Podcast: Play in new window | Download
Whether you’re just starting out or looking to refine your AI strategy, there’s a clear path to getting faster, better, and smarter with AI. In this episode, Dan Sanchez, AI Strategist at Social Media Examiner, breaks down practical ways to integrate AI into your workflow – helping you speed up repetitive tasks, improve content quality, and even use AI as a personal coach to level up your strategy.
Faster, Better, Smarter: How to Make AI Your Marketing Superpower
Let’s be honest—most of us are drowning in marketing tasks. Social media posts, email campaigns, content creation, data analysis, and on and on. The to-do list never ends, especially for small business owners and marketing managers trying to do it all.
I’ve been there. And I bet you have too.
But what if I told you there’s a way to get more done, improve the quality of your work, and become a better marketer—all without adding hours to your day or another person to your team?
That’s exactly what Dan Sanchez, Senior AI Marketing Strategist at Social Media Examiner, shared on my latest podcast episode. And unlike most AI conversations that leave you feeling overwhelmed, Dan’s approach is refreshingly practical.
The “Faster, Better, Smarter” Framework
Dan breaks down AI adoption into three progressive stages that anyone can follow, regardless of technical expertise:
- Faster: Use AI to speed up what you’re already doing
- Better: Leverage AI to improve the quality of your work
- Smarter: Employ AI as a coach to level up your marketing strategy
This framework isn’t about replacing you or your team. It’s about amplifying what you’re already good at and freeing you from the tasks that drain your creative energy.
Let’s break down each stage.
Stage 1: Going Faster with AI
We all have those repetitive tasks that eat up our time but don’t necessarily require our full creative brainpower. These are perfect candidates for AI assistance.
Dan suggests starting with a simple inventory:
“What are you working on on a daily basis? What’s a daily task you have to do? What are you doing every single week? Do you have to write a report? Do you have to post this thing? What are you doing monthly, quarterly, maybe even annually?”
For example, one marketer Dan worked with had to manually analyze customer survey data and create weekly reports—a process that took hours. By using AI to summarize the responses and identify patterns, that same marketer now spends those hours actually implementing insights from the data.
I’ve experienced this too. My team recently needed to analyze hundreds of survey responses for a client. Instead of painfully reviewing each one manually, we used AI to analyze the overall sentiment and pull out specific quotes worth human attention. What might have taken days took hours.
The key is identifying tasks that:
- Are repetitive
- Follow a clear process
- Don’t require your unique human judgment for every step
Stage 2: Getting Better with AI
Once you’ve mastered using AI for productivity, you can start using it to improve the quality of your work in ways that wouldn’t be possible otherwise.
Dan shared a brilliant example from his own work. He created a 5-day email course on AI fundamentals, but instead of sending the same generic content to everyone, he used AI to create personalized action steps based on each student’s job title, market, and product.
“Now every single lesson is customized, hyper-personalized just for them,” Dan explained. “Because the lesson’s going to look very different if you’re the CMO of a beauty company versus if you’re an email marketing specialist selling accounting software to accountants.”
The best part? Once he set it up, it ran automatically for hundreds of students without additional work.
Another powerful application is using AI to incorporate authentic customer language into your marketing. Instead of manually searching through testimonials when creating content, Dan recommends feeding all customer testimonials into a custom GPT so you can easily pull relevant quotes and mirror your customers’ language in your marketing materials.
This isn’t just faster—it’s better marketing because it speaks directly to your audience in their own words.
Stage 3: Getting Smarter with AI
The third stage is where things get really interesting. Instead of just using AI as a tool, you can use it as a coach or consultant to improve your own thinking and strategy.
Dan explained how he created a custom GPT loaded with his core values, mission, and life goals to serve as his personal coach. While initially limited, today’s AI models are now capable of leading through thoughtful questions rather than just providing answers.
“AI can lead you as a coach, as a consultant,” Dan said. “You tell it what kind of coach or consultant you want it to be. You give it the data and say, ‘Hey, how can I make this better? What considerations am I missing?'”
This approach resonated with me. I’ve been experimenting with creating a custom GPT based on my company’s values and mission, essentially creating an entity I can bounce ideas off of to ensure we’re staying true to our purpose.
