Podcast: Play in new window | Download
What if you could streamline your work, simplify everyday tasks, and stay consistent with your brand voice – all with AI? Brian Piper, author and digital strategist, shares how he uses custom GPTs to tackle repetitive tasks, optimize workflows, and even plan family meals.
Revolutionize Your Workflow with Custom GPTs: a Practical Guide
Custom GPTs are transforming how businesses handle repetitive tasks, maintain brand consistency, and streamline their workflows. Unlike standard ChatGPT prompts, custom GPTs let you create specialized AI assistants trained for specific tasks without repeatedly uploading context or instructions.
Why Use Custom GPTs?
The real power of custom GPTs lies in their ability to retain context. For example, you can train a GPT with your brand voice, product information, and strategic priorities. Once configured, team members can access this specialized knowledge without re-uploading documents or re-explaining requirements.
Creating Your First Custom GPT
To get started with custom GPTs, you’ll need a ChatGPT Plus subscription ($20/month). Navigate to the “Explore GPTs” section and click “Create” in the upper right corner.
Key specifications:
- 8,000 character limit for instructions
- Up to 20 resource files
- 512MB file size limit per file
Pro tip: Bypass the character limit by uploading a PDF with detailed instructions and directing the GPT to use that document as its guide.
Best Practices for Custom GPT Development
- Start with Testing
- Conduct 10-15 experiments with various content types
- Test with a small internal team before wider release
- Regularly update and refine based on feedback
- Fine-Tune Performance
- Use a 1-10 scale to adjust attributes like:
- Humor level
- Creativity
- Specificity
- Response formality
- Use a 1-10 scale to adjust attributes like:
- Build Specialized Workflows
Instead of creating one complex GPT, build multiple specialized ones:
- Brand voice checker
- SEO optimizer
- Meta description writer
- Content repurposer
You can combine these specialized GPTs using @ mentions in regular ChatGPT conversations, creating an AI-powered assembly line for content creation and optimization.
Real-World Application: Content Repurposing
Here’s how to leverage custom GPTs for content repurposing:
- Input high-performing content
- Specify target audiences
- Define preferred channels
- Request format-specific adaptations (e.g., “Convert this blog post into a 20-second TikTok script”)
Getting Started Tips
Begin with personal use cases to understand the platform’s capabilities. Create simple GPTs for repetitive tasks you handle regularly, such as:
- Email templates
- Social media posts
- Meeting summaries
- Content optimization
As you become more comfortable with the technology, expand to more complex business applications.
Remember: Custom GPTs require maintenance and regular updates, especially as the underlying AI models evolve. But the time saved and consistency gained make them invaluable tools for modern marketing workflows.
Want to explore custom GPTs for your business? Start with a simple, repetitive task and expand from there. The future of workflow automation is here – it’s time to embrace it.
Creating Custom GPTs for Marketing Success Episode Transcript
Rich: My next guest is an author, award winning international keynote speaker, and consultant. He’s been optimizing digital content since 1996 and has created online training programs for dozens of companies. He spent the last eight years focusing on data analytics, digital marketing, and content strategy.
And since 2021, he’s been diving into AI, Web3, community building, and the Metaverse. He wrote, Epic Content Marketing for Higher Education, and co-authored the second edition of Epic Content Marketing with Joe Pulizzi. He’s the host of the AI For U podcast, and that’s U with the letter U for university, which focuses on practical AI implementation and use cases in higher ed institutions.
When he’s not creating digital data visualizations, he teaches wingsuit skydiving and spends time with his wife and six children. I don’t know which of those statements is more incredible. Today, we’re going to be looking into creating your own custom GPTs with Brian Piper. Brian, welcome to the podcast.
Brian: Thank you so much for having me on, Rich. It’s a pleasure.
Rich: And of course we had a nice little conversation about snowboarding right before we hit the record button. Before we get into the meat of today’s topic, let’s talk about wingsuit skydiving. When and how did you get into that?
