Ep 72: AI in Marketing - Reality vs. Hype
Watch the YouTube video version above or listen to the podcast below!
Episode Summary
This episode explores the divide between the hype surrounding AI in marketing and its actual effectiveness on the ground. Hosts Alex and Dave kick off with the observation that many corporate leaders sell AI's promise of efficiency to shareholders, but day-to-day users often see little real gain.
They emphasize that AI tools are mostly used for basic tasks—email summarization, meeting notes, and replacing Google search—suggesting missed opportunities for deeper integration. The conversation unfolds into the psychology of work habits, internal resistance to change, and the complex dynamics of change management. Even when tasks are automated, if employees revert to old habits, the tech fails to bring true efficiency.
Alex suggests that AI is often best applied to bridge “unrealized opportunities”—tasks that currently don’t get done because of friction or lack of interest. Examples include helping with performance reviews, where tone and structure matter, or summarizing lengthy content like podcast transcripts for strategic insights.
Dave adds practical examples like streamlining UX research with AI prompts and highlights the importance of presenting AI as an enhancement tool—not a job replacement—especially in sensitive organizational cultures. The episode also tackles tech debt, admin overload, and multi-platform content fatigue, questioning whether AI's ability to “do everything” is always worth the cost.
Throughout the discussion, the hosts return to the importance of context: company size, team dynamics, and strategic clarity all shape AI’s effectiveness. They end with a critical reflection on AI ideation—how generative tools often fail to produce resonant content unless deeply tailored.
This is a must-listen for marketing pros, content creators, and business leaders wrestling with how AI fits into their operations—and where it still falls short.
Ep 72: AI in Marketing: Reality vs. Hype Podcast and Video Transcript
Dave Dougherty: hello and welcome to the latest episode of Enterprising Minds. We've got Alex and Dave here today. Alex, we'll kick it off to you for the topic introduction and we'll get going.
Alex Pokorny: You got like a smooth jazz voice going on, Dave, us all. Just ease
Dave Dougherty: Welcome to the radio.
Alex Pokorny: I love it. Comment with us if you want Dave to start that way every time and we'll find out. That would be pretty funny.
Exploring AI in Marketing
Alex Pokorny: yeah, so topic today was basically is the reality of AI in marketing. So we've talked a lot about various scenarios that we've been playing around with, have tried the tools, how they've changed. We've even talked about a little bit of, the messiness internally, but I really wanted to play out the whole situation.
So a little bit of that is, it even worth AI in this scenario? is it efficient to be taking an AI tool and spending the time to automate? That was concept for the beginning of this podcast. So then we'll get let's play out a scenario of posting a blog post. If your blog post process currently takes six weeks, but your AI tool helps you save a couple of hours. We're not really seeing the benefit, but where can it actually be used along that whole process, where are those opportunities and where can you finally get that efficiency, that kind of AI promises that are out there would actually realize.
Dave Dougherty: Right.
Alex Pokorny: let's get into it talking about at least of you'd mentioned about a Will Reynolds post recently, but also about is it worth the time to automate
Challenges and Realities of AI Implementation
?
Dave Dougherty: Yeah, I think there were a couple of things, in our pre-recording kind of meetup couple of things that jumped out was, a recent comment about how. The leaders of companies are really excited about AI and they're going to shareholders and talking about, oh, all these efficiency gains and optimizations, which of course is exactly what Wall Street wants to hear, right?
But then when you go to the actual employees who are using these things, they're like, nah, I'm not really seeing any gains now. Part of that, there's there on the surface. It's okay. That's interesting. When you go deeper into that, it's what are you using, how are you using it? At least from some of the stuff that I've read up to this point is that a lot of the self-reporting sort of surveys from people who are using AI right now for work, it's still.
Summarize my emails, get my meeting notes, or replacing a Google search with an AI thing. So you wanna talk about unrealized opportunity? That is unbelievable if you're just replacing your search behavior with ai that yeah. I want to not even go down that rabbit hole. Like you can't even talk about productivity if that's what your employees are doing.
