Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
AI In Content Marketing: Turning Process into Progress
192 Plays1 year ago
In this episode of the AI-Driven Marketer, Dan Sanchez is joined by Marc Angelos, a seasoned entrepreneur and content marketer, to delve into the aerial view of artificial intelligence in the marketing realm. They tackle the inherent complexity of AI, misconceptions held by company founders, and the evolution of marketers' roles in the age of automation. From the impact of DALL E on image editing to the nuances of voice-modelling AI, Dan and Marc provide a roadmap for integrating AI into business processes seamlessly. With insights on the "Eliminate, Automate, Delegate" principles, and discussion of custom GPTs for content creation, this episode offers a wealth of knowledge for entrepreneurs and marketers ready to harness AI's potential. Tune in to demystify AI's transformative power and discover how to stay ahead in the ever-changing skies of digital marketing.
Timestamps:
00:00 AI news show may change its name.
05:50 Content marketing agency owner discusses writing vs. video.
07:35 Learning about AI through fast trial and error.
10:11 Use AI to automate repetitive tasks for efficiency.
12:42 Using LinkedIn data for my own content.
17:13 AI can streamline creating a lot of content.
22:04 Blog posts for SEO are essential tasks.
25:42 Founders often act intuitively without realizing it.
28:59 Creative scenarios for future, automated by AI.
31:34 "Seeked marketing automation process, found guru executive."
33:36 Analyze, optimize, standardize, mechanize for work improvement.
38:08 Automatically enhance voices and match lip movements.
39:31 AI voice lacks human-like pauses and inflections.
45:19 Marketing work often involves discussing AI applications.
48:04 Workers need to adapt to AI technology.
49:49 Book identifies 3 distinct personas within individuals.
Recommended
Transcript
Introduction and Podcast Direction
00:00:05
Speaker
Boom, welcome back to AI in the Sky. We're covering AI news, what's going on for marketers. And I think this might be the last time we air it under this name because Mark and I have been talking and sometimes you just launch, I don't know, I just launch something and then you kind of make it better as you go.
00:00:21
Speaker
But I have a hypothesis on where things are going and where the show could go. So we even just talk about that live. Cause like, why not just do it live for everybody. Mark's like, I have no idea what's going on right now. So Mark asked me that question that you asked, and then I just refused to answer and I just pushed, uh, I'm like, Dan, how's it going? How's things been? How you been?
AI Hacks vs Learning: Risks and Rewards
00:00:42
Speaker
He's like, Oh,
00:00:43
Speaker
The classic and I gotta tell you I've mixed emotions this morning. I'm like so excited at the same time I'm just so sad this morning. I had a had a revelation this morning that So many people actually don't want to learn how to use AI Okay, this this I'm like I thought more people want to learn I'm like, you know, I'm starting to discover like people really don't want to learn they really
00:01:09
Speaker
want to be given quick hacks. They want to be given cheat sheets. They want the recipe, but they don't want to learn how to cook. They don't want to work.
00:01:19
Speaker
They don't want to work. So I'm like, interesting. I'm like, cause to me, the writing's clear on the wall. I'm like, people who only learn how to follow, like microwave, like only learn how to cook Mac and cheese and don't learn how to cook. I'm like, those are the, those people, you're going to be at the biggest risk for replacement over the next five years.
00:01:44
Speaker
How many people know that eating less and exercising gets you in shape everybody and yet weight loss pills sell like crazy still how is that possible?
00:01:56
Speaker
Because you just want to know, I'm Papa pill. And I don't need to bother learning AI want it now. Right? The AI thing is funny because it forces change. You have no choice.
AI vs Coding: Flexibility and Problem-Solving
00:02:06
Speaker
It's not like a nice to know. It's it's already here. It's just so sad because I'm looking at AI and I'm like, there's been many things that have been hard to learn marketing automation pretty difficult, especially in the early days, man. The amount of like crazy forms you'd have to fill out to be like when this happens, then do this and then do this and then do this and the amount of
00:02:24
Speaker
work you had to put into it. AI is like fluffy in comparison. It's actually very fluid. Um, there are some note, there's, there's things to know and there's places where it can go wrong, but for the most part, it's very forgiving. Unlike code, dude, you miss one freaking semi-colon and then the whole page breaks and you can spend hours, days, maybe even weeks trying to find that one place you screwed up in the code. I know. Cause I, I, I, I know. I don't know how to like,
00:02:51
Speaker
You use programmatic languages, but I code HTML and CSS. I design websites. And it was hard just to do that. Now I'm looking at AI. I'm like, this is way forgiving. AI will literally try to troubleshoot it for you. And it might not give you the results you want, but it's not just going to break. You can even ask it for where you went wrong and it'll probably help you try to figure it out.
00:03:11
Speaker
I'm going to use prompting as the example though, Dan.
Crafting Effective AI Prompts
00:03:13
Speaker
Some people, they all know like how you prompt the thing matters, but yet no one wants to learn how to properly prompt. We touched on this previous, but like there's a skill here that they're just like, well, just tell me what the prompt I should use. Yeah. And that's why the AI gurus are crushing it. I'm like, why is like, don't copy and paste prompts, learn how to prompt, like learn how to cook. It's not that hard.
00:03:33
Speaker
No, that's not, that's not what we want. That's why humans work because all the AI SAS people are figuring that out now and they're just making it easy and easier to use the tools. Right. Yeah. So that's what I have mixed emotions. And the reason that's, so that's the sad part of me, the excited part of me as I'm like, well, there's the opportunity for the freaking arbitrage. Cause a few people are going to learn how to cook and the rest, everybody else is going to be out. I feel like what you do as a listener, no matter who you are, you could just put,
00:04:01
Speaker
for AI as the end of what you do. Like here's how to learn such and such and such with AI. Here's how to learn fitness with AI. Here's how to learn writing with AI. Here's how to learn. It doesn't matter. The AIization of your thing, people don't want to know it. It's like they're looking for, just give me the cheat sheets. So that is opportunities, massive opportunity.
00:04:19
Speaker
So it, and that's why hence even I mentioned like, I might change up the show because I'm realizing like the majority of people don't really want to learn how
AI Education: Courses vs Quick Insights
00:04:28
Speaker
to use it. Like I put out one of the most robust posts I've put out on how to learn AI. It's like fricking crickets over there. I'm like, I'm like, that's, I mean, I've known this, but now I'm like, yeah.
