Leveraging AI

303 | From ChatGPT to Codex: How AI Agents Are Transforming Marketing and Business Operations with Dan Sanchez

• Isar Meitis • Season 1 • Episode 303

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 42:09

What if the biggest AI opportunity for your business isn't ChatGPT—but the tool quietly replacing it?

While much of the AI conversation has focused on Claude Code and coding assistants, a new contender is changing how business leaders think about productivity, automation, and execution.

In this episode, Dan Sanchez joins Isar Meitis to explore how OpenAI Codex has evolved far beyond software development. Together they reveal how AI agents can proactively find context, take action, automate complex workflows, and become true collaborators inside your business, not just chatbots that answer questions. 

If you're looking for practical ways to scale marketing, streamline operations, and unlock new levels of efficiency without increasing headcount, this conversation offers a glimpse into what the next generation of AI-powered work looks like. 

In this session, you'll discover:

  •  Why OpenAI Codex is gaining momentum beyond software development 
  •  The key differences between Codex, ChatGPT, Claude Code, and Claude CoWork 
  •  How AI agents proactively find context and execute tasks autonomously 
  •  Why project-based AI workflows are becoming essential for modern businesses 
  •  How Dan uses Codex for marketing, content creation, and process automation 
  •  The power of AI-accessible folders, files, and organizational systems 
  •  How AI can generate, manage, and improve business assets over time 
  •  Practical examples of automating large-scale content operations 
  •  Why business leaders should start thinking beyond prompts and toward AI-powered execution 
  •  The future of agentic workflows and AI-assisted business operations

About Leveraging AI

If you’ve enjoyed or benefited from some of the insights of this episode, leave us a five-star review on your favorite podcast platform, and let us know what you learned, found helpful, or liked most about this show!

Isar Meitis

Hello and welcome to the Leveraging AI podcast, the podcast that shares practical, ethical ways to leverage AI to improve efficiency, grow your business and advance your career. This is Isar Meitis, your host. And as you know, in the past few months, we've been taken away by this crazy Claude hype, as everybody was doing Claude Code, Claude Cowork, Claude everything. And it really took everybody away, and people almost started to forget that the real 800-pound gorilla is OpenAI with ChatGPT. And to be fair, they were stuck in a really bad place end of last year, but they've made a lot of changes, and the biggest change that they've done was they launched Codex. And Codex was launched specifically to battle Claude Code, which means it was battled to allow people to write code faster and better. And it became a very good coding tool, and exactly in the same way that Claude Code became a tool for many other things other than writing code, the same exact thing happened with Codex. So there's been a huge spike in people that are using Codex for writing code, and in the past few months there's been a very big spike in people using Codex for more or less everything else. Now, our guest today, Dan Sanchez, who's also known as Danchez, held a senior marketing position in multiple companies in multiple industries, always doing really, really cool stuff. How do I know he's been always doing very cool stuff? Because Danchez has been on my previous co- podcast back in 2022. So it's been, four years ago. Yeah, March of '22. So four years ago, before ChatGPT came out, and even back then he was very knowledgeable, and I was always very impressed with how deep he goes into understanding marketing and also how good he is in explaining the concepts behind these things. And in the past three years, he's been all in on AI and how to use AI to improve marketing. And when I say improve, it's not just doing it faster, it's also doing it better. And in the past few months, he's been doing this with Codex. So we get a two for one BOGO kind of thing in this episode. We're gonna learn on how to do marketing with AI, but more importantly, or as importantly, we're gonna learn how to use Codex for doing things that are not writing code, which means these capabilities, skills, best practices that we're going to learn today can be transferred to doing anything in business and not just marketing, which is really exciting. And hence why I'm really, really happy to welcome Danchez back to a podcast, not the same podcast, and talk about how to use Codex for business- Danches, welcome to the show.

Dan Sanchez

Thank you so much for having me on. It's fun to do it a second time. Different

Isar Meitis

show- I know, I know

Dan Sanchez

but same people.

Isar Meitis

yeah. Different show, different topics. You and I have done- Yeah a gazillion shows for ourselves, for other people, uh, during that time- Absolutely and it's, uh, you know, we, we run into each other randomly. Again, just a quick, uh, story on how this happened. We run into each other randomly. We both were speaking at Social Media Marketing World. They had this year an AI Business World as a- That's right mini conference within the conference. We were both speaking there. We run into each other in the hall like, "Oh, this is so awesome," and we learn we're both doing similar things, and so this is a great opportunity. So again, I'm, I'm very, very excited. I'm one of those people that was caught in the, in the Claude hype. I've been probably 90% in the Claude universe, uh, other than for some of my clients. So I have clients in Gemini, and I have clients in Codex, and I have clients in ChatGPT, so for them I obviously- Yeah do work in those fields. But, uh, I'm myself very, very excited to learn from you today.