So that would ultimately mean there's a lack of education, a lack of knowledge, lack of desire to really, dig deep into these things. And I know I've experienced that with a couple of my coworkers where they have their ways of doing it and they, they know it's inefficient, but it's their process.
So they continue to do it that way. Part of that is that psychological need to feel like, okay, I'm in a certain spot with this project. When you're running through your own process, you have a gauge of, okay, I'm 40, 50, 60% done because I'm at this stage of my process, to your, to the Will Reynolds nuance on top of all this.
He had an interesting discussion on one of the socials, either LinkedIn or X or something. Where he was, he pitched the idea of diminishing returns where your use cases get specific enough that it's no longer worth building it out because instead of realizing the efficiency gains, you are fine tuning and tweaking so much.
That you're no longer actually getting the output. Instead, you're managing, managing the bots. I know I've experienced this. I think, with some of your examples you've brought this up.
Alex Pokorny: Right?
Dave Dougherty: That's the kind of setup for our discussion. How has that resonating with you?
Have you experienced that with your coworkers? Anything that you've been reading?
Alex Pokorny: Sure. Yeah. There's a couple parts there
Short-term vs Long-term Gains
Alex Pokorny: . One is short-term versus long-term gains. I
Of employees always look at things from a short term, how does this affect my day-to-day job?
The boss says, maybe they're throwing out the buzzwords, just to get the trending kind of phrase to make Wall Street happy.
Dave Dougherty: Right.
Alex Pokorny: they're also saying that with some idea of long-term benefit. Yes, we're not gonna see benefits absolutely today, but over time we'll see this. there's a huge gap in between those two points of view is. actually implementing. And that's, I think part of the piece that you're getting to as well is there's a problem right there in the middle of of what AI could actually do for a particular business process.
How do you actually implement that and is it worth it Because there's pieces that you do on a routine basis that great. I have a new piece of software that makes it faster. But I only do it
It makes the report that I need to put together every month so much easier. I got that slide out in five minutes versus an hour.
Great.
Dave Dougherty: right.
Alex Pokorny: Is it worth getting a whole new tool for that? Is it worth getting the budget for that approving this whole process? No. No. Probably not. It's likely pretty unrealized benefit right there. But there's also the other piece that gets unspoken, which I think the bosses always get interested about, and the employees want. Is unrealized in between again, which is also opportunity. new things can we do because of ai and if AI changes things so significantly different, for instance, the time of writing blog posts if it's reduced so low or internally, no one wants to be the one to write it and it gets us over, that bridges that little moat that we have a little problem that wall. Then it's worth it again because then it is an activity that's not being done that
Dave Dougherty: Right.
Alex Pokorny: be done, and there's opportunity there to then take advantage of that and benefit of that. But there's also, throughout that entire thing, we're also talking about throwing tech on top of things.
Dave Dougherty: Right.
Alex Pokorny: doesn't, that just doesn't make things better. It creates more work.
Change Management and Resistance
Alex Pokorny: It's the change management, resistance, which I
Dave Dougherty: Right.
Alex Pokorny: totally seen. I did a project once we saved, it was like 20 something plus steps and we got it down to three and we found out that people were still doing the 20 something step version. That was poor change management.
It was, okay, we got this cool tech thing.
But was it communicated? Was it actually adopted? Were concerns actually voiced and heard and responded to? there champions to actually push people along? Because it's one thing to have some random person from some other department say, Hey, you should do this.
It's a whole different thing when it's when your colleague says, Hey, I found a faster way to do this.
Those are totally taken differently. One of 'em, you, it's almost automatically met with resistance. The other one is met with interest. And if you're this outside group pushing in pushing, you're gonna start hitting walls just because people don't wanna do it.
They just get scared of it, or it's just, it's more effort
Dave Dougherty: Right.
