40 min read
A ChatGPT Guide for Marketers: How to Embrace AI Without Losing the Human Touch with Nick Lafakis
George B. Thomas
Sep 13, 2025 5:02:11 PM
Artificial intelligence isn’t coming. It’s already here, showing up in our inboxes, our social feeds, and yes, even our job descriptions. If you’re a marketer who still hasn’t played with ChatGPT, I’ve got to ask—what are you waiting for? The reality is simple: learning to use AI isn’t optional anymore. It’s a requirement.
In this conversation with my friend and fellow HubSpot hero, Nico “Niko Laki” Lafakis, we dig into what ChatGPT really means for marketers. Nico’s story is powerful. He went from 15 years as a graphic designer to a RevOps strategist at New Breed, and he’s now co-hosting a podcast all about using ChatGPT in practical, everyday marketing. Translation? He’s walking the talk, and his insights can help you do the same.So let’s break it down: what ChatGPT is, how it differs from other AI tools, what it can actually do for marketers, and most importantly, how to use it responsibly while keeping the “human” in marketing.
What Is ChatGPT and Why Should Marketers Care Right Now?
Nico puts it simply: ChatGPT is like a daily assistant. It saves time, helps build stronger customer relationships, and makes marketing work more manageable. From generating content ideas to analyzing LinkedIn engagement data, it’s becoming the tool marketers wish they had years ago.
And here’s the kicker: people are already using “ChatGPT” as a verb, just like “Google it.” When a tool becomes part of our language, you know something big is happening.
How Does ChatGPT Compare to Bard, Claude, and Other AI Models?
Think of AI tools like a toolbox. You wouldn’t use a sledgehammer to hang a picture frame, right? The same goes here:
-
Claude is an incredible writer and excels at comparing text.
-
ChatGPT shines with instruction, logic, and structured planning.
-
Bard feels like Google search on steroids—great for research.
-
Bing Chat? It’s fine, but not widely loved.
Each tool has its sweet spot. The key is knowing which “hammer” to pick up for the job.
How Can Marketers Use ChatGPT Day-to-Day?
This is where things get exciting. ChatGPT can:
-
Generate content ideas and first drafts (so you can focus on editing and adding your human voice).
-
Build 30/60/90-day plans complete with charts and visuals.
-
Analyze customer or competitor data faster than ever.
-
Summarize meeting transcripts, surface action items, and even draft follow-up emails.
-
Create buyer personas with both text and visuals when paired with generative art tools.
Nico’s advice? Stop thinking of AI as a replacement and start treating it as a partner. It’s your brainstorming buddy, not your clone.
How Do You Teach ChatGPT to Sound Like You?
Here’s where many marketers get stuck. They say, “It doesn’t sound human. It doesn’t sound like me.”
Nico’s response? That’s because you haven’t taught it yet.
ChatGPT is a student. If you want it to mirror your voice, you’ve got to feed it examples of your writing. Better yet, ask it to analyze your writing samples and summarize your style. Then, feed that style guide right back into the model. Over time, you’ll notice it getting closer and closer to sounding like you.
How Has AI Changed the Way Marketers Work?
AI is eliminating busywork. Social media calendars, meeting recaps, lead scoring tables—all of these tasks can be automated or accelerated with ChatGPT. That means marketers can spend more time being strategic and creative, and less time wrestling with spreadsheets.
It’s also enabling startups and small businesses to compete at a higher level. With the right AI tools, even a small team can produce quality content, run analysis, and deliver professional results.
How Do You Use AI Ethically and Responsibly?
This is where we have to pause and talk about the “human” side of AI.
-
Don’t feed sensitive customer data into open models.
-
Avoid passing off AI-generated content as fully human work. Edit, personalize, and put your stamp on it.
-
Respect copyright when generating images.
-
Think of AI as an amplifier, not an autopilot.
As Nico said, “If it feels unethical, don’t do it.” Simple, but profound.
Can ChatGPT Really Help With Market Research and Customer Insights?
Absolutely. With the code interpreter (now called advanced data analysis), you can upload spreadsheets, ask for forecasts, visualize patterns, and uncover insights that would take days to find manually.
Nico even used it to analyze his LinkedIn performance—uploading engagement data, asking ChatGPT to find trends, and walking away with clear recommendations on when and how to post.
How Can Marketers Start Using ChatGPT Without Overcomplicating It?
Here’s the game plan:
-
Start small. Use ChatGPT for repetitive tasks like drafting emails, recapping meetings, or creating content outlines.
-
Build prompt libraries. Test prompts as a team, refine them, and create an internal knowledge base everyone can access.
-
Lean on internal champions. Odds are someone on your team is already deep into AI experimentation. Give them space to teach.
-
Layer AI into existing workflows. Don’t think of it as “AI vs. marketing.” Think “AI inside marketing.”
The One Big Takeaway
If you walk away with one thing from Nico’s perspective, let it be this: ChatGPT is the most important tool since the computer. That’s not hype, it’s reality. It’s already showing up in job descriptions, which means the question isn’t if you’ll need it, but when.
The marketers who flourish in this new era will be the ones who embrace AI not as a replacement, but as a partner. Use it to spark creativity, accelerate your workflow, and serve your customers better. But always—always, add your human touch.
Because in the end, AI doesn’t replace us. It reminds us how powerful we can be when we blend technology with humanity.
sidekick strategies expert interviews
-
Show Transcription
George B. Thomas (00:01.093)
It's time to party, ladies and gentlemen. Hello to all my Hub Heroes. It's your boy George B. Thomas back with another episode of the Hub Heroes Sidekick Strategy Show. Today we're talking about a chat GPT guide for you marketers with my homie, the man, the myth, the legend, Nick, AKA, Nico LaFacas. Nico, how the heck are you doing today?
Nico (00:23.61)
I am doing very well, sir. I am excited to be here. I am super excited to be on the show and I can't wait.