Whether you’re an email marketer looking for feedback on your campaigns or a content creator wanting to improve your writing, AI can serve as your on-demand coach—available 24/7 and never too busy for a quick consultation.
Custom GPTs: Your Secret Weapon
Throughout our conversation, Dan emphasized the value of creating custom GPTs—specialized AI assistants trained on your specific needs and information.
When I asked whether it’s better to create a custom GPT or just craft good prompts, Dan didn’t hesitate: “I almost always go the custom GPT route.”
Why? Because custom GPTs allow you to:
- Avoid repeatedly entering the same context information
- Keep client information separate (avoiding “cross-contamination”)
- Create specialized tools for specific recurring tasks
Dan recommends two types of custom GPTs:
- General GPTs like “brand bots” that contain basic information about a company, its products, audience, and writing style
- Specific GPTs that perform one multi-step task exceptionally well
One of Dan’s favorite examples is his “Showrunner” GPT, which automates podcast pre-production by:
- Researching guests based on their LinkedIn profile
- Suggesting episode angles and titles
- Creating question outlines
- Drafting guest emails
“It took what usually takes 30-45 minutes and condenses it down to about five,” Dan explained.
Preparing for AI Agents
Looking ahead, Dan predicts that AI agents—autonomous AI systems that can perform more complex tasks with minimal supervision—will become increasingly important in marketing.
He likens working with these agents to managing interns: “I would train them how to write a blog post and then I would monitor it. And then I would just let them go. And I would check in with them once a week.”
The key to success, according to Dan, won’t be technical skills but rather your ability to define clear processes that achieve consistent results. He recommends building custom GPTs now to develop that skill set:
“Start getting the reps of building these things by taking inventory of what you’re doing daily, weekly, and monthly and see what you can automate, and then do another one, and then do another one. And quickly, you will become the go-to person in your company.”
Getting Started Today
If you’re feeling overwhelmed by all this AI talk, take a deep breath. Remember when social media felt equally intimidating? Now it’s just part of the job.
The good news is that you don’t need to transform your entire marketing operation overnight. Start small:
- Identify one repetitive task you perform regularly
- Create a custom GPT to help with that task (or use an existing model like ChatGPT)
- Once that’s working, move on to another task
Before long, you’ll have built an AI-powered marketing system that makes you faster, better, and smarter—without requiring a computer science degree.
And as Dan pointed out, now’s the time to get ahead: “It’s still really early. I only expect it to cross the chasm and become mainstream, just with marketers, probably by the end of the year.”
Will you be ahead of the curve, or playing catch-up?
Faster, Better, Smarter: Making AI Work for You Episode Transcript
Rich: My next guest is the senior AI marketing strategist at Social Media Examiner, where he makes AI less intimidating and more practical for marketing leaders like you. With over a decade in the game, he’s all about helping marketers work smarter, not harder, using custom AI solutions. When he is not talking bots and strategy, you’ll find him running, homeschooling his kids, or diving into his next big idea.
Today we’re going to be looking at practical uses for your AI in your marketing with Dan Sanchez. Dan, welcome to the podcast.
Dan: Thanks for having me on.
Dan: Now, even though we’ve had ChatGP around for a few years, there’s still plenty of business owners and marketing leaders who are skeptical about AI. Similar, maybe, to how they looked at social media 10, 15 years ago. What are some practical ways that you feel marketers can start incorporating AI into their daily workflow?
Dan: It’s always hard when there’s a new thing on the table and it’s like a paradigm shifting thing. It’s not just slightly better, slightly faster way of the way we used to do things, but AI is presenting a whole different smorgasbord in front of us.
I remember when social media became a thing, business owners were like, “I don’t see how posting about what I ate for lunch is going to help move the needle for my business”, which is right. But it took us a while to figure out like, how do we actually apply this thing?
And now, and then it took time and people, smart people like Michael Stelzner and Pat Flynn, wrote blogs and figured out how we can answer our people’s questions, and then post that to social media and then it can help.
Now with AI, we have a whole new thing, and we’re trying to figure out how to best leverage it. So even though it’s been out for two years, we’re still figuring out all the use cases. Some of those ways have been self-evident. Like we write lots of content, maybe ChatGPT can help us write some of that content.
But I find that with most people, you got to break it down into baby steps. You got to start with crawling before you can get up and start walking, before you can actually take on a full-on sprint. You just can’t expect to go from laying on the floor with AI and then running as fast as you can down the field.