Alex Pokorny: Current process is easy, mentally, easily done. So I guess I've seen those different pieces. I'd almost advocate for an AI group or an AI team. Committee. I don't know. I hate the phrase committee because it seems like too bureaucratic,
Dave Dougherty: Well,
Strategic Use of AI
Alex Pokorny: but trying to find those opportunity pieces versus like job replacement things or job efficiency things.
Because honestly, those gains are probably pretty small the effort is probably way too high.
Dave Dougherty: right.
Alex Pokorny: the person will go after it if they really care. I don't think it takes an outside committee to be like, I'm gonna interview your job and try to figure out where I can automate it. That's gonna be met with resistance to begin with,
Dave Dougherty: And that's one of the things Yeah. Where people are using it for things they don't like. Towards the end of last year we talked about, people leveraging AI for their performance reviews.
Alex Pokorny: Yeah.
Dave Dougherty: Stop what you're doing and talk about everything you did for the last 12 months, or at least whatever you can remember you did in the last 12 months.
And oh, by the way, your pay is dependent on it. What?
Alex Pokorny: And
Dave Dougherty: Like
Alex Pokorny: hours to respond.
Dave Dougherty: Yeah, exactly. Yeah. So
Alex Pokorny: Yeah.
Dave Dougherty: that. That ticks off a number of things. Is it useful to the individual? Yeah. Because now all of a sudden you don't have to be like, man, am I using the right words? Am I gonna be punished if I sound apologetic? No, you just change the tone in this thing. Make it appropriate for, end of year performance review done.
Alex Pokorny: Yep.
Dave Dougherty: But the huge processes, like I started writing down for my role, all of the AI pilots that I've done, like with individuals in New York to be like, Hey, let's try this out. And it came from a simple premise of it sucks when if you ask somebody that you'll get. 5, 6, 7 ideas right off the bat.
Alex Pokorny: Yeah.
Dave Dougherty: Right?
Alex Pokorny: Yeah. No, that's a good way to pose it.
Dave Dougherty: Because then it'd be like, okay, so when I come in every day, the coffee sucks. All right, I can't do much about the coffee, but let's mark that down. 'cause you need a good environment for creative ideas, right? Or. If you're in a big, enterprise or a company that, isn't as up to date with its tech stack, right?
I have seven Excel sheets that track projects or track statuses of things that mostly say the same thing, but I have to spend time every week updating these things.
Alex Pokorny: Yeah.
Dave Dougherty: in those instances, I've stopped and I've gone to my boss and said, look, do you want me to actually work or do you want me to talk about working?
Because right now it seems like you're asking me to talk about working instead of working. And no, I don't want to fill out seven spread. Why? Why? So in that case too, I think it really does depend on the type of company you're in and the size of it. Because in the smaller orgs, if it was like, oh, I have seven tracking spreadsheets, great.
That's a project management problem. Let's get something like an Asana or a monday.com or something where people can. Enter in things and you can enter dependencies and, oh, hey, this part of the project's done, kick it over to Johnny. That's an easily solvable thing. However, you're in a big enterprise.
Do you want to go spend six to nine months advocating for a new tool to then do project management when you know they're gonna come back and say, yeah, we already have a project management tool. So just go use that and go. Yeah, but the problem is it sucks and nobody uses it, so
You know, we need something else.
I don't have the willpower for those kind of fights right now, and I don't think any of my coworkers do either,
Alex Pokorny: yeah that's a key point though.
Efficiency Gains and Practical Considerations
Alex Pokorny: I think that's one of the, if we're looking at efficiency gains,
information transfer, transparency between individuals and teams
almost always a gap.
Dave Dougherty: Yeah.
Alex Pokorny: is never done well enough. It's just, extremely hard to ever get to that point.
Dave Dougherty: Right.