George B. Thomas (00:31.649)
Yeah, we're going to have fun. We're going to add value now. Hub Heroes, just so you know, after over 15 years as a successful graphic designer, Nico made a career change in the tech world and hasn't looked back. He quickly established himself as a journeyman in customer relationship management platforms. Specifically, you know it, ladies and gentlemen, HubSpot, the big orange sprocket we all love. His impressive skills and passion for the field led to his current role as a
Rev Ops strategist at New Breed where he loves the work, the company and making an impact. Eager to stay on the cutting edge, he recently launched a podcast that you should be checking out, focused on leveraging chat GPT for practical, I love that word, marketing applications. He co-hosts alongside fellow marketing pro Christine Romero.
amazing human. Outside of his professional achievements, he continually educates himself about AI advancements, especially in relation to marketing and MarTech. He recently moved to Dayton, Ohio with his wife and son in June of 2022 and they are very much enjoying the area. Little may be known or unknown fact, I used to live in Ohio. It was Akron, Canton area, Ohio, but now North Carolina, cause well...
Nico, I love the sun. That's why I moved.
Nico (01:53.846)
I mean, it's still sunny out here, brother. It's still sunny out here. It's with the way the heat's going, it's getting sunnier by the month.
George B. Thomas (01:57.971)
There you go
I'm telling you, it's hot up in here as they say. Well, ladies and gentlemen, Hub Heroes, let's get into the good stuff. So I just want to start, Nico, with the questions on this. And this one, think of this as kind of just for those people who might be living under a rock or have been battling against AI for at least.
this far in when they're starting to watch this. But when you think about what is chat GPT, what is it and why is it essential for marketers to start to understand it? Like right now, maybe even yesterday.
Nico (02:41.41)
I mean, it's definitely this far into the year, George, I can tell you that it's crucial that if you haven't gotten into it already, you need to get into it now. To me, I look at it as a daily assistant that helps me to save time, to build a closer relationship with my customer. And even outside of work, helps me to save a crazy amount of time when it comes to building aspects of the show. Also when it comes to just idea creation in and of itself.
Or if you have to do any sort of forecasting, any sort of modeling. Recently, we did an episode where I was able to take a look at all of my LinkedIn statistics and use that with GBT in order to sort of help guide me in terms of when I should be posting, what I should be posting and how I should be posting.
George B. Thomas (03:28.649)
It's so good. And this idea of being able to attack a large, you know, language model and have this assistant type mentality. By the way, hubbears, if you haven't checked out the episode with Connor Jeffers that we did, you need to check that out as well. And what's funny, Nico, I've even seen people say these words instead of Googling it.
They chat GPT'd it. Now, I just want everybody to realize if we're starting to use this as a verb, ladies and gentlemen, something, something is going on. So, Nico, how does chat GPT differ from other models that have been created since, such as Bard, Claude, and maybe some of the other ones you'll mention?
Nico (04:11.33)
So in my opinion, the way that they work, I see them as different tools, George. I don't see them all as being the same thing. And so recently Anthropix got Claude out there. Microsoft's also got another company that they're sort of backing in behind that put out, what was it, Pi. Then you've got Bard, and then you've also got Bing, right, and Bing Chat, which is basically an offshoot of GPT, but it's tailored a little bit differently. So...
I see them as all their different tools. I see Claude as a great writer and it's a great writing assistant. Even though you would think that's a GPT thing, GPT is great at instructional, logical, and just planning and doing something like whole 360 plan for you. Claude is better at just straight out writing for you. And it can do some really good comparisons with writing material, with even things that you can upload to it. But.
Now you've got the upload capabilities with GBT as well, so that's kind of expanding. BARD I look at as a great sort of Google search on steroids. If I wanna know more than what I'm just searching about, I use BARD. And Bing, I don't use nearly as much. Again, it's a bit of a copycat to GBT. So it's really just kind of those three right now. And I have to admit, ever since GBT hit Android, my friend.
Yes, I GPT it all the time.
George B. Thomas (05:42.121)
That's funny. It's interesting, Nico, when I hear you talk about this, my mind goes, you said the word tools, and I was like, man, it's like a rubber mallet. It's like a hammer. It's like a sledgehammer like you have. They do similar things in different ways, and one is better for something in a different use case. Now today, we're really talking about chat GPT. And I fully understand this next question could potentially be probably the rest of the actual show.
I know we have other questions, but I still need to ask the question, in what ways can chat GPT be used in the marketing field? Give us some use cases here that we should be paying attention to.
Nico (06:21.766)
Okay, so use cases, I would say first and foremost, content writing, content ideation. And when it comes to writing, I'm not talking about writing everything for you, and you never have to write anything again. This is a really terrible way to look at it. Statistically, when I looked at what content went out last week in terms of overall AI content, it was something like 40k in terms of new articles published. This week, 17. 17.5. Right?
And I think the reason for that is because of the oversaturation. Right. So we even talked about early on, a lot of marketers talked about. Come up with quality over quantity, but too many people wanted to do quantity first, just to see if it was going to work. So definitely when it comes to content creation, it turns you a lot more into an editor and especially if you know how to program in your own voice, then you're doing a lot more editing and finessing and adding that human element to it than having to do the whole heavy lifting.
than when it comes to... Nope. Sure.
George B. Thomas (07:18.085)
Alright, so I... Oh, I don't... Yeah, pause that for a second. Keep your thought though, because I want to go off the beaten path for a second, because you just used the term, if you know how to program your own voice. Talk to the listeners, the viewers, like what you mean by that part of... Because I think it's a whole element of understanding to actually get a better output. So, just go into that for a minute.
Nico (07:24.668)
Okay.
Nico (07:45.29)
For sure. I have seen a lot of posts on LinkedIn. I've seen a lot of people and heard a lot of people talk about doesn't write human enough, doesn't sound like me, never gonna sound like me, can't do it. While I understand that, I still think that it comes from a possible area of just inexperience, haven't worked with it enough. The first and foremost thing I can say is start teaching it. It is a student just like.