Brian: I started skydiving 35 years ago in the military. And then back in 2001, I was with a buddy of mine in Norway doing some base jumping, and we saw a guy take one of these squirrel suits off of the cliff and have a 30 second flight while we were getting 10 seconds.
And I said, I got to figure that out. So then the next time I had an opportunity, I bought one of the wing suits and have been doing it ever since. And now I like teaching other people how to do it.
Rich: Tripling your ROI overnight. So do you, at the end, you pull a chute, or do you take the squirrel suit all the way down?
Brian: No, you still have to pull a parachute at the end to land. If you want to land soft and nice and go do it again, yes, you need the parachute at the end. Otherwise it’s one time use.
Rich: Okay. I get it. So starting with the basics here, let us know what is a custom GPT and how does it differ from just using prompts in regular ChatGPT?
Brian: So basically a custom GPT gives you the opportunity to create a mini model within ChatGPT that you can train and personalized for your specific use case or task. So basically, you’re just creating a smaller instance of ChatGPT that’s specialized for a specific type of project or task that you want to do over and over again.
Rich: If ChatGPT is all powerful, all knowing, omniscient, as it’s sometimes made out to be, why would we need a more focused version? Why couldn’t we just ask these questions the way we ask any questions in a prompt?
Brian: You absolutely could, but you need all that setup. You need all that exposition at the beginning to be able to say, alright, I want you to act like this, and here’s all the information, and here’s all the data.
So for instance, I have one that I use for personal use that’s a meal planner, helps me figure out what the family is going to eat for the week. And I’ve got it all customized with the appliances that I have, how I wanted to answer, the format we’re going to go through, it even gives me my shopping list in the order that I shop in my grocery store. And that’s all because I’ve uploaded a map and the picture of the aisles and traced my route.
So these are all those personalization things that you can add into the instructions and into the supporting documents. So now every week when I need to go figure out what we’re eating for the week, I don’t have to go back and find that prompt in my prompt library, or I don’t have to go grab that super doc and upload that into a new prompt.
I am basically training this very specific use case within ChatGPT, so I can just go to my culinary creator, click on that, and just ask it whatever questions without needing to give it all that background.
Rich: All right. So if we have a task that we’re going to be doing multiple times, possibly it varies a little bit, but that’s when the custom GPT really shines.
Brian: Yep, absolutely. And especially if you need to give it a lot of context. So you could do it with just a prompt in your prompt library for very short, quick, simple things. But anytime I’m wanting to add more information about our brand voice, or I want to give it all of our strategic priorities, and I want to give it this list of different websites that I want to have it go check out to get background data on. You can just put all of those in the instructions for the custom GPT, and then you don’t have to think about those every time you want to go use that. You just know that it already has all that information in it.
Rich: All right. Now, are there any downsides to creating and using custom GPTs, or are there any use cases where they might be less effective, or the outcomes would be less desirable than just using a standard prompt?
Brian: Yeah. The other thing about custom GPTs is you need to work on them. You need to refine them. You need to keep going back and testing and iterating on them. And especially as the model updates, you need to go back in and make sure that now with the new model, with the new update, it’s not making mistakes or not answering your questions the way that you want them to.
Sometimes you’ll give it very clear instructions, especially if you tell it not to do something. It will intentionally do those things. So sometimes you have to play around with taking part of your prompt out and then republishing it, and then going back in and putting that part of your prompt back in and saving that again.
They can be a little bit more work sometimes to refine and get them just right, as opposed to just a standard prompt that you can use. But I find them to be incredibly valuable once you get them fine-tuned. And then as new features come out in the models, like when the Canva feature got added to ChatGPT, you’re able to just go into your custom GPT and just click on that checkbox and add that new feature into your custom GPT.
Rich: All right, that’s nice. Now, we’ve talked about ChatGPT or custom GPTs quite a bit. That’s from ChatGPT, are there equivalents in Claude and Gemini and any other popular LLMs that are out there right now?