00:04:37
Speaker
People really don't want to learn. I was thinking about building a course. I might still do it just because there will be a few that do want to go deeper. And I think if it's specialized, I think that's a good idea. People want to like, I'm in the thing and you tell me how to do the thing with AI, uh, that that'll sell. But when somebody's been like a long course, I mean, like you're talking like, I'm talking five video course on how to understand the framework that didn't go and test and do stuff.
00:05:03
Speaker
I think with education, though, I've written a lot of these online courses for other people, you need to make it as tight as possible. Like nobody wants to go through 800 hours of video. They want to just give it to me in one, you know, 12 minutes. Yeah, I'm thinking like, well, it'll probably be longer in that five minutes, like 20 to 30 minutes. Yeah, we'll see. It's I'm trying to make it a mini course, but at the same time,
00:05:23
Speaker
Nah, I'll see if I can keep another 20 minutes. That'll be the goal. It's like, keep it, keep it as tight. You know what I'd suggest? I mean, just my own thoughts and in my humble opinion, do a mini course upfront and see what the response is. Cause they'll tell you then, and then you build it on the go as, as people want more things around the thing. Like they'll tell you, can you dig into this more detail? And then you do that. So Mark, where are you at in all this stuff? What are you doing to learn AI
AI in Content Marketing: Challenges and Innovations
00:05:47
Speaker
right now? Are you, are you, are you dabbling? Are you digging in?
00:05:50
Speaker
Right, so I run a content marketing agency and a lot of what I do involves everything from video editing to just writing. So the writing stuff I know pretty well, I like to think I do. I dabble in multiple
00:06:03
Speaker
GPT versions, let's call it, and just to compare the products. And they are very different. On the video editing side, it's a lot more challenging for me because that's not my natural skill set. So you and I have talked about this, Dan. The fact that these video chop up edits, Zencaster, they all kind of suck to balance with you. They always cut it the wrong spot.
00:06:23
Speaker
I'm finding that's the case across, and I want to name names because some of these companies, I'm friends with the people. But the bottom line is, I just experiment same as you're doing, and I find out that A, I'm better at certain things with AI, I'm better at it, and B, the machinery is better at that type of AI tool, whatever it may be. So there's a variance there in both skill sets and tool sets, and I'm understanding that this is really a lot more nuanced than I realized.
00:06:48
Speaker
There's a lot to it. The more I play with it, and really, I'm really finding that building custom GPTs is by far the best place to play and test and learn. I had a chat with a gentleman that I need to have on my show soon. He was an industry veteran, and he was a VP of marketing, I think, at Alokwa. Something like one of those industry techs enrolled out a marketing automation tool.
00:07:17
Speaker
It might've been Marketo. I don't know. It's one of those like older, bigger marketing automation tools. He talked about the learnings there. Now he's working on his own AI tech startup.
Building Custom GPTs: Learning Through Experience
00:07:25
Speaker
And I was talking about just the things I'd learned in custom GPTs. He's like, Oh yeah, that's what we learned last summer. That's what we learned in September. Oh, that's what we recently are wrestling with. I'm like, there's things to learn here. And.
00:07:37
Speaker
If you're willing to roll up your sleeves, you can actually learn a lot about AI without having to work with developers just from testing it and custom GPTs writing instructions and then running your prompts against those instructions and the custom GPTs. I'm finding there's like a wealth of knowledge. And the cool thing is like it's trial and error, but it's so fast to trial it. Like you can write some instructions and then try it. And then, you know, it either gives you what you want or it didn't.
00:08:02
Speaker
And then you learn like bit by bit, like there's no, it's even private. Like you could just run it over and over and over again to see if it gives you the results you want.
00:08:11
Speaker
And it only cost a $20 a month subscription. Just to play. It used to be hard to write websites. I remember the first time I built a website, I was like, okay, like, okay, code. I understand HTML, but how do I get it online? And it used to be super hard. You used to have to go like, okay, web host. I paid for a web host. They gladly took my money and I bought a domain on GoDaddy. How do I hook these two things together?
00:08:37
Speaker
Now the learning starts. It used to be hard. Now it's, it's pretty, they've made it easier now, but like, it used to be hard. You're like, Oh wait, I need a, like an index.html file that becomes the, I don't know. There's all these tiny little unwritten steps. And even if you watched the videos and read the tutorials, they'd be missing a step depending on your web host. Right. And I'm like, AI, not, not nearly as hard to go and play with and figure out. It's very, it's much more intuitive.
00:09:02
Speaker
Do you remember the early days of social where you'd be on platform X? I can't even say that anymore. You'd be on one platform and then you'd be like, okay, it doesn't work the same way on the other platforms. Like the algos are different. Okay. I didn't know that. Then it was became like, they used to be thing around posting times, posting length, like all that stuff had to be learned.
Consumer Trends: Quick AI Insights vs Comprehensive Learning
00:09:18
Speaker
And now it's kind of common knowledge, but we're in that phase with the AI where the only way to learn it is to go dabble in experiment, but people don't want to do that. They just want to be told what's the cheat sheet.
00:09:28
Speaker
Well, and even the ones that are even more hungry, I'm noticing a trend because I've had people have noticed I've been posting about it. It's funny, LinkedIn reaches way down, but still it's working because I'm like, might get a few hundred, a thousand, a few thousand. Sometimes people are still coming inbound asking me about AI. I'm like, okay. They're like, Dan, what do I got to do to learn more? Like, do you have a course yet? Hence I need to put out a course because enough, enough people have asked now.
00:09:52
Speaker
But a lot of people are like, hey, I saw this, I saw this course by, I don't know, some online MOOC somewhere. And it's like, Luke Matthews, you pen e-course you can take online for somebody that's not accredited or anything. But it's like, there are things like there's a hundred of them out there. They're like full courses. I'm like,
00:10:10
Speaker
But nobody wants to get a chance to. I haven't checked out any of them, but I just intuitively know that it's not going to get you what you want. It'd be way better to go watch three YouTube videos and then just start building custom GPTs to build out things that you're doing, repetitive tasks you're currently doing. See if you can get AI to do it, even if it's ridiculous. See if you can get AI to do it anyway. Even if you fail, and I've failed many times trying to build these things, the learnings in them are way better.