Dan Sanchez

Yeah. So I gotta say, like, when I started, I was probably a year after ChatGPT launched, and then I launched my own podcast, AI-Driven Marketer, 'cause I was like, "This is clearly the future." Like, I have to learn everything there is to know. And you-- we both know that, like, podcasting's one of the best ways to learn, but then document what you're learning for everybody else to come along with you. So, like, I did the journey just like you. We're both podcasting. I'm interviewing people. I'm, I'm showing my own stuff. Um, but about, I don't know, back maybe fall last year, like, the code, like, Claude Code is becoming a thing, and I'm just like, "Eh, I'm just not a Claude guy." I don't really find the code to be good enough for a new-- Like, I know how to do HTML and CSS. I'm good at web design. Like, I set up websites all the time. I'm really good with WordPress and all this stuff. But I'm like, the code just isn't quite sophisticated for me to be able to do more than maybe simple scripts that I need it for websites that I'm doing manually still, you know? Um, but then when Claude, it was, uh, OpenClaw came out, right? In the beginning of this year, things started- Yeah changing. Not only did OpenClaw start to take on its own trajectory, but Claude Code, around that same time, really started taking off, and then they launched CoWork. And all this is going on, and I'm like, I, I'm just busy doing other things. I'm exploring how to turn a book-- how to turn a podcast episode into a book and all kind-- I'm exploring other stuff. I'm kind of ignoring this whole other side quest going on. I'm watching OpenClaw curiously. I'm just kind of, I don't know. I don't know. The beginning of this year, it really got hard, even for people doing it almost full-time, like you and I, to keep up. Like, the paths diverged into so many different rabbit holes that I'm like, I don't even know what's going on in video right now. I just-- Part of me is just kind of like, "Screw it. I just can't keep up with all of it anymore." and I've always been a ChatGPT guy. I've liked Claude. It's fine. I just found that ChatGPT was a better tool for me in a bunch of different capacities. I also found that ChatGPT did have one thing over everything else in general, and that ChatGPT was better at dealing with ambiguity For some reason. And that's tr- that-- I think that's true across images, it's true across its content, it's true across its code. You could know less, and it could help you find the thing more and faster. I think that's one of its chief things, and why, one of the reasons why I recommend it for marketers and other, non-technical users, that are out there. I'm like, "Hey, if you, if you're not quite sure what you're doing, just use ChatGPT. It's much, much better." Claude is definitely more technical, especially if you really know what you're doing. Like, you're a coder using it to write code, probably should use Claude. If you're a writer learning to use Claude for a writing tool, you should probably use Claude. there's a bunch of different situations like that. But in general, I was sticking to ChatGPT. I would keep testing everything else, and I stuck to it. Little-- And then I, I think enough time had gone by and the Claude code thing was not slowing down. OpenClaw was continuing to go up. Codex, I started hearing from a friend that Codex was really good, and I... So maybe just a few months ago, back in March, I was like, "Okay," like I can't... It's been three months, and the news hasn't died down about this thing. I don't know about you, but sometimes I wait for a launch to happen, and then I wait to hear what ha- what, is, if, if people are still talking about it a month later, you know? Yeah. I'm like, "Oh, that's the one to stick, and now it's time for me to come in." I was a little late to the game 'cause I was busy. but naturally, as a ChatGPT user, I gravitated towards Codex. Yep. So I installed Codex, and there was a few learning curves, but it wasn't that hard. I also did Claude Code, and I was comparing side by side both the Mac apps, you know, and then you, you log in with your accounts, and I was using both, and I was like, "You know, Codex is just... I don't know what it is about it, but it's just more intuitive." It feels more like I'm in ChatGPT's super app. Codex feels l- definitely like it was designed for developers. I'm a little bit more confused. Cowork is kinda like a half step in between, but I was like, "Ah, I don't know." Codex was just all-in-one, and I found that now that I'm using it, I'd say about 90% of my work AI work is now happening in Codex, not in ChatGPT at all. Um, and there's a few reasons for that. But I think y- people on your show are probably familiar with the way Claude Code works, and Codex works in the same way, where you give it access to maybe a folder, maybe your whole computer, if you have a separate computer and you wanna do that. Um, but even if you just give it access to a folder It has access to everything in that folder. And the thing that's different about these harnesses they call them, it's crazy weird name to call these things, but whatever, that's the name.

Isar Meitis

Well, like ChatGPT is a great name.

1 GMT20260615-111834_Recording_avo_1280x720

This episode is brought to you by the multi-agent orchestration course. We are about to launch the July cohort, which is already sold out and a little bit over that as well, and we're almost sold out of the August cohort as well. So if you're interested in taking the course and you do not want to wait all the way through September, and trust me, you do not want to wait all the way through September, you should look for the link in the show notes and sign up as quickly as possible. We just finished a cohort and it was extremely successful. Everybody were extremely excited, was able to build amazing things. People got new contracts that they didn't get before. One person found a new job using the things they've learned, and so on. So people are making huge updates and upgrades to their career and to their businesses using what they've learned, real impact during and immediately after the course. So it does provide actual hands-on capabilities that you can use immediately to drive growth in your business and for your personal life and career. And as I mentioned, if you wanna learn how to use AI properly in this era of AI right now, so the first half into the second half of 2026, don't waste any second and come and join the course because just like all the other cohorts, this one will sell out as well, probably in the next week or so, and then the next time you'll be able to join us will be in September. So don't wait. Click the button in the show notes and come and join us.