Brian: Yep, absolutely. Claude has projects which are different than ChatGPT projects. So Claude has projects and Gemini has gems, and those are basically ways within those other tools that you can build these custom, personalized, little mini models.
Gems gives you the opportunity to play with the background a little bit more. So if you’re more on the developer side of things, that might be a better option for you where you can go in and you can adjust the temperature, the chaos level, or the different attributes of the model itself.
Whereas in ChatGPT, you’re given just the basics of ChatGPT. It doesn’t give you a lot of that background, but they’re very easy to build, very quick to build.
And my wife and I were cleaning out our basement and we had all these items that we needed to list on Facebook Marketplace. And it would have taken us hours to manually do each one of those one at a time and write the description and the title and figure out what the price was. And I made a custom GPT in about 30 minutes, and then all I had to do was take pictures of each of the items and it went and did the research, wrote the description, the titles, figured out what the other items like that were selling for, gave me a price, and then it actually put it all in an Excel file that you could just upload to Facebook Marketplace to list multiple items at one time. So I was able to list 40 items in about 20 minutes.
Rich: Wow. And then you got 400 people who immediately messaged you and said, “Is it still available?”
Brian: That’s right. Absolutely. Once we get that automated, then we’re really going to save some time. Exactly.
Rich: So thinking, putting on my small business owner hat or my marketing hat for a small to medium sized business, what are some of the more common use cases that I might be interested in, in terms of custom GPTs?
Brian: Yeah. So one of the great things that we see a lot of businesses use is creating a custom GPT that simulates your brand voice. So you can train it on a bunch of your writing, or your website, or your blogs, newsletters, whatever. So that then when you need a new idea, it can help you write a draft in your brand voice using the language that you’re used to using.
So a lot of times that can be very effective. Same with emails. So at the university, we have a lot of emails that go out from different leadership areas when something happens on campus. So now we can start off by getting a first draft of that message in the leader’s voice. And then that way the approval process usually happens much, much quicker because it’s in the language that they’re used to using that they’ve used in all their previous correspondence.
Writing proposals, creative briefs, job descriptions, all of those things that are time consuming and you need to provide the input about the specifics of the job. But when it comes to the formatting, the structure, all those sorts of different things that you have to spend your time crafting and putting together, it can spit out a first draft of those really quick, and you can just go in and tweak and adjust and correct things before you finally release it. So those are a few good use cases.
Rich: Yeah, and I can imagine if you have multiple people writing for you and you’ve created a custom GPT for brand voice, as you were saying, that you could also use it for editing. So if I’ve got six different people on my team or if I’m using some outside consultants and they’re writing for me, I can feed it through there and it can give me a version that is more in alignment with my own brand voice, correct?
Brian: Yeah, absolutely. And if you have multiple people across like your business that are creating content for you, if they use that custom GPT, then everything will start out right out of the gate sounding more similar, more consistent, more in your brand. So it’ll take less time on the editing side downstream for sure.
Rich: So as we’re talking through this, I want to make sure that people understand how they get to this. Can anybody use custom GPTs? Does it require a certain level of ChatGPT? What do you call those pay tiers or whatever it would be? And where exactly do we access it if we want to create our first one?
Brian: Yes. So if you want to create a custom GPT, you need the plus version. So that’s the $20 a month version. And then you can build as many custom GPTs as you want. If you want to use custom GPTs, you can use public custom GPTs just with the free version of ChatGPT. You can also use versions of custom GPTs that people publish and they share the link with you.
So you can set the share status of your custom GPT that you build to only be available to you, or you can have them be unlisted, which means they’re not public GPTs, but you can send anyone you want that URL, and then they’ll be able to go in and use your custom GPT. Or you can make them public, in which case they go into ChatGPT, Open AI’s library of public GPTs that you can use. So those are the ways that you can share them.