So I found that if people take it and use AI solutions to do what they’re already doing, they can essentially just go faster or improve the quantity of what they’re doing. And then they can use it, once they’ve done that for a while, they can use it to do better and actually improve the quality of what they’re doing.
And then when they’ve done that for a while, they can actually start to learn how to get even more out of AI by becoming smarter and getting AI to essentially not only just do your work faster and better, but actually start coaching you and making you better.
Rich: I love that. And I know you’ve got that framework of faster, better, smarter. So let’s take those one at a time.
So faster sounds like it’s the lowest hanging fruit. It’s the first place for people to start. If we’re just getting started with AI, what are some of the easy ways that we can start doing our job faster using some of the AI tools that are available to us now?
Dan: Sometimes people are doing the same things over and over again. In fact, we do this throughout our lives, right? We are creatures of habit. We do the same things in the morning. We do this, we get dressed, we brush our teeth. We usually do them in the same order. Shoot, you go through the shower and you probably wash your body or wash your hair first, and you do it the same way every time. We are creatures of habit and we do it with our work.
So the first thing I usually do when I sit down with something, I’m like, dude, what are you working on? Like on a daily basis, like what’s a daily task you have to do? Interesting. And I’ll move on and go, okay, let’s write that down. What are you doing on a weekly like ritual? Like what are you doing every single week? Do you have to write a report? Do you got to go and check something? Do you have to post this thing? What are you doing monthly, quarterly, maybe even annually.
But I try to take inventory of all those different things and be like, which one is the biggest pain to you? Which one takes the most time, and all of a sudden you find a whole out of that whole list, you’ll start to find some pain points of things that are reoccurring that you couldn’t have already automated it. Maybe you could have automated it, and you should have done it a while ago. Maybe it’s something you can’t eliminate. You have to do it. You can’t automate it. You can’t delegate it. You’re like, okay, this might be an opportunity for AI. So you start to pick apart the thing, and then you break that down into baby steps.
Whenever you do this weekly task, for example, maybe you have to write a report. Where do you get the information from the report? How do you analyze the report? Where do you then document this, and then who do you send it to? You start to break down the task and you start to find that some of this can be done with AI. Oh, that’s a time-consuming part.
You know, what if we actually copy and paste it and drop it into ChatGPT, we might be able to speed up the analysis. Not that you’re not looking at it, but AI might be able to take a first pass and speed up what you would’ve found and maybe even find something you might’ve missed. That’s an example of just taking a weekly item off of someone’s plate and speeding it up. It’s something they’re already doing, maybe doing it manually.
I remember interviewing a guy where he was just bringing insights from a survey that a few hundred people were filling out, and he would have to summarize it in the Excel sheet and then actually provide a weekly report to the higher ups on all the different survey fills. So they wanted to stay on top of it, which is great practice. AI has automated it completely. So he engineered it so every time a new survey comes in, it summarizes the sections and then does a summary of all the summaries in order to instantly create a weekly report of it.
So now he can get into actually diving in deeper and making the most of those insights and pulling them into the marketing. Because oftentimes they’d read the report and inform their guts, but not actually work it into the marketing. So now he’s spending more time doing that. So that’s an example of where you can use AI to go much faster, to free you up to do the things that only you can do, to do the things that honestly, it actually makes you more human because you can spend the time doing the stuff that you couldn’t do.
Instead of spending more time doing the manual monotonous stuff, I can spend more time actually engaging with my audience on social media, for example, in the comments and actually have real human conversations. In fact, just to put out a warning, that would be something I would never automate. For those of you out on LinkedIn automating all your bot comments, it’s obvious to everybody, it’s not working.
Rich: That’s terrible. Yeah. Especially because your bot’s having a conversation with somebody else’s bot, and really nothing is getting done.
I do love your example though. And I had a very similar situation when it … because faster is really about productivity. And as an example, we had to review a whole bunch of surveys for a client, which normally we would’ve had to go through by hand. Just hundreds of different people answering the same 10, 15 questions.
And with AI we were able to have it quickly analyze it, give us some of the overall sentiment, and also pull out specific quotes that were worth reviewing as humans. And to be able to pull this, what are the best quotes? What should we be looking at? So that’s going to make us more productive. It’s going to make our jobs go faster or the specific tasks we have. So fantastic. Love it. Let’s talk about how we can actually use AI to get better at our jobs.