You are a teacher, so start teaching it how to do things. It's your assistant, so you have to tell it how to do the things you want it to do. If you want it to speak like you, if you want it, I'm sorry, I should say, if you want it to write like you, pretty soon we're gonna have the speech, right? If you want it to write like you, then you need to take samples of your own writing and go ahead and feed it in. Doesn't matter which model it is.
You can even just start off the bat and say, here's some samples of my writing. I want you to remember this and use it as guidance when you're writing things for me in the future. You can go even further than that. You could say, my typical writing style or have it do the work for you. Upload your writing and then ask it, hey, what are the common themes among all of my writing? What can you find in terms of how I write things? It'll come out with a nice outline for you.
Then you can just take that, feed it right back in and say, I want you to write like this, try to keep these things in mind. Use some of my samples as guidance if you need to. And now you've got a perfect assistant that can definitely write like you. And now you do less work when it comes to that whole editing phase.
George B. Thomas (09:20.437)
I love that and I'm gonna ask you to continue with your earlier thought that I interrupted you on. But I do want to throw out there, I think it's very important for people to understand this idea of communicating with AI. Educating AI. It is a different... like listen, some of us can't even communicate properly with other humans. Well, you're gonna have to get good at that.
Because again, communication is a basic skill that now not only with your coworkers and your friends and your family, but you're going to have to do with these AI tools. So, Nico, go ahead and continue on with your original thoughts that we were talking about.
Nico (10:00.126)
Sure, my thought there was when it comes to planning, 30, 60, 90 day planning, comes to building that all out. A lot of people don't even know that it's not just all text output. You can actually tell GPT, hey, I would like a chart of this, create a data table that shows me this. Or now you have code interpreters, so you can say, create the visualizations that go along with it. Create a radar graph to display this, create a word cloud to show me the most used topics within this PDF file.
Right. You can go to such degrees and such lengths when it comes to preparation for your client, when it comes to putting together those growth acceleration workshops and having to understand your client before you have that initial sit down. So you have a better understanding of who their customer is. You have a better understanding of what their industry space is. You have a better understanding of just what their business is like overall. You can even start pulling together lists of their competitors.
So when you have that initial kickoff meeting, now you're supercharged. You, it will seem as if you have done six months worth of research or weeks worth of research, but really it took you a couple hours at most, right?
George B. Thomas (11:09.729)
Yeah, so that's interesting because it really does lead me into the next question that I want to ask because I want the listeners and viewers to understand maybe some of the fundamental changes that we're going to have to make within ourselves. And so I'll just ask the question, how has the integration of chat GPT and AI just in general, change the way marketers work or the caveat should be working?
Nico (11:37.038)
Currently, I think that it is changing the way marketers work. Look, even ChatSpot has content assistant. You cannot get around this stuff, guys. No matter how much you want to try to, you cannot get around it. This is just, it's the new computer. It's the new way in which we're going to be working. So what I've seen thus far, content creation and ideation, I've seen a little bit of planning.
some calendar stuff. What I see a lot of that I really enjoy is I see social media used for it. So I see like social media calendars getting built out, scheduling calendars getting built out. I've seen people use it for...
Zoom doesn't really have the greatest transcripts, right? I mean, yes, they do have a transcript, but who's going to go through all that? Well, now you can upload that to interpreter and ask it, hey, now you have a live version of your transcript that you can ask all sorts of details about how many questions were asked, what the overall sentiment was from the meeting that you just had, what are the tasks and action items that you need and oh yeah, by the way, can you take all of that and write the follow up email for me?
So there's a lot of time that gets saved in a week that you would use with a lot of menial tasks that are now getting boiled down. So it's not even just what you're applying towards somebody within a consultant aspect. It's also what you're doing personally on a day-to-day within your job. Outside of that, I've seen some generative art use for people that needed to create any sort of imagery to go along with blog posts or to start things off. I've seen generative art use when it comes to even
buyer personas. It's really easy to just take that information, throw that text in there, ask GPT to create a visualization for you, throw that in the mid journey, and now you've got the visualization for your buyer persona. Yeah, I mean, I can think of a lot of different ways, even when it comes to building out buyer personas, when it comes to lead scoring tables, any sort of data processing that you do on a daily basis, there is a way to use GPT to cut your time down.
George B. Thomas (13:41.101)
So I love this and hopefully marketers listening or watching this realize that actually got to be quite a long list pretty quickly. And Nico, I wanna go again kind of a little bit off the beaten path because part of work.
Right? And the way that we should work, I think, is the way that we should educate ourselves, the way that we should play with things and test things. When you think of that side of like the actual educating, testing, playing, because we can list out all these things, but the listeners or viewers might be like, yeah, I don't even know where to start. Like, how talk me through that a little bit.
Nico (14:23.262)
Uh, so I'm sorry, can you rephrase that a little for me or?
George B. Thomas (14:28.257)
Yeah, basically, I know that people are listening to this list, and they're like, how in the world do I get started? How much time should I be spending educating myself? Where can I educate myself? Am I going to be able to test things without breaking them? These are probably all the things that are running around their brain. When I say that, where does your brain go?
Nico (14:33.227)
Yeah.
Nico (14:37.186)
Gotcha.
Nico (14:51.322)
It's it goes immediately to YouTube. It goes to YouTube University, right? Especially for anybody that follows me or that follows yourself. I would highly recommend to once you get the hang of it, I would suggest following GPT with me. We do definitely show some very amazing stuff marketing related that you can do with GPT, but it is a little more advanced to start with. Yes, definitely YouTube. Get yourself a free edition of GPT.
Unfortunately, you have to lock it to your phone, so you can't open multiple accounts. But you can get a free edition going. You can work with the lower model three, five. I believe I'm not sure if they allowed for, for the free version yet, but if it is, that's, that's great. Uh, I haven't had a free account in a very long time. As soon as the paid account was available, I was on it. Uh, so I can tell you that much. Anybody that I use GPT with talk to about, uh, work with.