Dan Sanchez

Yeah, I know. It's like, but that's a- I think they're not very good at names, really that's a bad brand name. The harness is a bad technical name because it's like- Yeah like a climbing harness. Like what are we talking about? You know? Yeah. So it's the harness, it's the app that then the AI model comes into. I find that using either Claude, both... This is true about Claude Code or Codex. The thing that makes it so different that people coming over from your ChatGPT web app is that they do two things differently. They proactively look for context, and they can take action better than your native ChatGPT, even with all its plug-ins and integrations over on the ChatGPT side. Codex and Claude Code can look for context proactively in the folder they have access to, and they can take action. Those two things changes the game. They were designed that way for coding applications, but now it's turning out that those two, I guess you'd call them, uh, unique, unique features may- are just, was what we needed in order for AI to really become more agentic, for it to be able to have a- autonomy to be able to do stuff for us and think, uh, without us and essentially take action. So with that in mind, I can dive into Codex a little bit. I promise if you're already paying for ChatGPT, you're gonna wanna do this today. Like do not let the day go by without going to chatgpt.com/codex, download the Mac app. Yes, if you don't have a Mac, then I think... No, I take that back. There is a Windows app. So- There is just go to chatgpt.com/codex. Download the app, log into it with your normal ChatGPT account, and you have access to it. It's like, it's already part of your subscription that you're paying for. and if you go to it, and I can show you, like it's really a simple thing. There's just two concepts you need to carry over from normal ChatGPT and over to conte- o- and over fro- to Codex, and you're good to go. The main concept is projects. We all know it and love it from ChatGPT. In Codex, you're essentially forced to work in a project, so you can't... You can have a generic project and call it like my random stuff file if you want to. Um, but generally, we're usually, if you're using projects over in ChatGPT, shoot, if you're using projects over in Claude, it's the same thing. It's essentially a dedicated folder where all those chats exist and all the context exists. You have to dis- establish your projects and then you're good to go, and you always have chats within those projects. And that's the main, main like one thing you have to transition, that everything exists in a project. And I can share my screen here and kind of

Isar Meitis

go through it. J- just to call out. Yeah. While you're sharing your screen, two things. one is I agree with you 100%. While I do this in the Claude universe, I touch the regular Claude chat Maybe once every three weeks. Like I, I just don't go there. It feels so antiquated and, and not very capable. So very much like you, like once, once you transition, there's no going back. Like the level of results you're gonna get are significantly different. The other thing is for those who, those who don't know what projects are, uh, and, and Dan is gonna show in a minute, but they're on the left menu. They're called projects. They exist in the regular ChatGPT, and they're an extremely powerful concept because what they are, they're a bubble of context, right? So yeah, ChatGPT has memory, and it kind of remembers who you are, where you work, and what's your role, but that's it. That's all it remembers. And inside a project, you can give it files, you can give it videos, you can give it HTML pages, you can give it PDFs, you can give it Excels, you can give it whatever you wanna give it, and it has its own bubble of memory inside the project. Everything that happens in the project, it remembers for the project itself. And projects can be used as a project. I have a new project, which is probably what Dan is gonna show us. But you can also use it for a specific client. You can always also use it if you're planning a trip. You can always like, whatever is a new ball of context that you wanna keep separate from other things- Yeah and be able to focus on that, uh, that's the concept behind it, and it's translates very, very well into the agentic world because again, in that ball of context, the AI can go and find whatever it needs on its own, and you don't have to give it to it

Dan Sanchez

So what's cool about it is I u- used to run into this problem all the time, and I figured they would fix this eventually within ChatGPT, and they didn't. They ended up doing it in conte- Codex. So I don't know about you, d- I, I loved the canvas feature within ChatGPT. I do. And it's- You know, you'd ask it to create a doc, and instead of doing a normal chat window, it sa- it would create a doc called a canvas, and in the canvas you could edit it right there. And the frustrating part to me is I'd create all these assets in Canvas, and then what did you do? You copy and paste them into Google Docs- Yeah where it doesn't have access to that. But oftentimes those canvases I was making are things that I would want, like, to plug. I'd want it to ha- That, that would be context for future conversations. But no, I have to go and copy and paste what's in the canvas over to Google Docs, and if I wanted that available for context, I'd have to download that as a PDF, upload it to the project as a PDF, and then it would have access and it'd be stagnant, and if I wanted to change it, I'd have to go change that Google Doc, download it, re-upload it. With Codex, it fixes all of that because every canvas, or now it's just a file on your computer, can be accessed later. And again, it pro- it proactively looks for that context. It will even leave little notes for itself to be like, "Hey," "when you're checking this folder, go check these things first." Right? It leaves little clues for itself so that when you're working on the same stuff, it's almost like everything gets better all the time. All the little canvases, now files on your computer, are able, are available for it to access over and over and over again. It's the coolest thing. Let me show you real quick. I'm just on my computer, and I'm gonna open up the Finder, and I just put all these in documents. I have all these folders. but I will create documents like this Danchez folder, right? And this is just my one project, and you can see Danchez over here. This is just my personal brand. You can see it over here. It's-- I have it as a file in my documents, and I have it as a file over in Codex. It has all my chats underneath, just like ChatGPT. It has all the conversations. Now, the difference is- It also has access to all the files on this folder. I've actually organized this particular project according to, uh, Tiago Forte's PARA method, which is projects, areas, resources, and archives. Very useful organization method. I recommend it for a lot of people. If you don't know how to organize, a project or one of these folders for Codex, use PARA. it's a good way to start. You can even throw a bunch of stuff in there, and every once in a while proactively ask Codex, "Hey, can you take a look at all this stuff we've done and just kinda, like, find out a good system for organizing it?" It'll do it itself. It can make folders, delete folders, rearrange things, write files. But again, it has access to all these different projects, whether it's my AI-Driven Marketer show. And you see this little README? I didn't write this. It wrote that. This is instructions for itself on what to think about and how all the things we have going on for the AI-Driven Marketer, right? These little MD files is the, kind of the key thing, the little documents it creates for itself. And MD is just markdown, and markdown is just a really simple language that AI understands well, and it's easy for humans to understand well. It's essentially a markup language, and if we were running this through, like, an HTML converter, it would make, like, a header here, a header there, a bulleted list here. But it's pretty intuitive. Like, you can look at it and understand. It's not code. Um- Yeah,

Isar Meitis

I, I always tell people it's Word document that doesn't look fancy.