Dan: Man, this is when it starts to get fun because AI is different. So it starts to unlock new possibilities that you just couldn’t do before.
I remember when I got deep into Infusionsoft, because again, I was trying to make everything fast. I was trying to automate everything. It’s like this small business marketing automation platform. They’re one of the first ones to do drag and drop journey builders in their system. I had automated it to its fullest extent. I maximized it even to the point where people were like, this feels a little robotic. I had to pull it back. It went too far.
With AI, there’s whole new possibilities of what we can automate and still make really good. For example, I recently put out a course on AI Fundamentals to take people from just dabbling with ChatGPT to getting really good outputs. Like getting to the point where they could just one shot at one prompt to get exactly what they want.
So I did a typical five video, five-day email course. It’s sprinkled out over five days. One email. The video’s embedded until you click through and watch the video. So that is cool. And if I wanted to use AI to make it faster, I could have had a little summary of each lesson in the email. Cool. That’s faster. Not bad. Could have done that by hand. But AI could do it faster based on the transcript lesson.
Ah, but what if I want to make it better? Now I can use my CRM marketing automation platform called, High Level, and I find that HubSpot. And a lot of the marketing automation platforms are building this in natively. Now, I can actually collect some information about the students taking this little mini course. And in this case, I’m taking their job title, what audience or what market they serve, and what their product is. Just those three pieces of information.
So then I take the transcript of the video lesson, which they can still watch the video, but in the email itself, I’m actually giving back custom action steps based on those three pieces of information. So now every single lesson is customized, hyper-personalized just for them. Because the lesson’s going to look very different if you’re the CMO of a beauty company versus if you’re an email marketing specialist selling accounting software to accountants, right? That the custom, the action takeaways are going to be different, how they might work it into their daily lives. So that’s an example of how something’s getting better.
Now I set that up. I set it up last summer, I forgot about it, but hundreds of people have now gone through that course and gotten highly actionable, personalized thing. And that becomes the bonus part of the lesson of look what AI can do. So that’s an example of better, hyper-personalization is almost always better if you do it right and do it with the right places in your workflow.
Rich: I love that idea. Because we often talk about putting together best practices docs for our clients, and I’m concerned that they actually don’t read them, because who wants to read best practice docs? But if it was instead customized where, hey, I know that you’re in the manufacturing arena, or I know that you’re in the healthcare arena and having an example based on your course that’s specific to those industries, I could see how that would be much more valuable to the end user.
And again, once you’ve set these automations up, used AI or used your CRM like you mentioned, that process becomes automatic. And of course you can go in and improve it over time. But the bottom line is now you’ve taken care of that and it’s happening on its own. As you said, you almost forgot about it. It was running on its own. You were off doing other things, becoming faster, better, smarter.
Alright, so that’s better. Let’s go into the third pillar which is smart, smarter.
Dan: We’re going to the third pillar. Okay. Yes. I do want to offer one more example, because there’s going to be so many more case studies of this, but I find that I’m always hungry for more case studies, like what have people done to make things better, not just faster.
But one thing that I’m working on right now is actually incorporating all my customer testimonials, nice things that people have said about me, and I just have them in a big file. You can attach that to a custom GPT, for example, as just a file or maybe a custom GPT project or a cloud project. And every time you make content, you can always reference and go find some relevant client testimonials that speak to this thing.
It’s not something you would’ve done before, because you’d have to go and scan a thousand different records. And the keyword search isn’t quite right, but AI can go through them all, look for the right intent, and then pull them in as you need. So every time you’re writing a blog post about your product, every time you’re doing a product, any type of product marketing, you could be pulling in actual words. You could say, “Hey, use the language that my customers use when describing this thing” and then just go pull it from the data. It’s not something that you could do easily, and therefore most people just skip it.
That’s why personas are good. They’re just summarized versions of all that data, but now you can just pull directly from the data. Like the personas are still good, but now you can just pull directly from the customer’s mouths themselves whenever making your product marketing content. And that’s another example of how it can use this stuff to get better when writing a blog post or social updates or making a webinar and scripting it.