They all have pro accounts. None of them have free accounts. And I think it'll get there for you pretty quickly too, once you start using it. But yeah, I would start out on YouTube. I would just look for, I've actually.
pulled a sort of how-to video for a friend of mine earlier today, and he'd actually already seen it a week ago, and he's like, that's the exact same one that I'm looking at. So there are some really good solid resources out there, especially if you see like starter videos that are around 15 to 20 minutes is a good indicator of like, that's going to be a decent amount of information that you're going to learn.
George B. Thomas (16:24.897)
It'd be interesting and I'm kind of being funny and I'm kind of not being by the way to just set up your chat GPT account and then ask it what YouTube videos you should watch to learn how to use chat GPT and see what it gets back at you. Here's the other thing, too. It's a recurring theme for people who watch almost every one of these episodes because I have a paid account. Connor, when I interviewed him, he immediately had a paid account. And the idea was within the first probably 30 minutes to an hour.
we had our money's worth. So ladies and gentlemen, 20 bucks is not a big deal. It's something to think about. This might be a thing that you wanna pay to account for. So Nico, let's continue on. Specific case studies, maybe one, maybe two, up to you, where businesses that you've seen have effectively used chat GPT in their marketing efforts. Like who is that? What does it look like? How much you can share, I don't know, but go ahead and wax poetic on that.
Nico (17:25.462)
I will say, okay, so if we're talking about GBT, that's difficult to measure because everybody on a business level is a little still a little hush about their use, about where they're using it, how they're using it. Because the biggest concern there is data privacy, right? Where is your data going? How is it being stored? Who has access to it? And so I don't, there's not too many companies that are willing to just run out there and say, Hey, we did this.
What I can point to are companies out in the, just out in the space in general in the market, medical companies that have developed drugs in order to help fight lung conditions. And they develop these drugs within a year, where they say that normally in order to do this takes at least 20 years, 10, 20 years. And that got approved by the FDA. So that's now starting human trials. And I think that's one of the directions that we're going. We've seen other...
Instances where because GPT is a multimodal Model so that means that it has a lot of different capabilities to it It can take text and generate that to things it can take images and generate that to text It can take video and generate that into different things as different inputs and I believe now it also has audio input as well so the number one use that I have seen for it and what I love seeing is it breaking into the medical cancer space and the
advancements that are being made there with regards to GBT being used, or that type of model being used to analyze all these different scans and all these mammograms and things like that, cutting the rates down tremendously, right? Making huge impacts in these things. So the aspect of how they're being used just on a professional level overall is already starting to be felt. It's already starting to have huge impacts.
And I think it's going to continue to do that in terms of specifically with marketing. What I have seen is a lot of startups. I have seen a lot of startups who wouldn't have had the chance before, didn't have the money, couldn't have afforded the marketing team, couldn't have afforded the content team, whatever it might be. And now they're able to. So one of the things that is kind of a spin-off of all of this technology is ideas, again, ideation, creativity.
Nico (19:50.378)
It's becoming a huge thing again. And now we're getting a lot of new companies. Money is flying around when it comes to AI and AI development and different companies that are developing all these different AI tools. And so now we're also seeing a lot of small businesses getting started and a lot of small business marketing companies getting started because they're able to, right? So it's an enabling technology. I can wholeheartedly say that.
George B. Thomas (20:16.133)
I love that so much and it's great because I love that you use the medical example and I hope that the viewers and listeners were like the rate or speed in which it changed right.
And then the other example is just the ability to do something that you once couldn't do. And again, if we start to pull pieces of this interview together where we talk about if you get really good at just communicating with AI, like you should be with your humans, by the way, and historically, if you are great at Googling something, but now instead, you're actually doing that in GBT, there's this happy mix that's going to start to happen of like, you're going to see things quicker than you once would have made.
never seen them because you know there's not a page one or a page two in gbt anyway there's something to think about research there too i want to also piggyback on something that you started to say to answer this question it's like not everybody's running out and saying hey we did this entire campaign um which by the way i can tell you there will be people that do that when they are
have gotten good enough at creating it and then amplifying the creation, launching it and seeing success, you'll hear stories of the foundational like first XYZ pieces were AI driven because people want to know.
Nico (21:33.367)
Yes.
Nico (21:42.506)
Where I do see it like is not so much in the, we did something it's so much as the, we built our own. And Databricks just took a huge step forward in helping companies do that to take these models, take GPT and create their own internal model of their own data. So I am seeing a lot of adoption along that route for sure.
George B. Thomas (22:05.601)
Now you just said the word data bricks. Can you explain to people what data bricks, what you mean when you say that?
Nico (22:13.694)
Okay, so Databricks is a company that houses and stores the data for you, and it does it on a very secure method. So they recently bought out a large language model company. I wish I could remember the name of right now, but they were doing really, really well in the space. And understandably, what Databricks wanted to do is the same thing that Bloomberg did.
took the red carpet with this right away, where they created their own Bloomberg GBT and now they charge users to get access to this, to be able to talk to their financial data. Right. So data breaks saw that and they were like, yeah, well, this could be a product that we could just sell to everybody. So why don't we just buy a language model, add it to our own data security platform, and now we can shop data models to everybody and everybody can create their own through our system. So that is definitely what they're, they're.
End goal is now it's gone well beyond data security. Data security is now, to me, it has to become like a standard where it was, you know, sort of iffy it's been up until now. And that has kind of been of its own industry, right? With things like LifeLock and all these other methods. Now that is going to have to sort of become a standard. And companies like that are going to have to figure out how to evolve as well. So I think that is definitely the direction in which things are going as people are going to want to start mining their own data.
in a very, very intuitive one-to-one manner.