Dan Sanchez

Yeah,

Isar Meitis

yeah. it's just the text, basically, from the Word doc.

Dan Sanchez

And it leaves all these files. The cool thing is, is you can do more than just MD files. You could do images, you could do videos. there's a few other advantages. And then now that I have the f- the-- now that you kinda understand, like, it's folders and files in here, just know that every time you talk to it, it's gonna be able to proactively go and be like, "Huh, should I go and look through this file? I should go and look for some context locally on the drive." It can also go search the internet like the chat app can, but this one has access to all the files all the time, and that changes the game. But let me share a little bit what I

Isar Meitis

love about- Just to show people the practical, just to show people the practical aspect of it, how do you connect the folder on your computer to the folder, to the project inside of ChatGPT? Because that is a difference from, uh, from the regular chat.

Dan Sanchez

Yep. So you can either have the folder that pre-exist on your computer and then just connect it, or you can... It'll create a folder for you. Like, if I come up here to Projects and I click, like, hover over it, and I have this, like, l- new folder icon, I can click and say Start from scratch, in which it's gonna ask me to create a folder, and it's gonna have me name it and tell me where it goes. Or I can use an existing folder, of which I'm, like, just gonna pick an existing folder I have. Either one, and then you have, like, like I have on my sidebar, you can see I have a number of projects for different clients, a test project that I did for a demonstration, much like this one earlier. and I have a bunch of things going on here. Now, let me park right now, is everything I've said is also true for Claude Code. This is very similar to how you would be work- operating on Claude Code. So why did I choose Codex? Codex has a few advantages other than the one I, I mentioned earlier, where I think it deals with ambiguity a little bit better. If you know less about code and you wanna code with this, then use Codex. If you, are n- have ideas and you wanna help flesh them out, Codex is gonna be a better thing for you. If you just want a general tool, Codex is probably gonna be a better super app for you. but other advantages is it can generate images natively, which as a marketer is something that I'm doing a lot. If you want, you could do it in Claude Code, but really, if you wanna do it in Claude Code, you're gonna have to create a separate account with OpenAI and just code a little application that every time I ask for images, it's gonna have to go and use some tokens from OpenAI. It's a separate cost. But here, it's kinda connected all in with your same subscription, which is really nice. the other thing I love about Codex home blanket. There was like a few different things.

Isar Meitis

So yeah, I, one thing about what you said, I agree with you 100%. Like images, again, as somebody who runs in the Claude universe, uh, I did exactly what you're saying. I have a side app that knows how to call ChatGPT and Gemini actually, and create images so I can compare the two and pick from them. But it is a separate cost. It's not a high cost, but it is a cost, and you need that extra nuance that you have to do this, and the chat itself is less aware of what was the previous image and so on, which inside of ChatGPT, because it created image, you ask it to fix something, it knows exactly what's the something that needs to be fixed.

Dan Sanchez

Yep. I thought of my two other reasons why I use Codex. One, the, one of them is it's better at using your computer's hidden tools. Believe it or not, your Mac, like either Windows or Macintosh, but particularly Macintosh, is loaded with hidden tools you don't even know exist until Codex finds them for you. Let me give you an example. A lot of people are using Codex to edit video, and no, it doesn't need Adobe Premiere. It doesn't need CapCut. There is local video editing tools that it just has that you don't even know about, but that it can use it. It's almost like code-based editing, and it can figure it out. You could literally take this whole podcast transcript and be like, "One, I need you to transcribe it. Two, I need you to find the clips. Three, I'm not just talking," 'cause you're like, "Okay, make clips." Fine, it looks at the transcripts and just clips it on the front and the end. That's a pretty easy edit. But you can even ask it to be like, "Hey, o- of the 60-second clip, I want you to find the five seconds within that clip that's the most exciting part. It's the peak. I want you to copy it, not delete it. Copy it, move it to the front end, and then let it play through normally." That's a more sophisticated edit. You're actually taking something- Yeah duplicating it, and putting it on the front. It can do that without you having to open a single video file, without you having to use CapCut, Adobe Premiere. It could just do it for you and make it vertical and do all kinds of things via the tools on your Mac. I asked it the other day, I was like, "Hey, I have this like 100-page PDF. I don't wanna read it. I'd rather listen to it. What are my options for turning into audio?" It's like, "Well, you have a free option. We can use your Mac's audio tool, and I could use it to create all the audio files for free." And so it just did it natively. It wasn't that great of an audio transcription, so I ended up using OpenAI's audio tool to do it. But, and cost, that cost a little bit of money to do. But it was great that it actually was like, "Oh, like let's, let's try the free thing available on your computer." It does this stuff well because OpenAI was smart and hired a bunch of ex-Apple people in order to help design Codex to be able to actually control your computer in a really natural and efficient and effective way. So those are little Easter eggs you don't even know about until you're actually asking it to do stuff, and it's like, "Oh, well, we have this c- c- tool on your computer to do it." better at Claude Code at doing that stuff. And then lastly, this is a temporary one because this will change eventually. Right now it's just way cheaper. Claude has gotten really expensive because OpenAI has had a head start on creating a bunch of data centers, and Claude got really popular in the last four months to the point where they're, like, dying in their compute capa- capacity. Yeah, to be fair, they just gave everybody- And they're having to raise prices to deal with it