So I’m thinking those are small, little ways that you can get better. The course is a good one too, though. That one is a little bit more of a lift. But I will say like anytime you could do hyper-personalization, it’s good too because you do the work upfront, but now it’s set and forget, that thing’s just operating without me now.
Rich: Brilliant. Alright, now I’m ready to learn about how way AI can make me smarter. Yes. Hit me!
Dan: Smarter. One of my first things that I did when I started digging deep into AI, this was not quite possible, but now it is. It’s my first failed attempt at building a custom GPT. It was December of 2023, and I saw the writing on the wall. I’m like, AI’s going to be a thing.
And I had just gone through this course from Michael Hyatt from Full Focus about how to build your best life kind of a thing. And it walked me through all my core values and my mission and all these things and goals and all that kind of stuff. So I just finished all this stuff and I put it all into a custom GPT, all the docs that I had created. And I’m like, “Okay, your goal, custom GPT, is to be my life coach. Keep all these things in mind. Here’s my mission, here’s my core values. I want you to lead me by asking questions. I give you a problem, you help me think through it by asking questions.”
Honestly, at the time, it couldn’t do it. It just would try to come up with answers rather than ask questions. If it asked a question, it asked multiple questions instead of just one at a time. I literally just tried that same GPT again and it’s working. It’s like even though we’re still on the floor model, it’s gotten much better, and it can actually execute this.
AI can lead you as a coach, as a consultant. So you tell that it what kind of coach or consultant you want it to be. You give it the data and say, “Hey. How can I make this better? What considerations am I missing from this thing I’m trying to accomplish?” And you can even go big objective of, “Hey, I’m running into this problem. I’m not even sure I know how to define it, but here’s what I know. Coach me through this. Only ask one question at a time and help lead my thinking through this problem.” And you’ll find that ChatGPT is a very capable coach.
The one drawback to this is that it’s not like a human coach that can intuit things that it wouldn’t know to intuit from the situation. Like it might not intuit that, it sounds like you’re not getting enough sleep, even though you’ve only been talking about work. Where a human coach will be like, you look tired. I think that’s your problem. How do you work out, you seem run down, but I don’t think it’s a work problem we’re dealing with. A human coach will be able to make that connection. AI can’t yet, but you’d be surprised how much of us actually hire coaches and consultants to think through problems sometimes, but not nearly as much as we could probably leverage. But with ChatGPT, we can.
If you’re an email marketing specialist, you can now have it coaching you on your emails and giving you feedback regularly. If you speak for a living and speak in public or do podcasts like this, you can have AI look at your transcript and provide feedback on where you could have tightened it up and provided better illustrations. And that’s an example of using AI to make you smarter. Though I’m starting to build it more, even into the custom GPT that I’m using into all the different areas. Automating it so that it can actually help me get smarter, faster, and more regularly.
Rich: I love that whole idea. And I think I mentioned it with another one of my guests recently. But along those same things, I didn’t think of the life coach approach, but I started to create an entity for my company, flyte new media, based on all the things that I wanted flyte new media to be able to accomplish. My mission, my values, where I want to see how flyte can help other businesses and other people, and just be a force for good in the universe. And I basically have started to create a custom GPT around this so that I can query it and ask you questions.
Is this the right direction for us? If this is the fork of the road, what advice as the company, the entity of the company, would you give me to make sure that we’re staying on track? So I think that just being able to bounce ideas off of AI, even if AI is far from perfect, sometimes it’s just a good mental exercise. Sometimes you just need to talk it out. And if you create these AI, these custom GPTs, I think that’s a great way to do it.
Which is a great segue, because I do want to talk to you a little bit about custom GPT or Claude projects or Gemini Gems, whatever your preferred flavor is. How do you determine whether you should create a custom GPT for something, versus just a really good prompt stack? Do you have a mental trigger in your mind where you’re like, this really is going to be custom GPT versus, I think I can create a series of really well-crafted prompts to attack this problem?
Dan: I almost always go the custom GPT route.
Rich: Really? Okay.
Dan: I hate having a prompt library. I don’t want to have to copy and paste it. If I find that I’m asking AI, if I’m having to load a really long prompt more than once, I’m almost always going to make a custom GPT to do it.
