threads. But the reason that Chris and I got connected, just so everybody knows the context, was at the parent company of this podcast, this chiefai officer.com and we have had a manuscript that was 95% done on a book, the role of the chief AI officer. It's been done for almost six months and it was just one of those things that we couldn't get over the finish line. There was always something in the way and Chris and I got connected through the AI space and he was sharing the project that he's about to tell us about. And needless to say, I was very intrigued. And as a result of Chris coming on the scene, the Chief AI Officer book is a masterpiece and was actually published on Amazon as of last night, thanks to Chris. So Chris, why don't you tell us a little bit about what that conversation was that you and I had.

Yeah. So basically you told me about your Chief AI Officer program and I was really interested in the program and enrolled in the program and started going through it and kind of wanted to reciprocate for the work that you were doing with that and offered to write your book with my product, Ghostwriter AI, that's Ghostwriter without the E. And asked you about that and basically got some information from you about your courses and about your background. And with that night, pretty much I was able to write the book with my automation system. And basically we ended up with a book.

We.

Yeah, we had the first draft the next day and then your people went through it and, and fixed it and you, you guys made it your own and we were able to publish it last night.

So the process for somebody. Because it's a self serve. Right. Like I could go in there and if I wanted to do book number two, I would. You wouldn't need to be involved. Right. For the creation part. So.

Right.

So that somebody can kind of understand. Because if I'm driving right now and I'm listening to this and somebody's talking about writing a book with AI, here's what I'm thinking. I'm thinking that I'm giving it maybe a little prompt, but what I'm getting out the back end is synthetic. It's inauthentic. It's just like some computer generated algo generated series of words that happen to make sense, but that's not the output that's happening if you do it the way that you're doing it. So I want people to understand that it's not, it's not just, hey, I write me a book that there's. The human is involved in this process. The human is directing the, the, the, the output. So for the human who's interested in doing this, who's listening to this, what does that experience like? What is it? What Is quote unquote, writing a book with AI using ghostwriter. What does that look like?

Basically for in ghostwriter, I basically you, you come up with a book title. So you're filling out a form. It's a pretty simple 5.5 part form where you describe the book. So you have the book title, the book type. And right now I'm doing just authority nonfiction. So it's monetize your expertise by writing a book. And then 10 chapters max. So you can choose the number of chapters up to 10. And then you fill in the book description. And then there's a little AI chatbot that's, that's on the form, that walk. If you don't know how to, you can't come up with a book description. You basically do some ideation with that little chat bot. So it'll write the description for you. And you just cut and paste that into the book description field. And then you want to supply it with a book outline. And that little chat bot will help you write that as well. And then additional context, which is your notes. So if you have a personal story or you have specialized background, you put that in there.

How much context can I put in for that? Because if I don't know, let's say I was, I already had, I've been doing public speaking and had transcripts and stuff like that. How much could I feed it?

So it's. Right now it's 50,000 characters. Okay. And that's kind of the max amount that I can put in without breaking it, without getting. Making the prompts too long.

Yeah.

And then you can upload. The next step is you, you can choose your tone and style. So you can upload a text file that has your, your writing sample in there. And it's your author background. So you'd have an author bio. And again, the chat little chat bot will help you create an author bio. So it can either analyze your tone and style from your writing, you can specify a tone and style. So you select from option boxes basically what your tone and style is, or you can emulate an author. So like if I want to write like Tony Robbins, I just basically punch in Tony Robbins in there. And then the next step, the final step is defining your audience. So in the chat bottle, help you identify your audience. So it's like if you're in your, your instance, it would be like, you know, professional or business. Right. For the, for the, the niche. And then the final step it would be reading level. So you can specify, you know, if you want it. Yeah, if you don't want it to get too technical, you can, you can say right at a fifth grade reading level, which is a pretty safe reading level to make it conversational. And that's pretty much it. And then once you have all those things filled out, you basically hit the go button and it launches my automation and the automation will, it will generate an outline from the outline that you provided and it'll walk through that outline and write each chapter sequentially. So it's keeping track of the context of the book as it's writing it.

So it's got memory a bit.

Yeah.

Right.