Isar Meitis

they just gave everybody until sometime in July, I don't remember, like 50% off on, on Claude, which means I downgraded immediately to the $100 a month- Mm-hmm from the $200 a month plan. But, i- in general, I agree with you, and the, the general rumor is that they're, that OpenAI is gonna make their stuff even cheaper just because- Yeah they're both in like final stretch for their IPOs- Yeah and they wanna show that they're better than the other guys. So-

Dan Sanchez

yep.

Isar Meitis

So- it might get even, even a bigger spread from a cost perspective.

Dan Sanchez

Yeah. So ChatGPT is more competitive from a cross- cost perspective. so those are the reasons why I chose Codex. I just find it's more a, a much more intuitive tool for non, generally non-code work, though I u- am using it for code. It is very good at this. so if you, if I can, I'd love to talk about some, like, really interesting practical applications that I've done.

Isar Meitis

Yeah.

Dan Sanchez

Um-

Isar Meitis

Let's dive in.

Dan Sanchez

So I am... In fact, this is an old project. In fact, I'd rather just talk through it, but maybe I can pull up an example of it

Isar Meitis

while you're looking, two cents about this whole agentic universe versus those of you who are still using chat. And, you know, I've done multiple episodes on Claude Code and Claude Cowork, but the biggest difference is that it actually works with you rather than you tell it what to do and it gives you answers, and then you have to probably do the work yourself. So you brainstorm what you're trying to do, you find The right path with the AI. So the AI will suggest a path on how to do the thing you're trying to do, and then it can actually do the work, and you can define which checkpoints you want as a user to review the work. So as an example, uh, I haven't written a single proposal myself for a very long time. The AI does the research on the client, the AI takes the chat from, like, the transcript from the call, the AI develops the plan for the proposal, it writes the proposal, and then it shows it to me to review. Once I review it, it knows how to put the bells and whistles on it, right? It knows how to add, uh, my header and footer and table of content and, you know, all these kind of things- Yeah that need to be in a proposal. And it also knows, because it's an agentic world and it's connected to a lot of other tools, it knows how to put it in my CRM, it knows how to, create a draft email, it knows how to put it in the relevant, folder or create the relevant folder on my Google Drive. So all these things are agentic capabilities that allow the AI, again, like Dan just said, understand the concepts, understands the context, develop a plan, and then with your approval, create the actual thing that needs to be created.

Dan Sanchez

So let me show you the thing. I am working with a company called Trim Healthy Mama, and they are a supplement company, but they put out a lot of recipes for their supplements. it's a whole thing. So I have this whole site with recipes on it, and you can see they're, like, interesting recipes. They got nice photos. And I was like, you know what? What can I do here? I have about 1,000 recipes on this website, and I know this image generator can do some really cool things. I'm like, I wonder if I can create a process that takes every single recipe, turns it into an infographic. You know, you've seen those infographics of maybe, like, a flat lay photo with all the ingredients that goes into a recipe- Yeah labeled? And they're just kind of interesting little infographics. I'm like, huh, well, I have the recipe, with all the ingredients, and it's all well-structured and everything. I got an interesting picture of what the color and the thing looks like. I wonder if I can make a system that will literally take all thousand of these, generate all the images, and then just re-upload them all. With ChatGPT, this was impossible. Shoot, doing this manually is a freaking nightmare. Even if you can do all the images automatically, it's still a nightmare 'cause I have to download them, I gotta check them, I have to... We're just re-uploading a-- times 1,000, anything times 1,000 takes a long time. Five minutes- Yeah times 1,000 is a very is a long time. yeah. But I'm like, huh, I wonder if. So what I did was actually worked it step by step, and you can see I have this, like, whole dialogue. I'm not gonna get into the whole dialogue, but let me tell you, I'll show you little pieces of it. The first thing I did was I just took a web URL like this one And I just, this is how I recommend approaching Codex. If you're like, "Huh, I know it can write code, which means it can access WordPress." So let's test this one little baby step at a time, one baby step. And all I did was take the URL, and I was like, "Hey," I posted in here and just gave it instructions, being like, "Hey, take this URL. I want you to scrape the recipe ingredients. Take a look at the image. Download the image too, and I want you to turn this into an infographic of a flat lay." I described the picture that I wanted, and this was the result. It's like, can the... Can it generate the image? Yes. Look, it can generate the image. This is a different one. This is apple and walnut salad, and it has... Remember, the photo that it got was probably from a side profile, but it did a top lay, and it did a really good job of generating all the ingredients. If you can't see, like, every in- it's, like, a top lay, every ingredient in a separate little dish as if you were expecting it from a cooking show, and everything's got, like, handwritten or, um, text showing you, like, oh, this one's apple cider vinegar. This is mineral salt. This is- And it

Isar Meitis

looks really clean and professional tuna. Yeah. It's not cha- Yeah. It like, it looks like, like a design company actually made this.