Now it’s funny, I’m recording an episode on this now for my podcast, AI-Driven Marketer, about how to stack information and how to customize ChatGPT specifically, but you could do this with any of the models across the account. Because there’s different ways to customize it. You can customize it at the account level, which gets applied to all custom GPTs, all projects and all chats, and it has custom instructions, which changed recently.
If you haven’t been in the custom instructions of GPT it’s more nuanced now. It asks you what’s your name, how do you want it to address you, what do you want it to know about you? How do you want ChatGPT to behave? Do you want it to be witty? Do you want it to be skeptical? Of what you do you want it to challenge you more? You put it there and then you can load it up with like random custom instructions.
For example, the custom instructions, every time I ask ChatGPT to write a blog post, I almost always have five major points, and I’m like, blog posts should always be X, Y, Z. It should always have multiple types of chunks broken up, bulleted lists, headers. It should always be written into first person, directly to the reader.
There’s always multiple things that I do with the blog post. I don’t care which custom GPT I’m using across clients, across employer, personal stuff, I always want it this way. So I load that in the custom instructions. It’s at the account level and there’s also the memory feature, all the little things that it remembers about you. There’s different ways to load that up with helpful things.
But then custom GPT, I always alarm with information about you have client projects, you don’t want a client project information loaded in across the account, that wouldn’t make sense. You don’t want to cross contaminate different projects you’re working on. That’s when you use custom GPTs or projects in order to organize that information at the instruction level. So you don’t have to keep retraining it on the basics, oh, this is who this company is, this is their products, this is their market, here’s their audience, and I’ll act accordingly. I get tired of reminding it of all that stuff. So I generally have a lot of custom GPTs and projects laying around to keep those things separate and keep the chats simpler.
Rich: So your advice for agencies, or any company that’s going to work with multiple people or multiple businesses, is that we should be creating custom GPTs for each one of those businesses to, as you say, avoid that cross contamination.
Dan: Yes. And I have two different custom GPT styles that I use. And this gets more complicated with projects, but there’s essentially two GPTs that I build. There’s general GPTs, and there’s specific GPTs.
The general ones are, general, you could do a lot with them. Every client that I have or employer that I have, I build what’s called a ‘brand bot’, which is just armed with the general things it needs to know about that particular business. So I have a Social Media Examiner brand bot.
Rich: So what are some of the things that you would put into a typical brand bot? I’m guessing it’s things like, ideal customer, competitors, industry. Anything else?
Dan: I usually don’t put competitors in there. It’s usually just what is the business, what are their products, what is their business model? Who’s their target audience? What is their writing style like? What are the things that are really important to them? What’s their position in the market? It’s just all the really base level things. It’s the things you would have to give to a marketing manager if you hired a marketing manager. What would they need to know to get started? It’s that same stuff. It’s just like a human, you got to train up humans to do that kind of stuff. You need to onboard your AI to be able with that basic information.
If a client’s written a book, that’s pretty critical to the thinking. I upload the whole book. I just put the PDF in there. Hopefully it’s a more summarized version. If it’s a 60,000-word book, I might try to get an abridged version, because the context window is only about a hundred thousand words. So I put those in there too, in order to train it and keep it online with the message.
Rich: And Dan, if we can just dig a little bit deeper on this for a sec. So when you bring on a new client, like I know that you recently started working for Social Media Examiner. When you are doing something like that and you’ve got this new client, I’m just curious about your process. Are you interviewing, in this case it would be Michael Stelzner or maybe somebody on his team, how are you deciding or are you scraping their website? What are the steps that you take to really build the best possible brand bot that you can?
Dan: It became something that I did frequently enough that I literally made a custom GPT to build this custom GPT. Again, I don’t like to repeat the model over. I’ve done it three times. I’m like, screw it. I’m just making a custom GPT to go scrape the internet for it, because it’s really basic. It doesn’t have to be a lot.
So I go and I built the… now this isn’t a good example of a specific custom GPT, right? That was a general one, a specific one. It only does one thing and usually takes multiple steps to execute that one thing. So this other custom GPT I made is a specific one because it walks through the same process of building out the instruction profile for a brand bot.
So this one goes through, and it probably starts off with, like it just knows, the first thing you enter to get this thing going is just put the company website URL. It goes and scrapes the website. And then to understand what the company is about, its general product market audience. That goes and scrapes other websites about it to get reviews and different things, and then starts to build a profile and understand the business model.