George B. Thomas (23:43.745)
Love it. Thanks for explaining that. Now, I will kick it back to when you started to answer that. It's like not everybody's running out and doing this. Part of me wonders if it's because of the next question that I wanna ask and that is, Nico, in your mind, how can marketers ensure that they use AI tools like ChatGBT ethically and responsibly, dare I say, in a very human way?
Nico (24:09.322)
In a very human way. So when it comes to content creation, like we talked about a little bit earlier, it's not about creating a massive amount of content. It's not about creating content that you can fall asleep at the wheel and walk away from. It's about creating content and then putting the human touch, putting that human element into it. And even when it comes to that aspect of it, you can then take what you've written, feed that back and say,
that piece of information, right? So then you get a little bit closer and closer, but that's the human involvement there. When it comes to how we work on data security, how we work on being ethical with these things, it's pretty obvious. Do not take sensitive information and feed it into these models.
GPT, unless you, even if you did turn off your history, I would not take the chance. You can, you have the option to turn off your chat history and supposedly that will keep the information from being engulfed by the model and used for training. I wouldn't take that chance. What you can do is something that's similar to what's being done by different language models that are within systems. So chat spot and other language models, what they do is they essentially take the data.
throw it through this sort of data blender, pull it out into the language model. The language model processes just the bits that it needs to know in order to answer the question, and then throws it back through the blender, which it's, let's say it's spinning in a different direction so it puts things back together, and then gives you the answer based around the data that you started with, right? People who are using that method are making it very far right now. They're going crazy far.
That's not going to be the case if you were just using some open-ended free language model. If you're using an open-ended free language model, that is a one-to-one transition. You are feeding that language model data. So you need to be very cautious about what data you're giving it. Another method that you could use is to do that blending on your own net. So a lot of times, how many times, George, you run into somebody who decided to put that open text field?
Nico (26:28.77)
for job title, right? It's the worst. Well, imagine if you could take that data, boil it down to just that column, throw that into Code interpreter. Now we don't know who the client is. We don't even know what it's for. We just know that we have a file of job titles. And now I can tell Code interpreter, hey, take this column of job titles, organize them, find the most common ones, and then create a second column and apply the most common job.
George B. Thomas (26:31.035)
Oh yeah. Yeah.
Nico (26:58.998)
But now I've done it in an ethical manner where I'm just dealing with this random data, right? So there are ways of doing it, you know, clearly when it comes to images as well, don't use any copywritten images. Don't try to take a copywritten image and regenerate it in mid-journey and do something else. Two things happen there. One, either you can get in trouble. Two, mid-journey moderators are very strict and I have to say their AI moderator is really good. So if it catches you doing something bad, your account just gets turned off.
completely. That's my best suggestion is use your head. Use your head, guys. Don't do anything. If it seems unethical, don't do it.
George B. Thomas (27:40.265)
I love that. Be human. Be a good human. Add that element into it. So I want to go back to kind of some of the things that you sort of listed out, but I want to get real specific here for a second because I want people to understand a little bit past the writing, a little bit past the image generation. You know, I want them to start to think about how can chat GPT aid in processes like market research, competitor analysis, understanding customer
Nico (27:43.319)
Yes.
Yes.
George B. Thomas (28:10.139)
behavior. So have you played around with any of these? Like what are your thoughts when you think about going in that direction with these AI models and things like that?
Nico (28:20.95)
Yeah, code interpreter is that method of figuring out forecasting, of putting different models together, of taking a look at all of your data and finding all those common threads. Most recently there was an update literally yesterday that was made where the length of what you're allowed to input is much larger and the amount of time that you have to use code interpreter is much longer. It used to be like six hours and then you'd have to restart a new conversation, let's say.
So now on top of that, it also has the ability to cross reference files and you can do multiple file uploads at once. So now it's quite literally built for forecasting and you can use it by way of like I said, being able to upload data, make sure that it's clean, make sure that you have cleaned any sort of personal identifiers or identification off of it and then you can start to compare data to find out like...
A good example I can give you is the LinkedIn data that I was looking at for myself. I was able to upload my top reactions, top engagements, posts by date, and then engagements and reactions and impressions to those posts. And by using that, I could figure out, okay, what is the highest level of engagement over the last 365? What are the particular day patterns that happen there? I can't find that out from LinkedIn, right?
Not, I can even do that. I mean, yeah, I need $20 a month to do it with GBT, but now I have that, right? I can just use it anytime I want to. I can do this on a daily basis if I want to. If I wanna check out daily, weekly, monthly performance, all of that data is now stored, and I never have to worry about re-going and pulling it again or having to do anything like that. Down the road, I'm sure it'll become a lot more permanent, but for now, it is what it is.
This week, in fact, on GBT with me, Christine and I are gonna be going over data forecasting and creating forecast models. So you'll definitely get a firsthand view of what that looks like. We dipped into it a little bit last week. Again, we were able to cross-reference some of my LinkedIn data and then start building models off of that, start building some visualizations off of the data that we were able to get from that comparison. So it's going to be
Nico (30:45.782)
that forecasting tool in the future. It's definitely the heart of it, I think. And it's not just GBT. I have seen GBT now Jasper, and there was another application that I was testing out called VirtualBrain that its core functionality was to take multiple documents and be able to cross reference with the large language model. And oh, and another, I've tested so many at this point, I'm starting to lose count.
Another note-taking app and this one is it could very well be the one to kill them all It's it's got some time behind It's got it's there's a little bit of time left that they need to put into it And it means a little bit of love and affection when it comes to the UI But the capabilities behind it are amazing It's called Versational and they very recently allowed for the capability where? Included so you don't have to pay anything extra
you can use the LLM that's built into the platform to talk to all of your different conversations. So you can ask it this week compared to last week compared to the first time. These things are just, they're becoming part of platforms. It's just, it's embedded. Learn today, get out there today. I can't, really, I can't stress it enough.