Dan Sanchez

It's got the right logo at the top. It's got the right font. It's got a little call to action. This could be floating around Pinterest. Get the full recipe at trimhealthymama.com. So it has, like, a call to action in case the recipe gets disconnected from the URL somewhere. So this would be... This is hard to make if you were doing this na- uh, manually, but with a ChatGPT image generator baked into Codex, this is easy. Okay? So I'm like, "Okay, it can make the image. Ha." I'm like, "Cool." Now I'm like, "Codex, what does it take to, like, pull all these recipes, um, and create, like, a checklist for yourself?" It's like, "Oh, well, it would be awesome if I could get, like, your index file." I give it the index file. It creates a checklist, and I'm like, "Cool. What would it take for you to make, pull one, do it, and then upload it?" It's like, "Well, I would need access to your WordPress API." You're like, "Great. What do you need to do that?" And you're like... And it's like, "Sure. You log into WordPress here, go to this page, go to your user, and then generate the A- the API key. This is what I need. Tell me what it is." And I'm like, "Okay, here it is." It's like, "Great. Yep. Accessing your WordPress. Oh, give me this one more thing," and it's, like, helping you troubleshoot little issues it runs into. You're like, "Okay, got it." It goes, and it's like, "Great. I'm now searching your API. It looks like I have access to do this. Let me do a little test. Okay, yeah, we can do this. Do you wanna get started?" You're like, "Absolutely. Do it." And I'm like, "Now that you have the checklist of all 1,000 recipes, I want you to create a project Where we're gonna go through this and I just want you to do five, okay? So it manually goes and pulls five. And I had to talk through the steps. I didn't, like, just say, "Make the process." I'm like, "Here's the steps you need to follow. I want you to go and check it, pull the image, ch- do this. And when you create the image," I have very specific instructions, "make sure to re-reference the i- the logo file, re-reference the image file, and then pull the ingredients and make sure that here's the call to action for the bottom." So I had very specific instructions on what these images to look like. So I wanted it, there, there to be uniformity across the thousand different infographics. And it cr- we created that. And of course, it didn't just have that in the chat, it created an MD file. We created a whole project folder for this. It created an MD file with the step-by-step instructions. It created a different MD file with the checklist of all the recipes so we didn't redo the same recipe, right? And I said, "Great. Now walk through the checklist, do the first five." And it did it systematically. First, it would go and access the WordPress site through the index file, find the page, pull the image, pull the ingredients, do the work, create the image, uh, w- check the image to make sure everything was right, so it had a process for checking to make sure the logo was right and that they didn't screw up anywhere. And then it would take it and then actually upload it to WordPress and then check to make sure it got uploaded to the right WordPress recipe, right? So all that would take a long time. And then I, I didn't have to automate it to the point where it was like thousands and thousands. I could do a batch of 100 at a time and spot-check the hundreds. But it's a huge project that would take a very long time to do. But with Codex Took me about probably 30 minutes to get the process and kinda, like, check it and set it up, and then probably another hour to walk through a batch. And just-- And the cool thing is once you're walking through a batch of 100, you can set that thing, let it go, and then come back, let it go to sleep, leave your laptop on. It keeps working for you. Come back in the morning, 100 are done. Now, I did let it go for 100, and somewhere along the line, it had screwed up, so 100 were gone You know? It had screwed up 100 times. I'm like, "Okay," like, "Delete all those from the post. Go fix that all, and we're gonna have to re-render these." But it's like it, I didn't have to do anything. It could go and do the work of fixing all the things, and you re-render it and fix the one little bug. As, like, one of my favorite things about working with Codex is that every time you're like, "Hey, we didn't do the thing right," update the docs. This is what you need to do instead. Best-- One of the best things ever about working with these tools is that it can remember stuff. You can tell it and be like, "Hey, every time we run this process, make sure to add this note," or, "You did it, you did the wrong thing again. Make sure to double-check this every single time." emphasize it. Make it important, And this is how we delegate to humans. This is why- Oh, yeah it's such a big difference.

Isar Meitis

I wanna touch on a few things that I think are very, very important for people to understand, uh, both on the concepts that can be applied not just for marketing in this use case, but for everything, and also on the small little details. And I'll start with the small little details because they're connected directly to the last thing you said. those MD files, those Markdown files, are-- People hear about, you know, agents and skills and they think there's, like, something magical. It's, like, unicorns in the forests with spirits that make the agents work. The whole thing, the whole agent universe is these Markdown files. And what a Markdown is mostly an SOP, a standard operating procedures that's telling the AI how to do the thing. And if you think about it, what is the ultimate employee? The ultimate employee has a set of things they need to do, and they will follow the instructions to the T every single time. They're good,

Dan Sanchez

yeah.

Isar Meitis

And if they get stuck, they will try to figure out a way on how to get unstuck. And if they are still stuck or they think something is wrong, they will stop and go and ask questions. Yeah. This is, like, the best employee you can define and hire.

Dan Sanchez

Yep.