And it’s specifically going through and looking for all the things that I just discussed, and then summarizing it, going through some researching, summarize, and I think it takes three different passes on research and then create summary. And then from those three summaries, it’s last step will be like, “Hey, do these summaries look good?” It checks with you, and you’re like, yes, move forward. It goes and compiles the instructions that you can then copy and paste to build a brand bot. So that one’s a pretty. Pretty quick and simple custom GPT.
But every time I have a process that’s like multi-step like that, I try to turn that into a custom GPT. One of the best ones I made that’s very popular, because I released the instructions on it that you can copy and paste to make your own, or riff off of and do something more with, is one I call “Showrunner”. And it essentially automates all the pre-production for a podcast. So you might be interested in it because it might help you with your show, but it does a lot of the work that I was doing for every guest on my podcast. And it starts off with, “Hey, who are we interviewing today?” “Is this a solo show or is this an interview show?” It’s an interview show. “Who’s the person?” You’re like, oh, great. You want to drop their LinkedIn profile? I just go and copy and paste their LinkedIn profile, put it in there, just paste it in. It’s like, great. It finds that, analyzes that. Then goes and looks for them online and looks for things they’ve written out there, other things they’re associated with. Builds a summary. Is this okay? You’re like, yes, great. What do you want to interview this person about? I’m like, I’m thinking about going in this direction, or I’m not sure yet.
It’ll then recommend okay, here’s five different angles we can take for this episode. You’re like, huh, I really like angle three. It’s great, here’s 10 different titles for this episode to give it direction based on this angle. And again, it’s thinking about the consideration of the premise of your show because that’s in the instructions, the expertise of the guests. It’s a lot of things to think through, but it thinks through all of them and then comes up with 10 different titles. You’re like, Ooh, title v’s. Awesome. And also make this tweak based on that title, based on that premise, based on your show, based on their expertise.
Here’s an example, outline of the show intro questions. Outro. Does this look good? Great. And here’s an email that you can send to your guest with the questions in there. Let me know when it’s done. Paste the transcript below and we can start working on post-production. So that’s my show runner. It’s nice. All things you kind of work through manually, but it took what usually takes 30, 45 minutes, and condenses it down to about five. So that saves a lot of time. If you’re interviewing a lot of guests, which I don’t know about you, but I’ve been at the point where I’m interviewing multiple guests a week and you’re just like, oh my gosh, can we speed this up?
Rich: Yeah. I go through phases. This is a busy week, and then I might have three quiet weeks, but I can see the value and I definitely leverage AI.
I’ll usually, for example, you and I had a pre-interview chat, so I take the transcript of that, and I put it into both Claud.ai and ChatGPT, and I explain my show. And I probably should be creating custom GPTs and projects at this point, but I ask it to generate 10 questions in both places based on some criteria I give it. And then I pick and choose the best questions or rework them to be a little bit more in line with what I’m curious about.
But yeah, it’s definitely been a time saver. And sometimes it just takes some of that mental load off so that I can focus on being present during our conversation. Absolutely I see how that is valuable.
I did want to, as we move towards wrapping up, I did want to get your take on AI agents. Obviously there’s been a lot of buzz around them. We’re starting to see them almost ready to come out. OpenAI has made some announcements recently. What do you see as the role of AI agents for marketing as we look to the rest of this year?
Dan: Yeah, all the ingredients are on the table to start making these things. Like the reasoning models are good to then make pretty critical decisions and make them reliably. That’s the big thing that wasn’t available with the old models, or the one we use 4.0 now. But the reasoning models are pretty good.
And I think it’ll go through a series of progressions that I thought there was going to be a limiting factor. Because I’m like, oh, it has to be able to access data and then do stuff with it. If you’re in the marketing automation world, you have your inputs or your triggers that kick off the automation, and then the outputs or the things that you can do with the automation, like once the automation’s going, like what do you want it to do? You have to deal with all these APIs. But the operator thing they just announced recently. I’m like, I guess we don’t need APIs anymore because it can literally just do what a human does and just access everything through the controls humans use. So that’s huge.
My timeline for agents keeps getting shortened. I’m like, ah, we’re probably going to see something really basic by the end of the year. But with Operator now, I’m like, it’s probably this summer we’ll start to see some basic ones. I think agents are going to become massively important. I think they’re going to change the way we work.