As I mentioned, our guest today is Kurt Malley, an old friend, a brilliant marketer and owner of Black Box Social Media. So you know, you mentioned one thing in particular and that's a pretty common impediment for a lot of people. It the synthetic feeling of the output. They're looking to create some content or they're looking for some sort of maybe make turn it into a social post or whatever. But the output, it especially if you're used to working with the models, you're like, they didn't write that, GPT wrote that. Right. So this training them to be able to write in your voice is something like. That's a code that I've been looking to crack for a hot minute. Oh, and you're talking about it. Okay, good. With this video transcript, I'm going to ask you in particular, is it like a video of you doing a sales presentation? Is it a video of you kind of just like riffing with somebody? What's giving you the best result?

Sir, we have a system and a program that we have created that is really revolutionizing our business and it's revolutionizing lots of other businesses that we're working with right now. It's kind of like you've seen Alex Harbozi absolutely blow up with these short form videos. So I'll just give you the process right here, we'll talk about it and if you want me to dive, dive in deeper that I can and I'll train this whole model. So here's what happens. We either show people how to do this, we do it with you, or we have a done for you service. I'm not plugging my service. I want to show you the opportunity. I'm going to want you to plug.

Your service in a minute.

That's great. So what we do is we do an interview like this with a client of ours and we'll do it for 45 minutes to an hour, about every three weeks. Now what we're going to do is we're going to ask AI. So for instance, let's say, Dr. Daigle, I am interviewing you. And if I am interviewing you, I want to ask you 25 questions the AI is going to come up with. So I want to do a little bit of research on you, your topic, your audience. I want to ask some really good, you know, 25 questions. Now I'm going to remove my voice from the 25 questions. So after an hour long interview, it's just going to be you. I'm going to go ahead and I have this whole transcript of just you and it's just you in the video. So someone watched this RAW video. It's going to look kind of weird because it looks like you're just talking to yourself. Right? I like using Descript specifically. I'm going to go ahead and take the interview, upload it. I'm going to go ahead and remove Kurt Molly from the interview. Now it's just going to have Chris Daigle in the interview and that's it. Now I can hit a button in Descript to take out the ums and the ahs and the pauses and clean up the audio. Four buttons, two minutes. Done. Now I have all this done and I have this transcript. Because when you're inside of Descript, if you're not familiar with it, there's a whole transcript. You edit by the transcript, which makes it really, really easy to edit. Now essentially what we're going to do is we have this video and we have this long form post. So my next goal is before I start editing any of the videos, actually let me take this back. I'm going to go ahead and render the video. So I'm going to now download the video. So I have to download the video of you just talking directly into the camera. I'm going to give you four websites that we are building businesses on and other companies are building huge brands on. I take this video and I upload it to Opus Clips number one. Once it goes to Opus Clips, Opus Clips is going to automatically make these edits. Just like Alex Harmozi short form. The AI is going to look for hook angles. Is this a popular common topic? It's going to rank them from like 1 to 20, which the best short is it's just not going to automatically cut a minute. It is finding highly useful information because of the transcript. So now the video is added inside of Opus Clips. We use Get Munch, which really focuses over on YouTube videos. This is going to create YouTube, TikTok, stuff like that. The third thing that we're going to use is something called clap app. K l a p k l a p app klap video app. And then the next one is video IO v I d y o IO so if you take this one file and you upload it to all four of those sites. Chris Daigle a 45 minute interview can turn into over 200 different raw clips. They have captions on them. They don't really have any social media posts yet, but I'm going to go ahead and get to that. But they have captions on them. But here's the rub. We can do this all in an hour. We can record the video, put it all into AI in an hour, Everything's done. So what we do is let's just say, let's just say if we boil that 200 videos down to 100 videos, right? We just take 50 of them. Take the crap one, just throw it. We're not even worried about editing. Throw those away. Now you have a hundred videos like just from this interview right here. Mr. Daigle. 100 interviews would allow you to post three times a day for the entire month with just one interview. Harmozi was posting three times a day when he blew up his brand. But he was spending $88,000 a month to get the content out there. So here's what happens in the second step. In the second step, we now have all this content. So what we do is we build a GPT for the clients and I'm going to tell you how to do it right here. And it's a three part thing. Part one, we're going to do research on the client's avatar. Who is your perfect avatar? We want to figure out what are the wants, what are the needs, what are the desires, what are the trigger points, what are the things that they want to know about? Here's our perfect type of avatar. Now we don't name this avatar Wednesday or Tom. We don't have five different avatars. It is, here's the perfect market that we're looking for bucket number one. So that is going to be put in a notepad and we're going to upload that notepad of my perfect avatar into our chatgpt that we're going to create. So I'm just going to make this up. I'm going to create a GPT that is the Chris Day. Go ahead.