So you can have one person on your team that’s responsible for creating new custom GPTs, and then different members of your team testing those out, trying them. And when they need changes made or when they need adjustments to the model made, they can just go back to whoever made those custom GPTs and have those tweaked and changed.
Rich: So when we’re in our paid version of ChatGPT, where exactly are we going to see where we can create them, and where are we going to be able to see like the ‘app library’, so to speak of everybody else’s GPTs?
Brian: Yeah, so when you’re in ChatGPT, over on the left-hand side where all your prompts are right at the top, you’re going to see explore GPTs. So when you click on that button, then it takes you to a screen where you can either search for all the public GPTs that are available, or up in the upper right-hand side you’re going to see another selection that is my GPTs. So any custom GPTs that you have created. And then you’ll see another button that says create, and then you click on that and then it’ll take you into the creation process.
Rich: All right. Now you mentioned that it’s a little bit of an iterative process to create a really good GPT. Sometimes, you who are very experienced in this, it might even take you half an hour. Can you walk us through what steps should we be taking if we’re about to craft our first GPT?
Brian: Yeah, so we always tell people when you’re going to build a GPT, there’s a couple different ways to do it.
So it’s either a prompt that you’ve used several times, and you keep going back to, and it has a lot of value. Or it’s a task that is, has multiple steps in it, or it’s a repetitive task that you find yourself doing all the time. So those are when you would create a custom GPT.
And then you have to start thinking about, all right, what are the goals? What are the objectives that I want this GPT to deliver at the end? You want to be very clear up in the front of your instructions. I’m going to be asking you to create a Facebook marketplace listing for items based on a picture.
And then you establish the workflow. So you have to walk the GPT through here are the steps that you’re going to take. Here are the things that you’re going to create, the outputs that you’re going to deliver at the end. And then you give it as much information as you can, as many examples, as much data, as many resources as it needs to be able to create the best output.
And then you start testing. So initially you can test within the tool without publishing it. So you’re just figuring out if it’s actually going to give you the results you want. But none of the data that you input or none of the examples that you use within those tests within the custom GPT are actually saved in your prompts.
They don’t actually start getting saved in your prompts until you publish that GPT, until you actually create it. Then you can actually go in and start using it and getting output and those will show up in your prompt menu.
Rich: Is there any way to reverse engineer someone else’s prompt? If you see a custom GPT you really like, is there a way to find out how it was actually made so you might repurpose it yourself? Or is it more that’s in a black box and you might be able to guess at how they did it, but you won’t be able to see the recipe?
Brian: Sometimes you can just ask other custom GPTs. So it used to be when they first started coming out with these, you could ask custom GPTs for all the files that were used to train it and what the instructions were for it, and it would just give you that. They don’t give you that anymore.
Rich: Probably best. Probably for the best.
Brian: Yes. Because people were uploading their entire books, multiple copies of their different books that they had to train the model or their own little custom GPT. And then people were just getting the free copies of the books from grabbing those. So yeah, they locked that all down.
But a lot of times you can just ask the model what are the workflow steps we’re going to go through? What types of information are you using as resource data? So a lot of times you can get a sense of how the custom GPT was created from those. But a lot of times you just have to use the tool. And then if you want to duplicate that, you could start experimenting with your own model.
Rich: All right. Now you mentioned another time when we were talking about a 10-point scale you use to refine these GPTs. Can you explain that a little bit?
Brian: Yeah, so a lot of times you’re going to have a GPT that will either be more enthusiastic than you want, or it will have a higher level of humor in the output than you want. So what we’ll do instead of going in and saying, “use plain language” or “don’t be funny”, we’ll say, “on a scale from 1-10, I want your humor level to be a six.” And then you’ll look at the outputs. And if it’s still too funny, then you go back in and you say, “now I want your humor level to be at a three”. So you can give it a scale like that, and it will adjust itself up or down incrementally based on that scale. So there are a lot of different attributes you can add in there to adjust.