00:10:37
Speaker
So two things, though. Number one, education happens on YouTube. I agree with you. That's something people don't realize. The folks are going to YouTube to find out what they should do. But two is the time factor of I'll pay money to save time. Same reason that nobody changes their own oil for their car. They'll just bring it in for 13 bucks or whatever these days is 30. I think that people want to just be told, what do I do? So you're being approached, Dan, because you seem to be knowledgeable about it. You're talking about it.
00:11:02
Speaker
I've gotten the same inbounds where people are like, well, you seem to be in the space. Like, we're all in the space. Everybody's in the space. But what do I do? That's the opportunity.
00:11:14
Speaker
So we'll see what people do. I'm going to keep chugging along. Um, I'm hoping more people, I finally had a friend tell me, Hey, I just bought, I finally just upgraded chap GPT to plus Mike finally do it more. Like these are close friends that are finally doing, I'm finally convincing a few people to like, Hey, like this is the best 20 bucks a month you can be spending right now. Even if it's just to learn and play and dabble, if only, but of course you can build your own custom GPTs and it becomes really like you start automating your work. It's, it's highly, highly, um,
00:11:42
Speaker
I don't know, productive for anybody, whether you're in company or you're doing your own thing.
00:11:47
Speaker
We're in a window of time where a gap is going to form between the people who start to dabble and the folks who don't. And it's not something you can make up later because the disappearance of jobs and all the stuff they talk about. That's a real thing. If you're good with it, you're going to know how to create value. And if you're not, it's, you're up against the clock here. Are you using, have you built a custom GPT? Yes. So what have you made? Just my own, like for me, the, like a, a more, what do they do?
00:12:15
Speaker
So mostly they write content and I don't love them to be honest with you. Like I've, I've created an AI avatar of myself using Hey Gen 11 labs to use the voice. It was,
00:12:26
Speaker
close enough that if you didn't know me, you'd buy into it. And it was freaking me out. I'm like, all right. I would put it out into the market as a, Hey, this isn't really me, but I'm showing this as an example. In the end, I didn't. I'm like, I'm not ready to, this is, this was the conversation we had around who's your mascot. I was, I went far down that road, but as far as taking the data dump from LinkedIn and making my own, you know, take all my old content right. Like me.
00:12:50
Speaker
I've done a lot with that and I actually, it's not bad as a junior hire, if you will, in my company that kind of writes like I do and then I can save time. I still don't think we're there yet though for any of these things and forget about like text to video. That's not, I'm not loving that yet. It's cool in general, but to be specific, like meaning I wanted to do exactly what I wanted to do, it's hard to prompt it, right?
00:13:13
Speaker
Are you using templates? I'm taking some templates from the internet. I've done a lot of the digital writing courses out there and a lot of the gurus. But if you have experience writing, or in my case, most of what I'm doing is writing,
00:13:29
Speaker
Then you know what you want like you have a style and that's Where templates don't always match up to what you're trying to do. It's the same thing with prompting Yeah, when you copy and paste these prompts, I don't want to name names because you'll know the people but like you'd be like Yeah, they're good writers, but they don't write like I write and I kind of feel like I'm a better writer no offense So yeah, that's where you're making your own templates. I've been doing that
00:13:51
Speaker
I'm slowly, I feel like where I'm at with content with these GPTs specifically is that I'm finding it's a three-way attack.
00:14:00
Speaker
Um, and I, I put, it's really, it's really just one freaking long mega prompt, but I, in my mind, I have it separated out into three different major sections. There's the prompt, there's the template, and then there's the example. I find if I can hone those in and those, then I tweak those over and over again. And I just run the same command. Like it's usually for me, I'm a podcaster, I'm using a podcast transcript as the starting point.
00:14:24
Speaker
Right. Okay, so and this is this is where the fun begins and learning AI. So I'm learning it now. I'm like, okay, I have the transcript and the transcript is what it is. And I'm going to keep running this as the starting point for this custom GPT. I'm just copying and pasting it and letting it run to see if it does well. The things that I get to tweak over and over again until I'm really happy with it is the prompt, a template, and the example.
00:14:46
Speaker
What the template is here is roughly what it should look like. You know, it's the insert this here. You can even build mini prompts into a template. Like I usually put questions in there like, well, tell me about this. Tell me about this. Tell me about this within the headings. It's going to do as part of templates that I'm finding out work really well. And that's where I start to troubleshoot little parts. Like if the intro is a little off, I start to tweak the template and it's almost many prompts within there. I'm asking questions for it to consider. And then the example is what does it look like if I were to do this by hand?
00:15:16
Speaker
or sometimes i'm like what's what's an example look like when a i really crushed it did really well it doesn't really matter who made the example but the example needs to be the ideal standard and i find between those three things i'm getting freaking close like for it to just go live and be bam done with minimal minimal edits but that leads to a problem dan
00:15:39
Speaker
both marketing and AI, the more people get acclimated to it, it doesn't work as well in terms of responsiveness, like people themselves. If I getting the same messaging from everybody or you copy and paste the template from some guru and then everyone in that community is using it, it doesn't work the same way. And so it's an evolutionary force, whether you like it or not, you have to be evolving your stuff. So you're smart to be doing that.
00:16:03
Speaker
And I think that of the three you just said that that's great. The prompt, the template, the example, the example is the weakest link for most people. Like telling it what you want it to, it should look like this at the end. Cause you really gotta be clear on that.
00:16:18
Speaker
The template and the example actually look pretty close. Um, if it's a really good template, then it should, it's almost looks like the example. It's just the examples more specific and filled in with actual information. Right. Um, there's downsides, there's upsides and downsides to this one. You have to know, you have to want to have a, like, let's say you want to turn a blog, you're, you're essentially saying I'm going to create a series of blogs that all have the exact same format.
00:16:44
Speaker
Mm-hmm, which means as a podcaster I need to give it a template because I can't just you can't build a process like this and throw a Very different podcast episodes into it from a five minute solo episode to a 25 minute interview Like it's not going to be able to take the very different inputs and it's not going to handle well through the machine So you almost have to think systematically like what can I give it over and over again as a podcast episode? Tighten up the process and you're almost thinking like manufacturing at this point
00:17:11
Speaker
It's almost like assembly line as a marketer. You have to think i'm like, yeah, but what do I want a lot of?