Where are you developing? Are you using AI tools to develop the avatar or is that from the interview with the client asking Them.

You know, that's a really great question. Well, again, there's three buckets that we have for this. They can do it themselves. Done with you or done for you? Done for you. We'll ask the client, do it yourself. You just fill in 15 questions, put it into the AI. And Mr. Daigle, I'm going to give you another tip here. We take these questions and answer them to build a customer Persona. Then we go to PO and we ask five different AIs to come up with all of the best attributes to this perfect avatar, which gives me this huge list. Then I go to Claude and say, hey, Claude, condense this down to the best of the best. Mr. Daigle, we have emotional trigger points. We have why people buy we. There are so many things that the AI has figured out that again, I couldn't do this a year ago. That's file number one that we will upload as a resource into the GPT. Here's my perfect avatar. The second thing I'm going to do, I'm going to take that interview that we just had. That's about 45 minutes of just you talking. I'm going to take the entire transcript and I'm going to go ahead and put this into a notepad for a resource. But essentially what I'm going to say is this is the Persona and the speaking tone of Chris Daigle. Please analyze this transcript to write the following output in the speaker's voice tone. Written under a fifth grade level, but uses the same vocabulary as the speaker. Chris Daigle, in this video, that right there, because it was a half an hour interview, this is in your tone. So that right there is the second file. Now number three as the third file, what's your offer? You're a consultant, you interview, whatever that is. What is your offer? And we're going to add that as the third file. That's the third bucket that we're going to upload in the GPT. Okay, so now we have who's the perfect customer, who's our tone, and what our offer is. Now what's going to happen is every time that there is a short that is made. So let's say we go into opus clips. In opus clips, the number, the first short is going to say, hey, we chose this short. It's going to tell you why it's the best one out of all the edits they did. They're going to go from best to worst, but they're going to say the opening hook was good. This is really relevant. Here's a Framework that works really well. And then below that they're going to give you the transcript of the short form video. So what we do is in your GPT that we just made, we take that transcript of that short form video and we have, actually there's one other button we put in the GPT we say create for me from this transcript. Create for me three Facebook headlines, three Instagram headlines, three TikTok headlines, three Twitter headlines, three LinkedIn headlines. Then give me four hashtag suggestions because really we're only using four hashtags in all of our posts for each one of the platforms that ties into this content. So actually this is, this is actually another notepad I should say, for really the fourth bucket that we upload that says, hey, for your output, you're going to create social media posts in this, in Chris Daigle's tone, with this customer in mind, with this offer in mind. And they're going to produce all the headlines, the hashtags, and this is most important, they're going to give me, the AI is going to give me five posts to choose from that are written between 1750 and 2200 characters in your tone of voice below a fifth grade reading level. Interject fun facts when possible. So step number one, we record the interview. Step number two, I put it into Descript, remove all the background noise, everything else I have that transcript. Step three, I'm going to build that GPT. Takes me just a couple of minutes, maybe a half an hour. Step four, I'm going to take that file and put it into the four different AI tools to cut up all this content. Step five, I'm going to take all the transcripts from each one of these short form videos, paste it the GPT. It'll instantly give me all these headlines, all the descriptions, long form posts get shared more, viewed more and read more. Now we have the post made, but Mr. Daigle, here's the final part. We got a ton of videos. We now have a ton of social media post options to choose from. We still have to look through the options and post them. I use repurpose IO Repurpose IO Every time that we have a short that's edited, we put it inside of this Google Drive folder. The Google Drive folder goes to Repurpose IO and syncs with all of our social media sites, all of our business sites, all of our personal pages. The only thing it cannot publish to is my personal Facebook page, Instagram, personal business, LinkedIn, everything. So now, Mr. Daigle, I have 100 videos from one interview. I have all these social media posts written for the individual channels. I just have to schedule it out on Repurpose IO. So if you're doing regular interviews like this, Harmozi would post three times a day to blow up his brand and would spend 80 grand a month to do it for under $200 a month. I just showed you how to produce double the content that he did to put into reels. And I'll put this last thing to wrap it up. Reels are getting seen now more than anything else out there. Let me give you a perfect example. Last year HubSpot Hootsuite did a study and found that when you post organically to Facebook, a long form written post or a video, 1 to 2% of your audience will see it. So easy numbers, a hundred people follow you, one or two people see your message. Reels, shorts, these short form videos that are a month, a minute long that I just talked about. All of the platforms have the same type of algorithm, but it's like Spotify, they will find people who like your content. So this short form content, this is how Hormozi blew it up, can be seen by 30, 40, 50% of an audience that isn't even attached to you. You can start a brand new channel, have no followers or nothing, crank out a bunch of these one minute videos using AI. All of a sudden you have again, harmozi would spend $80,000 a month on content. You have the same strategy for under 200 bucks a month and it's going to take you a couple hours a week to pull off to look through this stuff. But again, this is 88 or $80,000 a month strategy that you can literally do for 200 bucks.