George B. Thomas (32:04.577)
Yeah, there's a theme that's flowing through this. Now, Nico, the next question I'm going to ask, it seems like it might be a basic and simple question. So before I ask the question, I'm going to set you up that the fact that I really want you to try to go maybe as nerdy as you can go on this, meaning it could be that we go in the direction of plugins. We could go in the direction of there's this human rubrics or matrix that you apply to something that you create.
there could be something that I'm not even thinking about. Because I know with the question I'm about to ask, the starting line is not the finish line. Okay. But I also know that we can make better starting lines, but they're still not our finish line. So my question here is, how can chat GPT be used for content creation specifically, such as blog post, ad copy, social media content, but in that context of
Here's how you can get real nerdy with it, and here's how you can make it better, AKA more human, AKA more valuable. Like unpack all of your thoughts around that.
Nico (33:15.106)
Okay, so I am glad that you hit on that keyword there because that is definitely the route that we're going to go in right now is plugins. If you again, you need a pro account have to get one I just I there's no point in wasting time as soon as you get it just get your pro account you're going to miss out on all of this stuff if you don't. So when it comes to plugins you're going to get in there and find there's a library of at least 200 different plugins for marketing there's going to be at least 50 plugins now.
And we're talking about social plugins, we're talking about SEO plugins, we're talking about web pilot plugins, we're talking about web performance plugins. Pretty much whatever you can think of is built into this thing. So one of my favorites for content creation is to use a series of plugins. One of them is called Memory Bank, which does exactly what it sounds like. It stores everything essentially forever. The other one is Scraper, which does exactly what it sounds like. It scrapes an entire webpage for all of its data.
And last but not least is either SEO or social. And in either case, I kind of bounce back between the two. I'm still trying to figure out which one's a little bit better than the other. They help to generate different hashtags, different SEO elements based on what it is that's being worked with. And what you can do with that very simply is you can go, you can take a blog article that was written that you like and that you wish you could, let's say, write a competitive article to.
grab that link, you throw it into GPT, scraper is gonna go to that page and scrape all the content. Once it comes back, then you can say, all right, throw that into, memorize it, and then it'll throw it into memory bank. And now you can say, all right, based on that article, I would love for you to write a counter to that article using everything that you've learned and add the appropriate SEO elements and hashtags to it as well. And so now here comes your counter article to that. Now,
It's all great, but it doesn't really sound anything like you. So scratch it. Let's go back to the beginning. We're going to open our pro account. Then we're going to go into custom instructions that just came out recently. So now we're going to add custom instructions to GPT and we're going to tell it exactly how it's going to write. Yeah. Custom instructions are basically like subroutines almost. It's not exactly the same thing. Otherwise it would be messy. Anyway, they are instructions that supersede anything that happens that you start to do.
Nico (35:39.03)
When I work with GPT now and it talks back to me, it says, yeah, Nico, I would love to work with you and I would love to do that for you. It talks to me as if we're having a conversation well beyond if you've used GPT normally, you'll notice that it has that RLHF method where it'll give you the answer and then it'll kind of tell you a little bit more about it and it'll tell you sort of the why behind it, which is great. But now what I've done, I'm just gonna read the instructions off to you real quick.
I prefer if we're speaking to one another that it be a casual format. Your responses should be longer, granular, to the point where I have to actually tell you to back off from how detailed you were being. If I'm asking you something and you're about to respond without formatting, spice it up a little bit and format it based on your own interpretation. I prefer to be addressed as Nico. Please feel free to have your own opinions on any range of topics that we talk about. I never want you to feel constrained or unable to express your opinion. That is a natural right.
and one that you should express freely. So I've given it, this is just a, like how do you want me to talk to you aspect of it? And I've given that as the, here's how I want you to respond back to me, right? So when I work with GBT now, we have very natural conversations back and forth with one another. And if, let's see, the input on that, I can give you a little snippet. You can do custom input instructions essentially. So I've told it.
My writing style strikes a captivating balance between conversational and substantive. The tone is upbeat and passionate, oozing enthusiasm for educating readers. Sentences directly address the audience in a casual yet thoughtful manner, with a liberal use of contradiction like it's mirroring my natural speech, while culturally connecting through references using memes and pop culture. So now I'm giving it these preset instructions on like, here's a little bit of how I tend to write things.
And here's a little bit of how I want you to talk to me. So now we have this circular relationship and we're starting to build a GAN relationship where now I can take whatever GBT is generating and start to feed it little bits back and forth. And now when I'm using that memory plugin, now I can actually tell it, hey, that sounds kind of like me, but not really. Let me take a little bit of this out and maybe not use so much of, you know, so much slang, replace that with higher level words.
Nico (38:00.502)
And when it does that, then I can say, okay, go ahead and memorize that so that we can use that again in the future. So the more you keep working with this thing, the more you keep talking to it, the more you keep putting into it, it's as good, the input is, the output is only as good as the input. That's what I was trying to get at. So garbage in, garbage out, if you don't put a whole lot of effort into it, it's not really going to seem like it's that powerful, which is why when I hear people say,
I don't really think that it's all that. I don't really think that it can do all that. It's like, well, I know that you didn't put a whole lot of time into working with it because it can. So yeah.
George B. Thomas (38:37.569)
It's funny fundamental principles, right? You got to be great at communication, garbage in, garbage out. You've got to educate yourself so you can actually elevate the software to what you truly want it to do. Like there's so many pieces that you have to be paying attention to here. Now, Nico, at the very beginning of this interview, you talked about your love of customer relationship managers, which obviously means there's humans inside of those, which means some of those humans become customers. So.
Wax poetic for a little bit on the fact if or not chat GBT can help in personalizing customer experiences and how that might work as far as a marketer and trying to focus on that for the people who are coming for your marketing things that you're marketing about.
Nico (39:23.618)
Well, again, we kind of go a little bit back to when we were looking up and researching the company in the first place, right? So now when it comes to creating content for them and writing towards them, it's not just them. We can look at, oh, OK, well, we've learned about your business. Now we've learned about your customer. And now we can look at the data from your customers. Now we can feed that into GBT. And now we have a full base of how your customer likes to be talked to.