Systematic Processes in AI Content Marketing
00:17:18
Speaker
Can I create now now you're trying to this this is where it gets really hard with ai because you have to envision like If I had a hundred blog posts on x and the topic changed, but the template stayed the same What would be valuable? And that's where you have to start thinking Ahead as far as like okay ai can make it easy for me to deliver a hundred blog posts on this topic Because i'm going to make an assembly line to do it and it's going to do it relatively well I mean a lot of
00:17:40
Speaker
Bloggers and seo people already know this because they have to deliver articles in the hundreds But what you can't do is build the process and then change the freaking process every time Right the amount of work it takes to do the process you might have well just written a blog post by hand Because the process takes a lot of time to build but it's kind of becomes like that MBA thing where you're like Like how much do I invest in automation versus doing manual? Yep, there's there's upside to investing in automation if you know you're going to be able to
00:18:10
Speaker
uh, make, make back the, the investment, right? Over enough units, but you have to have enough units to make it worthwhile. And that's where, I don't know, that's where I'm at. Like, I don't think content marketers are thinking about it. Like think, don't think in terms of two or three blog posts, think in terms of a hundred and all of a sudden you'll start to see where AI can really fill in the gaps.
00:18:29
Speaker
So I'll make two more points on that. Number one is the process versus the workflow. And this is, I'm making up these definitions, but the process is what you do for each stage. So for example, the prompting, the template, the example, there's a little process inside each of those, I'm sure you follow, whatever the Sanchez model is. The workflow is how do they integrate together? Because you might want to change the workflow later or add new components of processes into that thing.
00:18:55
Speaker
And I tweak that a lot, but for my own stuff, for example, in repurposing, uh, you want to take old content on a regular basis and refurb it and get it back out there again on a new spin. If you can, that's a process, but it's actually got to be put into a workflow in degrading with my current content creation. So the strategic networking together, uh, you know, linking of different ways of doing what you do, it's very unique. It's got to do with the person, what you're trying to accomplish, how you work.
00:19:25
Speaker
I tend to write better in the mornings and in the evenings. I found that out about myself. So like all that stuff comes into play and you have to map this if you really want to create a conveyor belt of production as a marketer.
00:19:36
Speaker
I do. I want that conveyor belt for every single type of episode I'm doing. I'm like, I'll have a few different episode formats, but I just want this stuff to happen on autopilot afterwards. You put your finger on it, though, when you say a factory or whatever you called it. That's the concept of a content process. Most firms don't really work it through to that nth degree. The content process is there needs to be content made and posted and analytics be done on it.
00:20:00
Speaker
That doesn't get into the nitty gritty of, okay, what is the knitting together of the workflow, which might involve, you know, zappy or central, whatever. Like you might really have to get into the weeds with this. I think companies have to start thinking like that.
00:20:12
Speaker
I Used to work at a college and the college employed students. In fact, there was a work college So every student had to have a job Which every department then had to take student workers or I mean didn't have to most departments were fighting over them because it was free labor They worked over 15 hours a week. What a great idea that is And it yeah. Oh, yeah, and you know, we had it wasn't free labor. We were disk I would highly I fact there at ours college. It was tuition paid you didn't pay tuition You did pay room and board so you're paying
00:20:41
Speaker
I don't know, 10, 12 grand a year on room and board, which is a freaking steal for a private college. Right. Um, but they had to work 15 hours a week and it was great. And I, the, the lesson that I'm learning about AI now is that I learned more about how to do AI based on what I learned to building internships.
Intern Management and AI Process Development
00:20:58
Speaker
Really? Because I had to figure out how to make 18 and 19-year-olds productive in the marketing department. Let me know the secret, would you? We had a draft process for selecting them, and I got that many years in a row. I got it wrong because there's a skill in learning how to discern which students
00:21:16
Speaker
are just reliable. But that aside, I had to learn how to onboard them. And I eventually started with one student. By the time I left there, I had a team of maybe 23 students. Wow.
00:21:31
Speaker
But how do you make 23 people with no prior experience in marketing or media productive within a month and then working where you're not micromanaging everything because that's what you have to do if they don't know anything. Uh, so they're actually useful to you and you can actually still go and do what you're doing as a, as a director or whatever that was the game. And so I learned.
00:21:52
Speaker
how to train them for different roles. But in order to do that, you have to break every single process down that marketing does into a very dry step-by-step process. Here's the process. Here's how we write blog posts here. You cast the large vision. You tell them like, this is why, one, this is why we write blog posts. This is the goal of the blog posts. We write blog posts for SEO. Here's how search engine optimization works. You have to teach them that. And then you have to break down the very specific steps.
00:22:19
Speaker
I had one intern that did nothing but do the keyword research validation, writing the titles and would essentially set up the Excel sheet for everybody to then pull their so their blog posts, the titles done, the keyword objectives done, the basic instructions already written for each blog post.
00:22:34
Speaker
And then it was like again assembly line. I had a blog I had each I had probably a team of five writers do nothing but write articles and they did the same thing every single day They'd come to work and get started on the new article. They were expected to write an article a day We filled a website called just disciple calm with about 400 articles over two years Wow, and it brought in probably half half a million page views a month at easy I
00:22:58
Speaker
But that speaks to the results. The whole concept reminds me of the McDonald's speedy system, whatever the heck was. You put the fries in for only five minutes, when the people goes off, you always make sure you put the mustard on first, then come the onions, then come the... People need to learn that. If you're learning how to break complex things down into small tasks, that is the exact same process you use with AI.
00:23:19
Speaker
It's breaking down complex process projects into small tasks because yes, it can do good It can do pretty well in small tasks. Like hey, tell me about XYZ event. Hey, do you a crossover between how? Super Mario and George Washington are similar. It'll write that article It's like they're they're unrelated but like it'll find the common point right about it. It'll do it. Oh
00:23:40
Speaker
But your mentality is structural. Like I'm, I can hear how you think, Dan, you're a systematic thinker. Most people look at AI as I don't need to do this. The reason I have AI, so I don't have to do that. It'll do it for me. So I'm going to set this thing hog wild and be like, write me some blog posts. That's what most folks are doing with this thing. Yeah. I mean, even there, you can't expect it to just go on like, Hey, I'm in X industry selling Y widget. Uh, write me a marketing plan. Right. But it'll give you.