Alex Pokorny: And that is probably one of the hardest pieces to realize unless you really focused in on this, like you really made a project out of it. Of I'm gonna interview a bunch of people about this particular process,
of how we actually do at this company. And then you would probably hear. I do my part and it's actually not that bad, or there's parts of it that aren't great, but it's overall it works. then I wait for three weeks for this person to reply. And
Dave Dougherty: Right.
Alex Pokorny: the other person is by the time I get to this, it's four layers deep. I forgot to log in because I only do this once a month and it takes me forever.
And I finally do the thing and it takes me 30 minutes. And it's like they wait two weeks. How many weeks for you to do a 30 minute task?
Start to realize like those gaps like that,
Dave Dougherty: right.
Alex Pokorny: That's a difficult one, and that, that's actually a really good, it's a good and a bad use of ai, so it's a good use of AI to try to make that process easier and simpler.
It's a bad use of AI because typically the tools aren't meant. For team-based sort of things. Even multiple people sort of things.
Dave Dougherty: right.
Alex Pokorny: GPT. Just now, you can start like a couple people all working on the same, chat stream basically. You can share it and everyone can communicate in with it.
Dave Dougherty: Right.
Alex Pokorny: Before that, it was all solo and so solo effort kind of stuff.
That's a gap.
Dave Dougherty: And I know for me, like when I was doing the individual, pilot projects, a couple of 'em.
Most of 'em actually I would tell somebody like, Hey, we could do this process this way with this prompt. And they go that's this department. Have you talked to anybody from there? I said, no. 'cause they're gonna freak out because that's their job. And that's 90% of their projects right now.
Should they be doing smarter, better, more important things? Yeah. Yeah. For example, the, one of the prompts was a UX thing. And the UX thing was, tell me, does this fit an informational, user intent or should this be moved into the catalog because it's better suited for, commercial and transactional intents,
Alex Pokorny: Sure.
Dave Dougherty: pretty basic, whatever that kind of project with humans.
With everything else going on in the organization is like a three month, okay, let's audit 20 pages.
Okay. How about, here's this prompt. Here's 20 links, 20 minutes, get a cup of coffee, come back. Now I have the responses right? And baked into the prompt are things of, okay, here's how you'd wanna rewrite.
The copy to better match the new intent, or, Nope, it's good where it is. Leave it the way it is. Like that to me is a perfect use case. The problem with that is presenting it in a way where I'm not, me or anybody else, is not saying, okay, this entire team. You don't need, we don't need you anymore.
'cause this'll just do that, right? No, this is one like research part of your job. This can now allow you to spend the three months doing more strategically important work. Instead of this like clicking around and reviewing and, oh, there's a broken image, or oh,
Alex Pokorny: that also points to leadership as well.
Dave Dougherty: Yeah.
Alex Pokorny: be a significant gap because are easy to find
because they know the task, the process, and they start tearing off little bits and pieces of it. Or
Dave Dougherty: Yep.
Alex Pokorny: a little bits and pieces to your point of you can start doing more strategic work. Oftentimes people aren't actually even set to ever get to that level.
The organization hasn't really made it clear what the strategic intent actually is. They just said, this is your job duties. Keep doing it. As people
It and keep repeating it, and maybe you can improve upon it, that'd be great too, but if you just keep doing it, that's fine. Instead of getting to that next level of saying look over our goal is beyond this little piece. Can you help us work on that?
Dave Dougherty: Right.
Alex Pokorny: I think you'd quickly find the individuals who really enjoy the minutia, and you'd find the individuals who like working more at that strategy layer, but you'd also inspire the whole group to be like, okay, we're doing this for actually for a reason, not just because my boss told me or some coworker asked me to do it.
Dave Dougherty: Yeah, and this is what's interesting to me is there's a reoccurring theme across my careers. When I was in music, you would have what I would call the scenesters, right? Where it's all right, I'm just gonna go play, and I like playing 'cause it's a good time and you get free beers and get to stay up late at night and it's just a good time.
Alex Pokorny: Sure.