Isar Meitis

This is exactly what AI does Through this agentic universe. So these MD files are just SOPs. The other thing they are, which I really, uh, like what you said as far as lessons learned, I have a skill built, which you can all build for yourself, that writes lessons learned across projects. So I have one that runs one- Yes write lessons learned for the project, but I also have one that says, "If you see anything that can be helpful for other projects," as an example, "When you generate images, never do this or always do that. When you connect to N8N, always connect in this way." yeah. When you are using API keys from third party, like anything that is not very, very specific to that project, it writes to a folder that all the other tools have access to as well, and they all have to read it when they get started. And so then you get a learning organization, not just this skill versus that skill, this project versus that project. You're getting an organization that learns from any mistake that is made in any project that's done. The other two things from a best practices on things that you mentioned that are very, very important. One is you work with the AI to figure out the right SOP. Just build the process. Test it in a regular chat inside of Codex. Like don't build the automation first. Just, just try to get one right. Then go to Codex and say, "Okay, awesome. the way we have it now is what I want in the end. Please now write the instructions. Ignore all the stuff that didn't work. Write the instructions that did work, and build it into a process in an MD file that we can follow every single time." Then build another thing that Dan said, build a test loop. Now that it's there, go and verify that it's the right file, it's in the right place, it's loaded to the right folder. That's right. It follows the guidelines. Like you can build all of these things into the process itself. And then run a test. And the test is not one. The test needs to be a bunch. You can decide what a bunch is, right? Yeah. Could be a five, could be 50, could be... B- but two or three is just not enough because two or three doesn't have edge cases, and edge cases is what's gonna break your stuff. So you need to go through enough to go into situations where, let's go and take an example and follow Dan's example because it's fresh in everybody's minds. Let's say that one of those folders that the AI goes to, for whatever reason, does not have the description of the ingredients. The actual original webpage. So now the AI goes, and it doesn't know which ingredients to include because they're not included on the webpage. It will make it up. So now you have something that is a glass of water, but it has 16 ingredients in the image afterwards- Yeah because the AI knows it has to make up ingredients. So these are the edge cases you're gonna run into, and there could be technical edge cases. Okay, I can only run a batch of 50 and you tried 100, and it's gonna g- Like there's many different things that could happen. and so you gotta think through these edge cases. You gotta test enough to figure them out and then build the right scaffolding in order to deal with them, and then you can do really, really amazing stuff.

Dan Sanchez

One of the things I love about the way Codex laid out its UI is that, one, we have all our projects on the left. We're very used to doing that in ChatGPT. We have chat, we have the chat interface in the middle, and then when you click on these MD files, they open up on the right side, which is very much how Canvas used to work. it's changed now, but, like, it used to open up on the right side as a doc. The only thing I don't like about it is that I can't edit it natively in this right side. That would be a improvement. Yeah. But now I can just open it up. It has a little button where you can just open it up, and mine opens up into Sublime Text, which is kind of a coding app, but it's fine, and I can just make changes here. Yeah. And that's cool. But, like, it's just little tiny things like that I find add up over time. It's just more intuitive and easy to deal with. have you seen-- this is kind of a bonus feature of Codex. It's completely unnecessary, but I, I still find I like it. It's the pets file. You can wake up. There's these little characters- What did you click? No, I did not see. Have you not seen these?

Isar Meitis

No.

Dan Sanchez

There's these little characters you can make, and I custom-made this one. This is based on a character I developed for my own show, "The Edge with Marketer." If you look at the cover art, it has this character I call 404. Actually, he named himself. But these fun little, like, things are these little characters. The cool thing about them is they stay on top over the window. So if I switch to doing something while it's doing some long work, it'll keep me updated on the work that's happening. Ooh. And then as soon as it'll-- as soon as it's done with something that maybe takes 20 minutes, 'cause not everything takes overnight, right? A lot of things you're doing make... It works for, like, 10, 15 minutes. But you go and answer some emails on Gmail, and then the c- the chat, it's giving you updates, 'cause no matter what, what window I go to, it's kinda coming with me, right? And then you just click on, and it brings back the Codex thing that it was telling you updates about. It's a funny, little, tiny UI thing that I think is kinda fun because it's this little funny character hanging around with you. But it's called the Pet, and it's just a fun, little nice addition they did. It is? They didn't have to make it fun, but they did.

Isar Meitis

Very

Dan Sanchez

cool. It just dances on the screen. one last thing is you may not... I'm doing this for a lotta knowledge work. I'm doing all the kind of stuff I used to do in ChatGPT, I now do in Claude for a lot of this work stuff.

Isar Meitis

In Codex.