00:24:09
Speaker
a really generic, boring marketing plan. The philosophy there, though, is iteration. I know it's going to come up with garbage. I already go in knowing that. And now I'm just going to keep on going around and around on it until I like it. Whereas you can save yourself a boatload of time if you just said, this is the way it should be at the output. And you basically handle, like you said, with the junior employees.
00:24:29
Speaker
But if you had a step-by-step process for coming up with killer marketing strategies, and maybe it can't do it all, but you take into consideration what chatgpt4 can do, which is a lot. It can, there's, it's multimodal. It can deal with images. It can search the web. It can learn a lot by itself. And you're like, okay, step one.
00:24:47
Speaker
Who do you like it becomes a conversation where it kind of walks you through the process once it learns your product in your industry It comes up with a hypothesis on who your different targets are you talk about it like oh, yeah Yeah, that's that's definitely more of the market it goes and does research on that market because that's the next step of the process to understand their pain points and their frustrations by pulling blog posts talking about their pain points and frustrations it comes up with an analysis of that and
00:25:10
Speaker
Based on that it starts to think because you've taught it to Do an analysis on its pain points and how the product might be able to map to those, you know So now it's starting to actually do the work of coming up with a marketing strategy Is it the same as a marketing like genius to do it? No, but it's probably way better than what a normal founder would be able to do Yeah, but again, you have to break it down in the baby step processes and thinking about what it can it can't accomplish now and then have it run through a method and it be in that marketing strategy would be way better than
00:25:40
Speaker
if you would just ask it to write a marketing strategy. I'm going to take the counterpoint on this, Dan. Real talk. A lot of folks don't actually know why they do what they do, especially at the founder level. When you're busy, you wear a lot of hats. You just kind of know, like I was in frontline sales for many years. And if you tell me, you know, what's the process for selling? It took me a while when I found in my company to break it all down. What do I actually do here? How do I build rapport with the prospect?
00:26:03
Speaker
That's the same thing with marketing where you have to be like, okay, what do I do? And my process is built around what I would intuitively know, but didn't realize I did. And that's work. Like you got to sit down. When you want to get in shape, you eventually come to know like your macros and carbs and my, I work out better in the morning. Like you don't know that at the start. So you take some generic approach and it doesn't really work great.
00:26:27
Speaker
Well, I don't know. I think that's that, I guess that's really the challenge and why people struggle with AI is because they don't know how to teach what they
Articulating Processes: A Key to AI Success
00:26:37
Speaker
do. They don't know how to explain the step-by-step process. Even if they're really good at it, there are people that are genuinely just freaking good at what they do. They just can't, they don't know what their step-by-step process is. I'm going to go philosophically deeper on this. You might, I'm curious your response to this. I think this ties to the school system. You're coming out and you're used to just learn how to do a job.
00:26:55
Speaker
do the job well, you get paid. When you're doing an AI now, especially if you're a founder, but even if you're a senior marketer, et cetera, you got to put on the hat of like creative entrepreneur now, which is where you are. So you find this easy, but most people aren't used to thinking like stepping back and saying, okay, what is the functional purpose of my role as an employee?
00:27:14
Speaker
What do I need to have happen? And how do I get there in my head? And break it into steps? As you're saying, it's second nature for you, Dan. But a lot of folks are like, I come in, I push the button, I pull the lever. And you can't be that way if you're trying to really master AI. It starts with you before you get to the machinery. That's the problem.
00:27:35
Speaker
What are you doing? And then what is AI capable of and what can it do that you are currently doing? And even if it can't do it, how can you force it to do it anyway? Just to see. Right. There are certain things that I realize I'm better at than the machine. And I know it's, it knows everything, et cetera, but like, it takes less time for me to iterate this because I know it intimately well versus the other thing I'm doing with the AI. It takes me longer. It's me, but I'm just like, I'm better off if I just do this piece and pawn this off on the machine.
00:28:04
Speaker
So my point is like that's different for every person. I think there's a learning curve there with how you interact with AI.
00:28:11
Speaker
Yeah. I almost wonder if you could like hire people to help you break down what you do into processes because I sat down with a couple of people now, like building GPTs and I'm like, tell me about your process. Like you're doing this manually. You even delegate it. Like break it down, break it down. Now this, and I'm like, I just got hired to do build a GPT and it's, it's elaborate. It's, I love, I love, I eventually I'll have him on to talk about what we've built.
00:28:37
Speaker
I love it. It predicts the future. How do you like that? I'm sure this is about the financial markets, but keep going. It's funny enough. We are using it. It's not for stock trading. It's more for companies to predict where they're going to be, but a lot of their clients are in the financial space. It ends up being talking a lot about that.
00:28:58
Speaker
It really just builds up creative scenarios of the future and But they had a process for it, which is why I knew we could build a
Efficiency Gains Through AI Automation
00:29:06
Speaker
GPT around it because I'm like you have a documented process you have a clear input like you're starting with X and you're finishing with Y and you're always putting these in your presentations and it takes your team about two weeks to build the in-between and
00:29:21
Speaker
I'm like AI can do that in about 20 minutes, even faster now. So it's like, um, we just have to hone in what, what's like, what's the assembly line for what you're doing to come up with this in between so that it doesn't take two weeks. Cause AI can automate it and be done with, and you could probably get this whole process done in less than eight hours. So what became what was a two week process is now an eight hour process. And that is how AI is taking over jobs too. But at the same time, they want to become more efficient. So it makes sense.
00:29:51
Speaker
You made a point here. I want to press on a bit. You said, I wonder if there's a role where people can help you know what you do. This is Tony Robbins. Just not to just use an example. People would know.
00:30:01
Speaker
He'll up on stage, we'll chant whatever he says, where focus goes, energy flows. Like people are like, oh, I didn't realize that. Or like motion creates emotion. These are like tidbit sound bites that people walk away and think, oh God, I've learned so much about myself. But you need a Tony or whoever it is to teach you that. You happen to be really great making these GPTs with this company right now in your case, because they had a process.
00:30:25
Speaker
But that existence of a process, it's there anyway. Most companies don't know what it is, or they think they know because you're the high up boss and didn't realize that seven levels below you, something's happening in the works that you didn't even know happens. You know, like Alice runs some batch file at the end of the day because it just makes her life easier. She didn't tell anybody and it's been super helpful for the process. Like, but you wouldn't know that those types of things, like you really got to know the process, how it's done, and then break down and then build it into an AI thing.