Dave Dougherty: you had the jobbers, which is, I don't really care what I'm playing. It's a job. I'll come in, I'll play the part, I'll leave, whatever. Then you had the band leaders who were like, all this is a reflection of like me or our work, or the band's work or whatever. This is how we get to the next level.
This is the plan, the ro, like you had those individuals and it was pretty apparent because you're like, oh yeah, don't ask Joe to do this because he just comes in, plays his bass and leaves,
Alex Pokorny: Yeah.
Dave Dougherty: Funnily enough, that has absolutely played out in all of my quote unquote normal jobs too, right?
Alex Pokorny: see that. Yeah.
Dave Dougherty: So that's where part of my brain immediately goes to of the people who just like doing their thing,
Alex Pokorny: Yeah.
Dave Dougherty: how many of them are the jobbers that just want to come in, play the tune, leave. Coach, soccer, watch, football call.
Good.
Alex Pokorny: 75%
Dave Dougherty: Probably.
Alex Pokorny: Yeah.
Dave Dougherty: that's the huge bell curve. Then, of the people that we're talking about
Alex Pokorny: Yeah.
Dave Dougherty: where it's the, Hey, I wanna do more strategically, important work.
Alex Pokorny: Yeah,
Dave Dougherty: maybe the top 15%, 20% maybe.
Alex Pokorny: Yeah.
Finding meaning in work. Are incentivized so many different ways and finding meaning
Dave Dougherty: yeah.
Alex Pokorny: for that top group. They won't find meaning in work unless they're working at that level.
There's a big chunk of people who are absolutely finding meaning and work by, it's a place that pays my bills and allows me to pay my mortgage,
Dave Dougherty: right.
Alex Pokorny: is meaningful to them, and that's it. Or they have some, distinct, meaning that is just, it's about their title. It's about where they've reached and saying I've accomplished college, which allowed me to accomplish this role, and I find meaning in the fact that I get to continue. Be whoever I am and do this kind of work. 'cause that's self-fulfilling in itself. Yeah. You're gonna find a chunk that just, they're only going to care when they're at that strategy level because they feel like they're implementing at that level. They're contributing. They're affecting things. They're, maybe it's a self-important title, to be honest. Whatever It feels like that's a
Dave Dougherty: Have you defined this for yourself yet?
Alex Pokorny: No, I think it's rather contextual for myself, but
Dave Dougherty: Yeah. I think for me it, I prefer the strategy. I feel I like feeling that momentum which has been really interesting to me, like with these AI pilots, right? Where it's the, oh, hey, we can, I can help you out doing this, and this. Yeah. But I gotta make this PowerPoint and I just need to bring all these things together, and you can just tell their brain isn't set up to be like, take this data stream and this data stream.
Put it into one thing,
Alex Pokorny: Yeah.
Dave Dougherty: right? Even though that's absolutely what they're doing, but they're just doing it really slow and click by click. So I don't know that. Especially in the enterprise world, it's gonna take a lot longer for any of these productivity gains to be realized.
Alex Pokorny: true.
Dave Dougherty: Especially when it comes to the creative like marketing sales side. I think that the easy things are implementing the kind of machine learning, the AI analysis of the data streams you already have coming in or. There's some non-marketing use cases that I think are really interesting like materials analysis and things like that.
Dave Dougherty: But I, again, just like with marketing automation or anything like that, it's gonna be the smaller orgs who can build. Their organization and culture around these apps or these ways of working that will realize the most productivity out of these things. How's that resonating with you?
Alex Pokorny: No it's, I think, spot on. It's always a question of
where does this fit? Can we actually fit this in and does it actually make sense? And you're gonna have to ask that at every level. Smaller company also will have individuals wearing, 10 different hats.
Dave Dougherty: Okay.
Alex Pokorny: to take on the effort of pulling in an Asana tool?