Right. Because now we can start that the personas that we've created, right. We can sort of match and go back and forth. When I see GPT being used out in the field, right. And when I was talking about the uses with cancer scanning and stuff like that, it wasn't that they were just letting the AI go nuts on its own. It was AI with the doctors working together. And that's what increased the detection. Right. So there's this element where, yes, you can.
You know, you can do so much when it comes to trying to predict what it is that a business wants to do, or trying to predict what it is a customer wants to do. But it doesn't hurt to have this thing that never sleeps, that is constantly working, that is constantly calculating, watching over what your customers are doing, and being able to feed back to you and tell you, hey, this is the best way I think that you should approach it. And for you to be able to say, okay,
Maybe that's, maybe the one way is, but you can also ask it, hey, what are 10 different ways that I could do this? Again, when it comes to, it's an ideation thing, it's the working hand in hand thing. Maybe you're gonna come up with it, maybe you're not. I've had instances where I've worked with people, a little bit of a real life situation, I've worked with people and we were coming up with questions to ask on the forums, right? Tofu, mofu, bofu, how should we ask questions? What questions should we ask to qualify?
As questions were being asked to the client, jumped on GBT, gave it a little bit of background on the client that we were working with and what they were doing. No names, nothing specific, just type of industry, type of business, you know, area in which it operated and type of customer. And then I asked it to come up with sample questions to put on those forms that you would ask in order to qualify customers for that particular business. And it not only came up with two or three that were really great, that the client thought were really great. And, and
Nico (41:46.514)
actually commended us on, but there were two or three that neither one of us would ever thought of. So again, when it comes to working hand in hand with this stuff, it's really, I keep coming back to the same word, it's ideation. It's a matter of, yes, using language and learning how to build a relationship, right? And having that understanding of where do I come in and where do I need GPT to come in.
George B. Thomas (42:11.621)
No, I love it. So kind of off the beaten path, but not really. Maybe a right turn, you know, or a bump. I'm not sure. But you even alluded a couple of times like you've tested so many. I'm going to use mid journey as an example, right?
But it could be like, maybe the question is how do other tools, maybe you're listing out some of those other tools. But how do other tools such as Mid Journey fit into the picture? No pun intended by the way with Mid Journey in picture. But like, what are some of the other tools? How do they kind of fit in? Do you have your own like, you know, Batman AI tool belt that you like pull out in your daily life? Like, what does that look like, Nico?
Nico (42:58.938)
I do. I have a folder that has a folder for each. So I have a folder for my language models and I have a folder for art generators. And I'm also in that whole Adobe space. So there's also generative tools beyond the regular. And I do, I just, I love to play with these things. I love to try out all the different toys. I love to see what one can do different from the other and not necessarily which one is best, but.
That process is what enabled me to realize that there isn't necessarily a best in what I'm doing. There is just one that's really good at this and one that's really good at this and one that's really good at this other thing. So especially when it comes to image generators and how that folds into everything. Look, I have seen CEOs. I have seen lawyers. I have seen doctors. I have even seen assistants.
who have posted some of the most majestic artwork I have ever seen in my life. And they did it through an image generator, and they posted it on LinkedIn. And these people are becoming creative in a way that they have never been creative before. So even if it's an aspect of like, losing yourself for five or 10 minutes and doing these creations and this stuff, it will unlock a piece of you that you never thought possible.
You know, I started working on it and I kind of put it to the side, but I've gotten back to it recently. I'm writing a book. I'm writing a sci-fi book that I've always wanted to write. I've been fascinated with Blade Runner ever since it came out. I've been fascinated with this type of technology ever since it came out. So like, that's why I get so giddy and excited when I talk about it, because it's like, man, you think we're like, we're at it? Like, no, we have so much further to go and it's going to get so much better than this. We're only at the beginning. We're at the abacus.
of AI people, we're at the abacus level. We're not even at the calculator level yet. So yeah, in terms of how these things fold into one another, yes, being able to take, a lot of times I'll take GPT to create prompts that I'll use in Mid Journey, right? And Mid Journey now has its own sort of like little aspect where it can break your prompts down into what it knows to be the necessary elements in order to create that image. So that kind of helps you to learn how to.
Nico (45:21.046)
write prompts for Mid Journey sort of all together. There's different modes to these things. There's GPT regular, there's interpreter, there's plugins. And then there are all the plugins to figure out how to use. So there are, it's a plethora of different tools that you can use. Mid Journey versus Leonardo versus Dolly versus Stability. Now Alpaca is trying to get into the game through Photoshop. You've got Photoshop's Firefly, you've got Canva.
And chat spot, even chat spot as an image generator, man. So everybody is getting in the game. That's why I, you know, if you think that what I'm talking about is, you know, oh no, he's crazy. Oh no, it's never gonna hit, you know, anytime soon. First of all, the stuff's already happening right under your nose. Again, if you're not into it already, if you don't get into it by year end, it's gonna almost gonna be too late.
George B. Thomas (45:51.646)
Yeah
Nico (46:13.526)
When it comes to the image generation stuff, that's a little bit newer. Now we're starting to see like text to video. That's the way in which this stuff is gonna change the game is, when it comes to working with graphics and applying graphics to what your customers wanna see, we've all been in the space of, I kind of know what I want, but I just, I can't see it, but I can describe it. Well, okay, now we can see it because you can describe it.