00:30:54
Speaker
I guess that's the trick. How do you become better at process? I hate that. Like I run a company now and so I have to have that down, but it's not my natural skillset. Like I'm a ready fire aim kind of guy. And so I think that people need to become more processed, systematic thinking, more oriented towards orderly. My wife is great. She's very orderly. This is just like we're opposites. You, I can tell Dan, you're very systematic in your thinking.
00:31:21
Speaker
Folks like you are going to do really well with this because you take fish to water with this. And your average, I don't want to move on to sales guys, but like your average ready, firing person is going to be like, ah, I don't really know. Like what's the, just give me the shortcuts.
00:31:34
Speaker
I used to be obsessed with marketing automation. So I would ask people because I knew there was, I'm like, surely there's a process for process development. I asked a lot of people, even people that were killer, like they were excellent at creating processes. And I often got like, there isn't a process for that. I finally found, I did finally find a seasoned retired executive who was like,
00:31:56
Speaker
high up in the png line i was working for a non-profit and he was just kind of he's one of those people that just kind of like just shows up and just kind of like just kind of i don't know like he's he's such a guru so high up that when he shows up you just come up with a pen and paper and a room full of execs and just take notes as he talks right but i didn't think there was a i was convinced that there wasn't a process for building process but my boss at the time he's like what's your process for building process because the conversation we were having he's like
00:32:22
Speaker
Oh, yeah.
Framework for Process Development
00:32:23
Speaker
Analyze, optimize, standardize, and mechanize. I was like, there is a process. I have it in my notion for every time I have to bring it back up, I just pull it up under notion now. Analyze, essentially document what currently is happening. What is the current process?
00:32:40
Speaker
Optimize how do you make the process? How do you think you might be able to make this better? Standardize how do you make sure it's like optimizes like building the plan for how to make it better? Standardize is making sure everybody knows what this plan is and it's actually like
00:32:53
Speaker
Documented what the new plan is and everybody's educated on it and then mechanize final step automating it Right. Where can it be automated? That's the process. So this is the process and of course He would know because png goes through this a lot. They have a lot of manufacturing plants. They have a lot of processes They have to mechanize working with the targets and walmarts of the world extremely process driven otherwise just-in-time manufacturing and delivery and fulfillment wouldn't work in that world and
00:33:19
Speaker
Right. Right. People talk like if the supply chain broke, Walmart would be out of groceries within a few days because it's just in time. It probably only got there two days before you bought it, maybe. And to run a just-in-time business at Walmart is a good example. You really have to have that system down.
00:33:37
Speaker
So what do we learn from this? I honestly think like every job can run through this process. What do you analyze? What are you currently doing that's delivering value to your employer or for your own business? Like what are the processes? And even bullet, like it could be as simple as like a five point bulleted list. How could it be done? And then optimize, how can it be done better? Standardize, how do you make sure that it's actually happening to that? And then mechanize, how do you automate points of it so that you're focusing on the points that can't be automated? Right.
00:34:04
Speaker
I think that's a really valuable tidbit right there. What I've heard it is ID8 automate delegate. So what do you think through what you normally would do? Automate what you can, there'll be some piece you either can't or don't want to, because you're better at it. And then can you delegate that if you have to? So there's a whole, yeah, this whole concept of thinking systemically, it's a great point you're raising. Yeah, I mean,
00:34:28
Speaker
I mean, that reminds me of like one of my favorite parts of the four hour work week kind of had, it's funny, it was sitting under my nose the whole time. Like the process, the similar process is in the four hour work week. Right. Like I forgot his book is that eliminate. Yeah. Eliminate, automate, delegate.
00:34:46
Speaker
delegate. Yeah. So that was his, that was his thing. What are you doing that you shouldn't be doing? Yep. To improve a process, just get rid of crap that doesn't need to be done. Especially at the bigger company levels. Yeah. How much wasted time are they? Like you, you work the companies I have. You're just like, are you kidding? This, this is the part of the process that we need to be doing this. Yeah. Yeah. I, I think that it requires you to sacrifice some sacred cows and that's a
00:35:17
Speaker
problem inside most companies. I don't think I could work at a big company for that reason. I tend to be like going cross department and different lines and going all rogue. This explains why the smaller firms are more adept in general, but specifically with AI. A guy like you, you're all over this. These large, large companies, they're looking for what AI-based tool can we integrate into our existing product suite, but they're not really trying to AI what they've got. They're productizing
00:35:45
Speaker
the technology versus let's leverage the technology.
00:35:49
Speaker
So if you're listening to this far, key points that I'm kind of remembering back now are like learn how to leverage this technology. If you can, like learn how to cook rather than just follow recipes. Yup. Learn how to drive process. Learn what your process is for delivering what you do. There's multiple frameworks. We just rattled off like three of them. You can go back record a rewind and go through a couple of them. All of them are lots of overlap between them. I do want to dump into like a few news, news pieces, because there are some interesting things that have happened in the AI world.
00:36:20
Speaker
Over the last two weeks, that'd be fun to cover. It wouldn't be a week if AI wasn't reinventing itself every couple of days. I know, it's happening fast. OpenAI had some updates, very exciting. One is like, again, they're starting to do this thing where they announce things and don't release it to anybody because
00:36:38
Speaker
They're scared because it's too powerful. The one I, the one that was the most noticeable is voice modeling. Do you see this update? I did. You're like, Oh, like you can feed it literally just a 13, 15 second voice clip of your voice, just talking, reading anything and it's ability to match the sound of your voice with any text read. Just reading through any text is remarkably good. Now you're like, wow. Like it had some samples and it was so freaking good. It's going to be incredible.
00:37:07
Speaker
They're doing what the social platforms did when they all would copy each other's features. OpenAI is stealing what 11 Labs had basically, you know, was the forerunner of this, but they're doing a better job of it, Apple style. But yeah, I mean, the voice, the voice mimicking, like all of this in my mind leads to the problem of
00:37:24
Speaker
who owns what content is real. And that to me is like a blockchain conversation, but we'll talk about this by the time.
OpenAI Voice Modeling Updates: Implications and Concerns
00:37:30
Speaker
I feel like that this is going to be copied. If someone has a better widget, open AI is the hub that's going to be like, okay, we'll take that and we'll just stamp that out better version.