Probably not. And the number of individuals involved on a project is five. So is it really that hard to coordinate with them or ping them? No, it's pretty easy to
Right? So you're gonna always have these contextual layers and situations that you're gonna have to fit. Whatever piece to it. I think it's difficult at the outside in to any organization,
Of the biggest difficulties with applying AI as an efficiency gain or actually anything as an efficiency gain, how do you actually ever see. Inside a company and inside a process, inside an organization, enough to realize there's an opportunity here, there at the next place, right? And the
Dave Dougherty: right.
Alex Pokorny: aren't necessarily always set up for doing things like that.
Limitations of AI in Content Creation
Alex Pokorny: Instead, they're more set up for these singular tasks and the res prompt and the response.
And honestly, it's a text-based res response or a code-based response. Image-based is getting there, but even
It's still a prompted response sort of thing. So it's still limited.
Where could it go anywhere? It
Dave Dougherty: right.
Alex Pokorny: takes a lot more effort to stretch it into that and place it in that and envision that even it should be going there. I
Dave Dougherty: and that's where I think as people with SEO backgrounds, a lot of our conversations with AI immediately went to programmatic SEO
Alex Pokorny: yeah.
Dave Dougherty: and the questions of should you, should, it woulda, coulda. It's like just because you can create a blog post for 12 things to do in this zip code doesn't mean you should.
And I feel like that's a lot of the, a lot of the AI situations early on, whereas, oh.
The Overwhelm of Multi-Platform Content
Dave Dougherty: We can do all the things now. We can do the LinkedIn stuff, we can do Twitter. We can repurpose one thing into 29 forms. Now we can, and you're like, do you even wanna be well known on LinkedIn? Do you even wanna be an influencer?
In that sense, if the answer's no, why? Why are you doing that?
Alex Pokorny: You have to
Another one, admin another one. Figure out who has access to it. Remove them when they leave the company. Like
Many, keep up with any kind of legal change, privacy, policy change, and everything else that you have to do with each one of those pieces.
Translations become a question, how are you actually
Support questions on LinkedIn? Like it's
Dave Dougherty: right.
Alex Pokorny: any honestly releasing any of this tech too. It's who's gonna admin this thing, who's gonna maintain it? It's more tech debt, every piece of tech that you release. So is it worth it? That's an excellent question.
Dave Dougherty: That's where from, go ahead.
Customer Communication Challenges
Alex Pokorny: oh, just a quick aside of the comment about the information being sent out also to the customers. I was just thinking about this recently, had a bad experience with the company and I was just thinking about how few people actually are involved in the communication. Approval stage at a company. Like I knew of a corporation
Dave Dougherty: Yeah.
Alex Pokorny: send even the same individual up to eight to 12 emails per day because it was coming from different groups and departments within the company. But
Dave Dougherty: That's insane.
Alex Pokorny: email address was tied to so many different, basically personas that they
Dave Dougherty: Yep.
Alex Pokorny: hit all the time constantly. And it's so funny, if you take. One step back and say, okay, what is actually our customer's full experience with our company? it's not just that we show up to trade shows, just that we have a website that people can Google and they can find it. It's so much communication out and so little oversight of does this actually, is this meaningful or is this actually high quality?
Or is this just crap that we're posting? Because we wanna say
Something every month,
what a poor metric. When you're considering like, this is the one time your customer's actually having a, an experience with you
Is that poor, like maybe less is more? I
Dave Dougherty: Yeah, absolutely. And I, I remember having a a meeting with a user, or we were watching a user experience testing thing. And like any company we're like, what's your relationship with PDFs or the website? And it turned out that they were only using the website to download the PDFs after they had already purchased because they needed it for the legal requirements in, their country or whatever.
Right.
Now my, my own personal feelings about the consumer journey I can't necessarily get into, but that to me was such a miss, where it's just like, why would that not be part of the post-purchase experience of, Hey, thanks for your order. Here's all of the documentation associated with the products that you just purchased.