But if you find something in the tone or the voice of the model, that’s not being effective or if it’s hallucinating a lot, you can say, “I want you to turn your creativity level down to a two, and I want you to turn your specificity level up to a nine”. So those are the sorts of different attributes that you can give it based on the feedback you’re getting, the output you’re getting.
Rich: Brian, is there like a master list of these attributes or are you just using your own personal sense of style or your own ideas about how to tell it to be funnier? Which I don’t believe it can be funnier. It may try to be more humorous, it’s never going to be funny. But can you… is there a master list, I guess, is my question?
Brian: I don’t really have a master list. It’s really situational based on if there’s a general part of the output that I don’t like. If I think that it’s trying to be too humorous and trying to put too many jokes in there, then I’ll use that as the gauge of whether or not to increase or decrease it. So it’s very situational depending on what the output is and what you’re trying to adjust.
Rich: And as far as you know, there’s no master list at OpenAI that they’ve created with hundreds or thousands of different attributes that basically could be toggled on and off for 1 through 10?
Brian: No, it’s not like Midjourney that has very specific attributes. Yeah, nope.
Rich: And then the 1-10 rule or the 10-point rule, is this something you’ve come up with? What if I said from 1-100? Or is that something that’s hard coded in and it’s 1-10?
Brian: Nope, no, I think you can use whatever scale you want. I got that idea initially from Ethan Malik, who’s a Wharton professor, AI expert, posts a lot on LinkedIn. And he was saying that’s one of the best methods to use to adjust the output instead of just saying, “don’t be so funny”, or, “try to use more plain language in your outputs”.
Rich: Now at flyte, I have my own personal plus account. And then we also have a team account for everybody else. There’s just things that I don’t feel comfortable sharing with the rest of the team, which is what… And I started first, so that’s why I have my own one.
What are some of the ways, what are the best practices for being able to share custom GPTs with your team so that we’re all using the same playbook?
Brian: Yeah. So if you have a ChatGPT teams account or an enterprise account, that’s even easier to share it because then you can just share those custom GPTs with other people on your team. So you don’t even have to make them unpublished. You can actually just keep them. Either you can share them only by yourself or with your team, or you can make them unpublished or private.
And that’s actually what Moderna did. They bought GPT enterprise for their entire 6,000 person staff and they gave them all access to that. And within two months of adopting ChatGPT enterprise, they had over 750 custom GPTs across the company. So people were not only creating custom GPTs for themselves, but they were seeing other custom GPTs that other teammates had created that helped make their job easier. And so then they were looking for opportunities to collaborate together and create these custom GPTs that would help everyone at the company.
Rich: All right. Now I’m sure a lot of businesses make mistakes when they’re first getting into custom GPTs. What are some of the biggest mistakes you’ve seen that companies make when they start entering this arena?
Brian: Yeah, I think the biggest one is just not testing enough. I think a lot of people, when we first came up with our Meta description writer for website pages, we have a bunch of different SEO custom GPTs that we’ve created. So one just writes Meta descriptions. And we came up with it. We tested it a couple of times and then we sent it out for a bunch of people to use.
And we started getting feedback about it making up things that were on the page. So then we developed a process when we were testing these things, and we came up with a workflow of we want to do at least 10 or 15 experiments with a variety of different types of content, different types of output we’re looking for, and kind of pressure test it a little bit more before we start releasing it.
And now we release it internally to a small team to test it. And then once that testing comes back and we get those adjustments made, then we release it to a larger internal team. And depending on the prompt, then sometimes after that goes through some testing, we’ll release it publicly.
Rich: I want to backtrack for a second, because you mentioned something about how people would be putting up their books and things like that to custom GPTs. Are there limits on the amount of data that it can manage? Because it seems to me like I’ve heard different things.
And in fact, I recently had a book that I was uploading, and I asked it, I’m like, “How many pages in a row do you think you can actually work with before you start forgetting things?” And it , “50”. And I said, okay, let’s do it in groups of 50, summarize each one, and then set those aside, and then we’ll work out the whole book together.