00:37:40
Speaker
What I'm excited about for this for marketing is that, especially for podcasting, it's going to do a better job of auto editing your, your, your video file or your voice file. And of course it'll match, it'll move the lips to change it, but like people don't speak grammatically correct. And that's okay. People are used to kind of hearing people just run a rumble through a conversation.
00:38:01
Speaker
like even me like rumble through a conversation i could go back and auto correct that and put like maybe ramble or something like more appropriate yeah like it'll just fix fix that and post automatic because it'll be able to match your voice match your lips to it and just gonna it's gonna smooth everything over everybody's going to sound way more intelligent and it'll be it'll be that's not even the stuff ripping off like of course we're scared that we're gonna make some politician sound like he said something and everybody knows that's a fear and there's that that is the official things are going to sound more articulate
00:38:31
Speaker
Like you will be able to just ramble through you talking through something in a solo episode, hit like improve, and you were just going to sound way more articulate than, than you actually are. Yep. Which again, coming back to your point, that becomes a live sessions and public speaking become way more authentic now because everyone's going to be faking it through recorded stuff. A hundred percent. I actually think that the human side of making mistakes or stumbling over your words becomes the appeal.
00:39:00
Speaker
Yeah. So real fast on that point, they were the first thing they did with the ability to vocally inflect the way a human would um, and ah, right. The machine now hesitates and pauses and talks the way a human talks versus talking perfectly. And that's, you're basically programming in stutters and pauses. And like, let me think about that fillers because it sounds more human, but it doesn't need to. And yet that's what people want.
00:39:30
Speaker
Yeah, it still sounds, I mean, it sounds human, but it doesn't speak exactly like humans do because there's more pauses, some well, and so well intentioned and some not. Um, there's all kinds of little voice inflections and oddities that AI just tends to kind of like the writing, right? The writing.
00:39:50
Speaker
Doesn't have a lot of flavor to it right sentences are all kind of similar in length instead of having like long mixing it up or short and like Having more drama to the writing, but it's just like the writing even though it's all grammatically correct and sounds appropriate Mm-hmm. It's still not fun to read right? It's too repetitive
00:40:10
Speaker
And not even his content, just in its style and its ability. Like you've, you've seen the copywriter thing where it's like, this writing is boring to read because it all has the same length sentence and they all sound the same. And then it goes and writes this like fantastic paragraph with like a, like medium, long, short, and then like really short, like two, two word sentence. And it sounds like a crescendo and you're just kind of like, ah, I wish I had it in front of me, but you know what I'm talking about. I know exactly what you mean. You've seen this social post that's went viral a few years ago. Yeah. I think it's gone viral a few times because it's just that strong.
00:40:40
Speaker
Like the power of writing. You can't do that yet. It'll get there. Yes.
00:40:45
Speaker
No, I agree with you. Also, one of the quick side point, OpenAI makes the announcement, doesn't release it right away. You mentioned that earlier. That's a marketing play in my mind because you get the power of the one is the release and two is the announce. But anyway, keep going. Yep. No, I definitely think they're smart because, you know, Anthropics coming out with it's like, oh, we're just above, we're now better by a margin, but like we're now better. So they, I think OpenAI is just throwing out more news like, well, we did this. Well, we did this, but we're not going to release it to you.
00:41:15
Speaker
Just to stay top of mind Dali now has editing which is very exciting like To the point where Dali's continuing like it's falling behind mid-journey Which is so exciting because to some degree Dali's a better image tool than mid journey is Mm-hmm. It's certainly more approachable for most people. Oh, yeah anytime Dali makes an edit I'm like yes, cuz it's just easier to work with and I honestly feel like I wish mid journey worked more like Dali but
00:41:39
Speaker
It doesn't. That's a really big point though. Like the UX matters. If I, if it's hard for me to use it, I'm not going to use it. Well, let's open AI is made AI to make it easier to make the stuff to intuit, to intuit what you actually want in an image versus mid journey. It's AI for image creation is good, but it's AI for understanding what you want is right. You have to know exactly what you want to get it to be right.
00:42:04
Speaker
So that's exciting because marketers, most marketers can't design and most marketers can't make these things unless it's a highly done, well done template in Canva. So Dolly's going to blow it open for marketers to actually be able to do their own design. Whether that's good or not is another question, but they will.
00:42:21
Speaker
I saw a first mark capital put out this infographic of the AI landscape. Have you seen it? No, I haven't seen it. I'll just describe it to you. It's not even worth like really looking at because it's like, you've seen like the SAS, the SAS debt diagram with like all the different categories and subcategories and all the logos in it. It's like that, but it's already like massive with like a thousand logos in it. You can't even read the logos. They're so small just to show like the AI landscape. And I'm like,
00:42:50
Speaker
Yeah, it's primarily startups right now. Well, it's all startups, but it's all the other legacy tools that just added either just tapped into open AIs, API, like everybody's in AI company now is essentially the point.
AI Startup Landscape: The New Norm
00:43:02
Speaker
That's how you get money. Every software company is AI company. Okay. Got it. It's just, it's the new standard. It's the new standard. Yeah. It's a good point. By the way, on that point real fast, you're right. They're not all really AI companies. Like it's the, it's on us. The onus is on the user to be like, okay, is this company really, is this a legit AI tool or is it's GPT wrapper, which they mostly are. It's not really an AI company. Like you can pipe in AI into anything.
00:43:30
Speaker
What would you call them if it's like, if it's primary primary use case is essentially GP at GPT wrapper, but without the AI, its functionality is, is dead, is, is nothing. I wouldn't call it an AI company. Like I would call it a GPT wrapper, which is kind of a, uh, an insult and I mean it as such. That's the fact that you can API in your AI into what your existing widget does is great, but that doesn't make you an AI company.
00:43:56
Speaker
It's no more so than if I use code and now I'm a code writer. Yeah, of course. You can make it into what you... This is a marketing play more than a business development play. Yep. What would you call them then? Because they're not going to lead with GPT wrapper. Right. I would probably call them... Officially, I have to think about this as a marketing term attached, but off the record, if I'm talking to you behind the scenes, they're like imitation AI companies.
00:44:26
Speaker
That's really what they are. They want to be AI and they want to be evaluation with an AI number, dollar sign, but they're not developing artificial intelligence software inside their company. There is another term that's emerged from Gartner called IA.