Dan Sanchez

But... Yes, in Codex. But I'm also using it to code, and it's so much fun. Now, I think where the, where this really gets fun is for marketers and other business leaders who don't code a lot, is to build the little integrations you wish existed that didn't exist before. For example, it's very common for marketers to create social content in something like ChatGPT. But what happens to it? You might even put it in a canvas, but you're gonna have to copy and paste it and move it over to its, um, final destination. It's kind of a pain in the butt, right? So I use HighLevel as my kind of marketing platform. It does, all the email stuff and CRM. It does, it does a bunch of different things. But one of the tools it has in it is a social media management and scheduling tool. And I was like, "Great." I'm like- Codex. Every time I say post this to social, I want you to do it through high level. Help me figure out how to make that happen. Of course, it's like, okay, well, we can work through the API. Here's-- And we're gonna have to code some code files and then, and, and just make it so, and write the MD file to say, like, every time you say this or make a skill for... I, I probably wouldn't do it as a skill 'cause different clients would need different tie-ins. I've actually done this for myself and a couple clients, um, so that when I'm working in their folders, it only accesses their high-level account. But every time I say post this to social, it works through the API. Because of the code it's developed, it had to build its own little, not a lot, but like a few code files in order to access the code, tap into the API, schedule it within the tool, or post it immediately, depending on what I specify. So now every time I see something, I'm like, "Great, send it," or, post it to LinkedIn." It just posts it. It's just done. I don't even have to do anything. I don't have to copy and paste it. Shoot, you can put that on a loop so that every time I run through a process, like maybe after I walk through a podcast, after I finished a podcast and it's got the transcript, I'm like, "Great, run the process." And the process includes turn it into social posts, turn it into a blog post, make sure to upload it to YouTube, so it's got all the files, and then it just runs through the process. I don't have to tell it to do all the individual stuff. Again, all the process is written down in an MD file, so when I say, "Run the process," it's like, "Oh, he means this MD file. Let me check. Okay, MD file says access high level. Okay, hit and do this. Okay, oh, for the blog post, I'm gonna have to access WordPress post." So I could just give the command, and it goes and accesses the code we've already developed. So code is part of the repertoire here, but you don't have to be a developer to do it, and these are, like, lightweight. I'm not building an application. You don't need applications anymore. Codex is the application. It's just developed using code now to extend the functionality. So I'd like... I know there's plugins. I probably need to get into plugins sometime, but I don't really care because I can just build my own plugins. I can build my own integrations. Any tool I have that has an API is fair game now. it's the easiest way to integrate with anything and everything in the way that you want them to be done, and now you just tell Ch- Codex to do it, and it can get it done for you.

Isar Meitis

Yeah, amazing. Uh, two cents on that. there, there are obviously benefits in connecting third-party plugins because they're plug and play, and then the AI can use them. There are disadvantage in doing that. One of them is that many of them has a gazillion things you don't need, and then it just fills up your context window with a lot of instructions and a lot of pieces- Yeah of code that it don't actually use, and that's one reason not to. A- and I'm not saying don't, I'm saying there's, there's- Yeah there's reasons to and to not do this. The other reason, if you are going to, uh, get plugins from third parties you don't know, then you're running the risk of these third-party plugins doing things you didn't plan for them to do, such as copy your data to a third-party server and use it whenever they wanna use it, or whatever other malicious thing that may, they may wanna do with it. Which I assume is not the case for probably 80% of plugins and connectors in the world, but it's very hard to know one from the other. And so that's another very good reason. If you just need a simple connector, again, to post something, to do something... and as I mentioned, expand this beyond the marketing world. Think about your ERP system. Think about your HR platform. Think about whatever other software you're using in the company. And I'm not saying go ahead and now develop this, especially if you're in an enterprise. There needs to be guardrails and governance and things you should and shouldn't- All the stuff connect to and change and- Yeah and so on. I'm, I'm putting that aside for a second. But from an... the having the ability to do it, it's at your fingertips. I don't write code. I've never written code, not even HTML like Dan did. I just don't know. Not CSS, not nothing. JavaScript, any of this, I don't have a clue. That's doesn't matter anymore. Now, to be fair- Not at all people who know how to write code can do more faster.

Dan Sanchez

Yes.

Isar Meitis

But People who didn't write code and don't understand code can still do incredible things. And all the stuff you don't know, and Dan just mentioned that before, "Okay, tell me how to do this." "Oh, go here, go to your calendar, go to this, go to the menu, go to this. It will have a place to create a, generate an API key. Generate the key, don't give it to me, go and place it in this folder, put it in your keychain, and then I can use it." I'm like, "Oh, okay. That, that I can do. A monkey can do that. I'll just follow the instructions." And so you can really do more or less any knowledge work you can do on a computer. Codex, Claude Code, Claude Cowork. Now there's, as of this week finally, Copilot Cowork, which is basically Claude Cowork inside of Copilot. Uh, Gemini is coming out with something very similar. All of these is gonna have all of these functionalities built in, and you pick the one that works better for you, that you already have access to, that you trust more. It doesn't really matter. They all can do these things, but they're all a whole different universe than just write a chat and going back and forth and copying stuff and moving it around on your own. And so whichever platform you choose, go and upgrade from the basic chat to the next agentic level, uh, because it will change your world- Yep. And your view on what AI can do dramatically.

Dan Sanchez

This will be the default way we all use AI a year from now. By, like, I- Until the next thing. I would, I'd bet big money on it. We will all... even if they have to change ChatGPT, whatever we're all using in the future, it will be like what we're doing here with Codex right now. I agree. Just like, remember, it was only a year ago that we had ChatGPT o3, which was the reasoning model attached to internet search. Oh, yeah. Think about how much of a game changer. That was only a year ago, and now it's default. Everybody does it. It's just part of how AI generally works now. You don't even think- Yeah about the old way where it couldn't check itself on the internet. Yeah. That's just default. That was brand new a year ago.

Isar Meitis

Yeah. Crazy. Insane. Dan, this was fantastic. Great concepts and great real practical examples. If people wanna follow you, work with you, learn from you, what are the best ways to do that?

Dan Sanchez

Best way to do that is to find me on AI-Driven Marketer on any podcast platform and linkedin.com/in/digitalmarketingdan. Connect with me there. Those are my two most active platforms.

Isar Meitis

Awesome. Thank you so much. This was absolutely fantastic.

Dan Sanchez

Thank you.

Head Shot

Isar Meitis

Host