But so what are some of the limitations? As we’re putting up more and more documents to these custom GPTs, when should we know that enough is enough, and how do we know if we’ve gone over?
Brian: Yeah. So within the actual instructions themselves, you’re only allowed to put in 8,000 characters. It’s an 8,000 character limit in custom GPTs, but you can get around that by creating like a PDF document or a word document with all of your instructions in it, and then upload that as a resource. And then within your instructions, just say, “Use this document as your instructions.” You’re limited within each custom GPT to 20 files to be uploaded, 20 different resource files.
Rich: So if I want to get more in there than the 8,000 characters, and I do have a book, if I’ve got the book as a PDF, can I just say, “Read this book before you answer every question” or “Just make sure you’re familiar with this book”? Like, how exactly does that work? And is that a legitimate cheat code?
Brian: Yes, absolutely. So I have uploaded copies of my books as PDFs into ChatGPT. There’s a 512-megabyte limit per file. So you have up to 20 files, up to 512 Meg per file. And then pretty much anything over that, it’s just going to tell you you’ve exceeded your limit for file size or for the number of files you’re creating.
Rich: Okay. Getting back to like, you had mentioned a few GPTs and a popular one would be his brand voice. And another one I know would be buyer persona. So if we wanted to create something where it was always going to look at everything we wrote and say, is it, or is it not in your brand voice, how’s to improve it, but also is it written for your ideal customer persona? Would that be two separate GPTs or could you create one GPT that was trained to review all content you were about to send out, and then provide edits based on those two things, buyer persona and brand voice?
Brian: So we would create them as separate GPTs. So we would have one that was our brand voice. And then we actually have multiple different GPTs for each different buyer persona that we have, whether it’s potential undergraduate student, international student, potential faculty, all are different custom GPTs.
But then what you can do within ChatGPT, you can @ sign within just a regular standard prompt. You can use @ sign and it’ll bring up a list of all of your different custom GPTs. So that then you can select, you can ask it some questions about the content you’re working on, you can open up your brand GPT, you can @ sign that and say, “Does this content fit within our brand?” And it’ll give you feedback on that. And then you can @ sign one of your personas and say, is this content relevant to you? How can we change it? How can we adjust it to answer your questions, solve your problems.
So that’s how you can do these virtual focus groups with multiple different personas, or you can bring several different custom GPTs together. So we’ll do that with our SEO GPTs. We’ll have an article that we’re working on, we’ll have it write the Meta description, have it write page title options, have it go through and optimize the piece of content for specific keywords that we’re trying to target. But each one of those is a separate custom GPT that we’ll reference.
Rich: So you’ve created basically an assembly line GPT here. Where basically you might require five people in the real world to do this particular task because they each have specialties. You’ve done the same thing replicating it in AI, and then you’re basically just adding them or tagging them to bring them into the process, to bring them into the assembly line to get that final output that you’re looking for.
Brian: That’s exactly right. And this is really setting people up to have the mindset that we’re going to need when agents really start hitting the scene next year. So agents are able to string multiple tasks together and access different tools, different browser windows, those sorts of things. So like I always see GPTs as the training wheels for agentic AI.
Rich: So as we’re crafting this, are you putting that into a new prompt or are you going into one of your custom GPTs to start this assembly line? Like when you said you were tagging the different ChatGPTs, can that be from a new prompt, or does that have to be in a different GPT?
Brian: It has to be in a new prompt. You can’t actually prompt a custom GPT from within another custom GPT. You have to be in a root level, new ChatGPT prompt, and then you can bring in whatever custom GPTs you want.
Rich: Excellent. Okay. So that’s how we can assemble our assembly line and we can do a whole bunch of things right in a row. And like you said, that’s probably in the future going to be something that’ll be taken care of by agents, but maybe that’s the conversation we’ll have next year.
I also know that one of your favorite GPTs is to repurpose content. So can you kind of just walk us through what that looks like, how you make that all happen, and then what the output looks like?