That's pretty incredible. We have a pretty robust content production and distribution. But this is, I'm going to have my team listen to this as soon as this is off the press. I'm excited to invite Jesse Hughes from Leonardo AI eye to the podcast today. And we're going to talk about something that's been on my mind for a very long time. Jesse, why don't, before we get started, why don't you tell everybody a little bit about yourself?

For sure. Hey, Chris, my name is Jesse Hughes. I'm the senior creative technologist at Leonardo. I'm also very lucky to be an artist in residence there. So I have, you know, a 10 year career in the creative industries, been working across that space of emerging technology and how that kind of intersects with like the arts and the creative industries for a long time. And then Last year, I think, as we were all aware, Generative AI really, really had its moment and it raised a lot of attention. And it was this kind of incredible moment as a creative where I went, oh my gosh, my practice has now just expanded tenfold and I just couldn't stop exploring how my practice could now just expand in every vertical that I hadn't even considered before. And so that was what kind of drew me to Leonardo in the first place. So for those who don't know, Leonardo AI is an incredible Generative AI creative toolkit for content creation. What I would say the reason I really enjoyed LEO was it was everything all in one place. Back when I was learning and trying all these different things, you would have to go to one platform to make an image and another platform to upscale the image and make it high quality. And then you have to go to another platform to add motion. You have to go to another platform. So it was just my subscriptions, guys. I'm paying for Netflix, for Spotify. Like I can't have all these subscriptions. And yeah, so that's kind of where I ended up finding Leonardo. And the, the platform has grown just so much of a community within such a short time period. So Leo has the third largest discord community in the world. So that's like, that's massive. A group of people who are all here to be like trying to learning off each other. And in 18 months we have grown to have over 15 million users. Like that is what that is incredible. Right?

Congratulations.

Thank you. I think we're currently Australia's fastest growing SaaS, which is absolutely insane. Wow.

Honestly, it's been the easiest of the tools that we've used, so that was pretty, pretty refreshing as well. So you mentioned a couple of things. The growth has been phenomenal.

Yeah.

Why?

I think it's just what you said. It's easy. Like it's easy. The accessibility is truly there by design. Like we are the team was and it's something that we're actually currently working on, on making it even easier because we are, we are at this kind of like crossword between do you give people like complete creative control? Which is our end goal. It's one of our like pillar values at LEO is creative control. So by that I mean the ability to literally like go in and edit like specific in the same way, like you have, you know, high end creative software like whatever, Photoshop or whatever. Like, like, like detailed.

Yeah.