00:44:44
Speaker
intelligent automation. I've heard this. And I'm like, that kind of makes sense as a category of what a lot of these would probably fall into because they don't want to be called GPT wrappers yet. The automation wouldn't work if it weren't for the AI component, which they're just outsourcing to anthropic or open AI or something like that.
00:45:03
Speaker
Yeah, so this begs the question of what makes a company an AI company, to your point. The IAG term, I've heard that thrown around. It's cute. I get why they do it. It's still confusing, though. This is my earlier point a second ago. There's no line between what is AI and what is not. I told you I've done marketing work for firms that said to me, to my face, we need to make this an AI spin. And I said to them, is there AI in there? And they're like, well, I'm sure somewhere. So OK, what is the definition of AI?
00:45:32
Speaker
If it's an advanced algorithm, is that AI we ran at this on wall street all the time. It's a trading black box. It, it learns as it goes. So there is some machine learning there, but it's not definitely, it's definitely not AI. Not as we see it today. Yeah. You kind of have to train your own model. Right. And then it's kind of your own AI ish. Right. Even if you're just training it on your 10 books that you've written.
00:45:55
Speaker
Right. Exactly. I think that this, this debate won't be solved and it's fine, but like what makes an artist, what makes you a content creator? That's the definition. Like what is that is my mom is a content creator. If she's posted on Facebook, like, so any, my point is that that definition of AI, it's going to, it'll always be fuzzy and companies will always try to grab onto it.
00:46:16
Speaker
Yeah. And like Alex or Mosey says, like you can't really sell AI because you, I mean, it makes for good YouTube videos and content platforms like this one. But at the end of the day, you'd never sell AI because people don't want AI. They want the outcome of AI. You're still selling that. But I can market spin of I'm the AI of ice cream or I'm the AI of lawnmowers or whatever it is.
00:46:37
Speaker
Yeah, right now it's a grab. Hence marketers are tired of the hype because that's what it is right now. Everything's a PR stance and a positioning stance of having AI right now while it's hot, but now everyone's getting tired of it. Yeah. I think actually that's a really good point about whoever's doing marketing right now. There's going to be a pushback against IA masquerading as AI, and I think that people need to take into consideration that there'll be eventually a consumer assessment of what real proper AI is.
00:47:07
Speaker
That's what I love. I love the new territory because everything's still wet cement right now. Before digital marketing was digital marketing, we called it e-marketing. We called it internet marketing. Are all these things like e-marketing, internet marketing still has a little flavor of those Dan Kennedy bros.
00:47:30
Speaker
And we called it a bunch, we called it a bunch of different things. And then digital marketing kind of became the turn. So we still have a ton to learn about like where this new era is going to take us and AI and how we're going to be calling it and how we're going to be addressing it and labeling it and categorizing and organizing
Documenting Processes for Effective AI Integration
00:47:47
Speaker
the whole thing. But
00:47:49
Speaker
We'll see where it goes. I still think your point was very valid about coming up with the process internally. Like if there's some takeaway from this, it's make sure you understand what you do internally and then put that down on paper, almost as if you're hiring an employee and then you can work with the AI.
00:48:04
Speaker
Yep, everybody should be doing this for your job right now because if you don't do it, somebody else is going to do it and then you'll be the one who loses the work because somebody else already did the process which eliminated your work. But if you're the one who builds it, then you have the advantage of knowing all the ways that it didn't work and did work and can keep up with the development of AI in order to make it better.
00:48:26
Speaker
And your job now becomes managing all the things that, like if you were a copywriter, you are now a copywriter manager managing all the copy of all the AI. That's kind of my hypothesis is that all frontline workers will essentially be managers in the future, managing their AI teams. Which takes us back to the point though, you have to think like a business owner about your own role.
00:48:48
Speaker
Yep. Um, kind of brings me back to the E myth. One of my favorite books. It's essentially the, the, the best book ever written on process, because it's so approachable and it's all story driven about why process is so important. Yeah. That's a great book. Finn has not read that. It's a great, I mean, it's made for like small businesses, but I'm like, anybody should read this for their work. Cause you are a small business. Even if it's you, you and your career, you are a business, you're just being hired by an employer to
00:49:16
Speaker
You have an, you have a customer of one called your employer. Yeah. One of my whatever we'll talk about that book later. That was a great book. It was one of my big influences. I love that book.
00:49:27
Speaker
Yep. So I think I need to round up books like that because I think these are the types of books that are the undercurrents for really learning how to leverage AI. And it's process creation. And the best way to learn process creation, in my opinion, is one of them is learning reading the e-myth. Right. Because that's really the why behind why process automation or why process development in general is so powerful.
00:49:49
Speaker
So one of the key points in the book is that inside each person, there's three separate quote unquote people, right? There's the visionary, the operator, the manager, the technician. And so that book forces you to analyze yourself and break down which one you're strong at, which one you're not. And that mental walkthrough is what AI will demand of each person. Whether you've read the book or not, I think the book is going to be helpful to your point, Dan. But it's about coming up with what should happen
00:50:18
Speaker
What do I currently do? And then you can optimize with a machinery of AI. Dang, I'm glad you just broke that out because I wasn't even thinking about how I knew the book talked about process, but I forgotten that the major point of the book is that there's those three roles that everybody, every business underneath to have.
00:50:34
Speaker
I can tell you now, like I'm, I've worked in finance, but I'm a words guy, not a numbers guy. So you need to know these things about yourself when you're going to build a tool into do things that you do. Like you got to know this is the parts of me that I'm trying to put into the machine and not because the entrepreneur needs to dream up coming up with the system. I almost think manager needs to document it so that AI can do it. And then technician actually has to have the skill to know whether what the AI does is good or not.
00:51:02
Speaker
Right. And then there's the iterative process that you always talk about where it's, how do we optimize this? What could we do better next time? It's entrepreneur, you got to play. Right. Try out stuff. And you got to be willing to suck, right? You're going to come up with something for your own company's process and it won't be a hundred percent, but it's better than we've got. And then you'll figure it out, but that's, you got to start.
00:51:21
Speaker
Yep. Well, I'm inspired. I need to go make some more custom GPTs and document the ones that I've made for others to follow along with. Mark, thank you so much for joining me again on this episode. Pleasure as mine, Dan. For everybody else, if you miss it live, it's on AIdrivenmarketer.com on the podcast. And of course, I think it lands in my LinkedIn feed or Mark's LinkedIn feed later on. Yeah. You guys can see the clips.