Brian: Yeah, so as we’re creating content, we’re always looking back on existing content that we have that’s performed well, that’s worked really well for a specific audience. And then we want to make the most use of that because your content is really a product, right? It has value, has long term value.
So just because it worked really well for prospective undergraduate students two years ago doesn’t mean that it’s dead and has no life anymore. So we’ll go back to those high performing pieces of content. We’ll put them into our repurposer GPT and say, what other audiences, how could this content be used for an international graduate student audience? What channels are they most likely to find this content on? And it’ll tell you, it’ll give you information about that.
And then you can say, all right, now, based on all that information that you gave me, take this content and convert this on, website story into a 20-second TikTok script. And then you go to one of your international student interns and have them reformat that into their voice and their language and then deliver that content.
So we’re looking at existing content that we have and looking at how we can repurpose, retarget, and redistribute that on as many different channels as we can for whatever audiences those are going to work for. And then we make sure that we’re tracking the performance of each one of those so we’re getting a lot of value out of the content that we’ve already spent a lot of time working on. It’s already all original, brand-new content. And now we’re putting it out there for different audiences and reusing it.
Rich: Now is this all text based or are we at the point yet where we can upload an audio file from a podcast or video file and have it work on that, or do we always need to move things into a transcript first before we start playing around with it?
Brian: So it depends on the tool. So some tools like Gemini has huge context windows that you could upload entire videos to. ChatGPT that you can upload audio files too, but I still tend to go with the transcript just because they tend to be more accurate if you go to a third-party tool to get the transcript than if you let ChatGPT do it.
I’m sure that’s all going to continually be evolving and adjusting. But most of what we’re doing is taking our text, our website content, and figuring out how to repurpose that. Because that’s where we have the most data around content performance. That’s where we tend to have the most content already existing.
Depending on your brand, depending on your content that you’ve created so far, you may have a ton of YouTube content. Grab those transcripts, figure out which videos are performing the best, grab those transcripts, throw those into ChatGPT. Or take the video, throw them into Gemini and ask it questions about the content that you have there.
Rich: Interesting. Now, if someone’s listening right now and they have not yet started with custom GPTs, what’s one actionable piece of advice that you can give them or one simple custom GPT that they could create to get started?
Brian: I always tell people, I started out creating custom GPTs for my personal use, my personal life. And I’ve got about a dozen that I use every week and they save a ton of time. And just by going through that process, and it’s not something that I feel like I have to share with other people at work or it’s not something that I’m like, I don’t know if this is really going to improve my performance at work. I haven’t tested it out enough. But if I’m just playing with it by myself at home.
My meal creator, I’ve got thousands of uses on it in the year and a half since I created it, and it has saved me a ton of time. So figure out something that you do in your life that’s repetitive or data driven. And then, whether it’s your workouts that you’re doing, if you want to add some variety into those.
I’ve got a custom GPT that looks at the weather for the week and helps me figure out what cardio I want to do. So lots of different options in there. But I’d say start with something personal, start with something easy, experiment with it. Once you create your first few, you’re going to start getting a feel for how easy they are to build and the power of them.
Rich: Awesome. Brian, this has been fantastic. If people want to learn more about you, check out what you’ve put out there, where can we send them online?
Brian: You can go to brianwpiper.com or you can find me on LinkedIn at @BrianWPiper and most social channels.
Rich: Awesome. Brian, thank you so much. Really appreciate talking to you today.
Brian: Thanks for having me on the show, Rich. I love the content you create and all the knowledge and insights you share.
Show Notes:
Brian Piper is an author, award-winning keynote speaker, and digital marketing strategist who has been optimizing content since 1996. Be sure to check out his website to see how he’s helping businesses get discovered, leverage AI, and drive results. He also has a number of excellent books covering content marketing, as well as a podcast. If you connect with him on LinkedIn, be sure to mention you heard him on the Agents of Change 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.