I used to run an internship where I had about 24 interns in my department. If you’ve ever managed interns, it can be really difficult if you don’t have a very well-defined, standard operating procedures to plug them into in very fast and effective onboarding for them. I think working with agents will be a lot like how I was working with interns. I would train them how to write a blog post and then I would monitor it. And then I would just let them go. And I would check in with them once a week on the stuff.
But I had a very clear process for them to do and they’d have to think through it in order to execute their task. And I think we’ll have all these agents that we can do, “Hey, like what are regular tasks?” Kind of taking inventory of what you’re doing regularly. What can an agent now take over that’s more than just a task but is actually a regular project that it can think through and make steps and actually execute against? Because it can now access the browser, it can do anything you can do because it can get your credentials and do all that kind of stuff.
So I think the key there to be prepared for agents as they come over the next six months to a year, is figuring out how to create those standard operating procedures. Because your ability now to delegate and to train, and is train it and tell it like what excellence looks like, is going to be the difference maker.
It’s not going to be your technical skills. It’s going to be your ability to define a process that gets a repeatable, great outcome, and then delegating it to the agent. But that’s not a technical thing. That’s just you actually being able to break down a project into its most basic tasks. And the best way to start training for that now is to build as many custom GPTs as possible, because that thinking is along the same lines. What can the custom GPT have access to? What can it do? How can it offset your work workload?
Like I did it with pre-production. How many more times can you do that? With custom GPT, you’re going to get some reps into the building processes and working with AI so that when the time comes and you could do much more sophisticated things, you already have a mental model for what AI’s capable of and how to leverage it in small ways so you can jump to bigger ones.
So that’s what I’m encouraging everybody to do. Build custom, especially these specific custom GPTs that are multi-step. You need to start building those because the AI agents are going to need the set of instructions. Like you can’t delegate to agents unless you can delegate to humans, or delegate to get it off your plate.
Rich: Which is a skill unto itself. But we’ll leave that for another day.
As we wrap up, what is one piece of actionable advice you could give to our audience to get started on faster, better, smarter, when it comes to their own AI?
Dan: To swing back through what we already talked about, take inventory of what you’re doing regularly and start to figure out what, maybe take a low lift thing that you’re like, ah, I think I could do this with a custom GPT. Do it. Start getting the reps of building these things by take inventory of what you’re doing daily, weekly, and monthly and see what you can automate, and then do another one, and then do another one. And quickly, you will become the go-to person in your company.
Once you’re the go-to person, people will start to come to you and you’ll start to get ahead, and you’ll start to get a reputation for AI within your company. And once you have that lead, you’ll continue to stay ahead.
Because people will keep begging their problems and you’ll keep, I don’t know what it is, but once you have the lead on something and people start coming to you, it’s easier to hold the lead because everybody keeps coming to you with more interesting problems. You keep getting more and more experience and you’ll stay ahead in that topic.
So now’s the time to get ahead. It’s still really early. I only expect it to cross the chasm and become mainstream, just with marketers, probably by the end of the year, maybe. As in to where like most marketers are doing it regularly or think it’s a meaningful feel thing that they use on a daily, at least weekly basis. So it’s still really early. Most of the population doesn’t touch AI.
Rich: All right, Dan, this has been incredibly helpful. For people who want to learn more about you, connect with you, where can we find you online?
Dan: Linkedin.com is my social media platform of choice. So you can go to linkedin.com/in/digitalmarketingdan. Or you can find my show, AI-Driven Marketer.com and subscribe there or on YouTube.
Rich: Awesome. And we’ll have links to all of those in the show notes. Dan, thank you so much for coming by today.
Dan: Thank you so much.
Show Notes:
Dan Sanchez is an AI Marketing Strategist at Social Media Examiner, where he helps marketing leaders make AI more practical and less intimidating. With over a decade of experience, he specializes in using AI to streamline workflows, enhance content creation, and drive smarter marketing strategies. Be sure to connect with him on LinkedIn, and check out his podcast.
Rich Brooks is the President of flyte new media, a web design & digital marketing agency in Portland, Maine, and founder of the Agents of Change. He’s passionate about helping small businesses grow online and has put his 25+ years of experience into the book, The Lead Machine: The Small Business Guide to Digital Marketing.