Like we have the capabilities of the tools with that within leo, but then we also want to cater for the people who are coming in for the first time, this is their first time doing image gen. How do we make that like a seamless experience? And so we're currently working on, yeah, some, some really exciting things which are launching quite soon on how we can be even easier. But I think there was probably that and then the other element of the images are fantastic. And that is the thing. It's like, like you said with your experience, Dali, we can't. Gen AI will not break through to the mainstream if the images are subpar. And I think that is something that's very important is that I use the word indiscernible when I, I give my lectures. And it's this idea of like, AI works in a commercial context. If the imagery feels indiscernible, as in a, as in an audience isn't looking at that and saying, oh, that's an AI. The other thing is like, audiences are incredibly savvy. They're also incredibly on edge about trying to like, identify these kind of things. So it's a sense of like, well, what instance are you using it in? Like, what example is this? Like, use cases, this being used in, Is it appropriate? Like, like where can you, you know, do multi layered stuff? Almost like if, like what, what is the use case and how do you want to build upon that is what I think is really relevant. You know, like, I have a, I have a friend who's, she's a founder, she's, you know, building her business and she wanted this beautiful, like stock photography. They were going to charge $500 per photograph for her. Yeah, right.

Ridiculous.

How in the world, like, honestly can a startup or found or the general public, like, that is not an affordable. It's just not affordable. And beyond that, it's not photography that is the specific stuff that she wants for her brain.

She had to settle.

And it was a sense of like, well, now we have the capability where we can define the models, define the scene, define the color tone, define the grade, the new tools that we're rolling out. We even have the ability for style. So if she has, say, a brand guideline of what her photography style is, we can, you know, go in and make sure that all of our images reflect that same style, which is a really, really helpful commercial tool for.

Yeah.

For people who are trying to build their own, their own stuff.

What would be first steps for somebody that was going to be trying it out this afternoon after this podcast episode?

Yeah. Cool. So I do it the way I teach in university, so I would start with real time canvas or like say Leonardo, there's like a few, like real time canvas is just a good way of being like, I'm going to have a doodle and have that instant kind of like adrenaline hit of being like, oh my God, it's making what I want. And so like start with your little sketch, do something silly and then play with the different models. So like the ray trace, the illustration, the sketch, like just so you get an idea of when, when the system knows what it is that you want, it will give you what you want. And if you're getting not very good images, it's because you're not probably not telling it the right things or you're not using it correctly. Like the results are phenomenal when it works, when you can articulate things correctly. So I would say probably start there and then going into the image generation area. Like image generation is where we is like bread and butter. I mean we also do like motion. I haven't got into all our other features because there's a zillion. But you know, we're able to do motion as well. But let's just keep it simple for today. Yeah, yeah. So you would, you would type in a prompt. And I think when you get into prompt writing, using AI to write prompts is really helpful. It's a feature that we're going to be rolling out publicly I think soon. Soon. I am obsessed with it since. So I think, I think it might only be available to me right now, but this improved prompt. So I will write like a simple dodgy prompt. I don't know, whatever.

Like in Leonardo.

Yeah, in Leonardo. So directly into the prompt bar. So like a woman, it's got a.

Prompt improver that's coming.

Yeah, Chris, very cool. Game changer. I will write maybe like 10 words that aren't very good. Oh, here, here we go. Okay, type of prompt. A woman hiker on a mountain side photography. I don't know. This is my excellent morning brain. And oops. And all of a sudden if I press improve my prompt, I'm going to get this. So it goes from a woman hiker on a mountainside photo to. A resilient, adventurous woman stands proudly on a rugged mountainside, her gaze fixed at the vast horizon ahead. The photo captures her, a mid height with windswept hair and a determined expression. This stunning photograph showcases the beauty of nature and the spirit of exploration. The vibrant colors and sharp details bring the scene to life, making it a truly captivating image that conveys a sense of freedom and awe.

Much better picture, I'm sure.

I think that's a bit more detailed than a woman stands on a mountain.

Yeah. Thanks for tuning in to Using AI at Work. Don't forget to subscribe for more conversations about how to use AI at work. And a special thank you to our sponsor, Chief AI Officer for empowering businesses with AI education and training. Visit their website for a free AI Readiness Assessment and AI Strategy Guide to help you get started using AI at work. That's www.chiefaiofficer.com. so thanks to our producer Evan Desaunier for making this episode possible. Follow us on Twitter at the handle using AI at Work Work and visit www.usingaiatwork.com for free resources to help you harness AI in your.