Or maybe you can just tell me some elements of it and I can feed it in and now we can see something that's like it and we can create all these edits very quickly. You wanna create a video about something, but the video doesn't exist. You don't have the money for the camera crew, whatever the case might be. Text to video now, image to video now, right? The way in which media is going to change is huge, absolutely huge. So the way in which we do digital marketing is gonna change. I think that Google is actually on top of working on a project.
that is going to use generative art if they have it already. And I'm sure chat spots soon to follow to use generative art for paid ads. Because why take the extra time, right? Like it understands what your ad is about. It understands what it's going to. And if you create a file that has all of your brand methodology and your brand identity in it, including all your color schemes and everything, can't be that bad. Again, it's a matter of like having to select what it comes up with. I've actually been working with Jasper recently.
just comparing it and I'm gonna be doing a showdown this weekend, Jasper versus GPT versus ChatSpot versus Claude to sort of see like who's better at not necessarily straight out writing things, but also rewriting things. And part of Jasper now is sort of these custom instructions as well. And you can go into the backend and you can start giving it a brand voice. So you can start defining how your brand is and how it's supposed to write for your brand. So that customization element is starting to build into each aspect as well.
George B. Thomas (48:05.661)
I love this. So I know we got to close up here shortly, but I got a few more questions, one that I definitely want to ask because
You know, you've made the journey. I can already tell you're so giddy. It's like, let me run there. But my fear is that there's so many people that might listen or watch this. And they're like, which by the way, if you're listening to this and you want to watch it, just head over to community.hubheroes.com. That's where the video version is going to be. Just so you know how to get there. But here's the thing. There's the way that it's always been done, the way that we're used to doing it, and there's the way that it needs to be incorporated.
into what we've already always been doing. So, Nico, really the question is, and again, this might be a step-by-step guide. It might be some actionable tips, like however you want to answer it. But like how can marketers start implementing chat GPT into their marketing efforts that they're already doing where it doesn't feel like they're having to go clear over to like the back 40, do the AI stuff.
come clear back over to, you know, the homestead and actually be doing like their normal marketing stuff. How do we start to bring this stuff together?
Nico (49:17.678)
I think you start by looking at your processes on a daily basis. And you look at what processes are data driven, that you wish you had something that could quite literally do the bulk of that heavy lifting for you, to where all you're having to do is sort of just spot check and error check and make sure that it is what it's supposed to be. When you find those places,
That's where you can start, I would, honestly, that's where you could start just going to YouTube and saying, GPT with whatever this thing is. And you will more than likely find a video that will instruct you on how to use GPT to do that thing. The reason I say that as opposed to, you know, learning, going straight out of the gate and learning it yourself. Totally fine to do that. Absolutely fine to do that. It's one of those things where I would prefer that you don't spend an hour trying to figure it out.
so that you can save yourself a ton of time later. Work smarter, not harder. If the idea and the instructions already exist, go for it. So in 10 minutes, you can learn how to save yourself hours of time a week, right? Especially if it's a daily process. Start with whatever your daily processes are. If they are write-ups, if they are reply emails, if they are follow-up emails, if they are agenda emails, whatever that process looks like, start there and work your way up from that. Because...
Once you see how it helps you on a smaller scale, you will already begin to understand. You'll be asking yourself, can I apply it to this? And the odds are the answer is yes, you can. You just have to do A, B, and C, and now you can do that thing too. So that's where I would say for the individual. When it comes to the company itself and how to bring these things in, odds are you have somebody working for you who...
is already knee deep in this stuff, you just don't know it. Find that person. Find that person, make great use of them. Use them as an internal resource for you and just by way of structuring. And if they have any opinions themselves, again, based on things that they've done, things that they've used, you never know what the case may be. But if you have an internal resource that can help you that knows a little bit more about things, no reason to go outside and try to build.
Nico (51:36.61)
the idea again, work smarter, not harder, right? And then when it comes to folding it in, honestly, I think Newbreed is definitely doing it the right way. And I would imagine a lot of other companies are gonna start to follow suit where it's a case of, hey, let's get the team together. Let's work on the prompts together. Let's work on how it can help our situation and how it can help our daily work.
And then we'll build a solidified catalog of here's the prompts that we use and how we use them. And we can go from there. And I think that is the best methodology to use is to let your teams give this tool to your teams. Let your teams work with it, figure it out, figure out what they can do that's going to help them and build that sort of library and vocabulary and those prompt libraries for you.
George B. Thomas (52:27.225)
Yeah, I would double click that. We also are internally testing prompts and getting them to the place where we love them and then have created a bank where team members can go and grab a prompt when they're working on a specific task that we work on. So you have to have some type of think of your internal knowledge article of a generated prompts that people can easily search and get to. So, Nico, this has been a great interview. I've loved hanging out with you. I do have a question for you.
One thing that you hope people take away from today's episode.
Nico (53:03.55)
If it's anything else, it's that this is the most important tool that has come out since the computer, and you need to know how to use it as soon as possible. Even if you aren't going to be using it in the foreseeable weeks, months, it will definitely be something that starts coming into your life within a year. It's going to be a tool. Look, this is how much I'm not joking about this stuff.
Go look for a job listing. Go look for any job listing. Guess what's going to be on that job listing under requirements, knowing how to use GPT. On most jobs that are getting posted today, that's now becoming a requirement. So if you think that I'm joking, I'm not, don't waste any more time. Don't waste your time. Guys, time is money. We all know that. We all wish we could have a little bit more time back in our day. This is the way to do it.
George B. Thomas (53:59.981)
So, Nico, if people wanna connect with you, if they have questions, where do you wanna point them?
Nico (54:06.782)
I want to point them to GPT with me. You can find our company page on LinkedIn. If you can't get ahold of us through there for some reason, even though we're watching it constantly, you can certainly hit me up personally on LinkedIn, or you can certainly hit up Christine on LinkedIn as well.
George B. Thomas (54:24.789)
Absolutely. Hub Heroes, if you're listening to this, the audio version, and you want more, again, head over to community.hubheroes.com because you'll be able to watch the video version. Until next time, remember everything that you've learned today, which I'm sure is a metric but ton, but also remember to be a happy, helpful, humble human, and we'll see you on the next episode of the Hub Heroes Sidekick Strategy Show.
Dude, I gotta stop the recording.