Alex Pokorny: That'd be
Dave Dougherty: That would be such an easy reach out,
Alex Pokorny: Yeah.
Dave Dougherty: And then you would have the better customer experience now
AI in Content Strategy and Execution
Dave Dougherty: . I think we, we had promised a specific use case and a specific thing and this ties into what we were just talking about, where the tech debt piece of it is interesting because throughout my career in writing thing, ghost writing for people doing e-commerce stuff, newsletters, any of that stuff, you sit down and you say, okay.
I'm gonna write a blog or I'm gonna write a newsletter, and then you forget, oh yeah, I need the subject line, I need the titles. I need a meta description. I need the social posts that promote the thing that I just did. Oh, I need to have blog header to make it stand out. And oh, make sure that you have the brand image on the header so that people, if they share it, also do that.
Oh, and if they share it, you need the open graph image in addition to the, whatever. So then it's like your one thing has now become 20,
and if your definition of done is you launched a thing, you better get those 20 things done.
Alex Pokorny: Right.
Dave Dougherty: So for me. Let's just let, it's a thought experiment.
So let's not get into whether or not you should be blogging in 2026. Let's just assume that you have decided that is going to be strategically important for you.
Alex Pokorny: Elephant in the room. We're that guy.
Dave Dougherty: Just completely ignore it. Yeah. Yeah.
Alex Pokorny: Just sit
Dave Dougherty: Yeah.
Alex Pokorny: Don't worry about that.
Dave Dougherty: Yep, that'll be the next episode. What should you be doing? The the, where AI has helped me with this scenario is not only, the strategy side of things are coming up with the ideas but also some outlining. I will do all of my own writing because I enjoy that.
So that's my own caveat, which is a big one. Do you enjoy doing the work or not? For example, with this podcast, going through that notebook, lm, that we've kept going on every single episode, that we have to be able to go to it and say, Hey, we want to tighten our messaging around, like what our main content pillars are.
Help me identify the topics we keep coming back to across 70 episodes,
Alex Pokorny: You're
Dave Dougherty: right?
Alex Pokorny: transcripts. Yeah.
Dave Dougherty: Yeah, so then it does that analysis, which honestly that could have been a, business school capstone project in terms of the amount of work reading through 70 transcripts. But then to be able to then immediately have four content pillars.
With typical subtopics that come up underneath those four pillars was a huge help because now, okay, here's what we favor. We might want to add, a couple extra things just for some extra flavor, more in depth discussion. That's been massively helpful because now I'm not doing that huge research thing and it.
And it doesn't have to be very correct. It just needs to be directionally correct. Because then it comes into the actual execution of it too.
Alex Pokorny: That's a very helpful caveat.
Dave Dougherty: Yeah.
AI's Limitations in Understanding Context
Alex Pokorny: There's precision is always a difficulty with any AI system because you have to trust the output entirely,
Dave Dougherty: Right?
Alex Pokorny: I don't think you can. So there's always a point there where either you are double checking the results, proving it out, or doing it all over again yourself.
Dave Dougherty: Right.
Alex Pokorny: can find those scenarios where you're looking for that directional thing. Also, in your example, you're talking about 70 episodes worth of transcripts. That's a ton of content to read through,
The benefit with AI in that situation also was that you're having to do heavy load kind of analysis and lifting
Dave Dougherty: Right.
Alex Pokorny: likely would not do, and especially at any other organization, they wouldn't do it. So this was an an unrealized opportunity
and it was a slightly unbiased third party. I giving you feedback on the kind of current common topics versus someone likely going to the most recent couple and saying we talked about this recently, and I
talk about a lot, making pretty broad based assumptions versus something that would be much more precise and say following 15 episodes for about this, the following 14 were about this.
Dave Dougherty: Right.
Alex Pokorny: That's helpful. No, that, that's a good example and I think people can do that and expand that also against blog posts. Content that they have, data that they have reports analysis, might even be multiple PowerPoints and say what's, common across here? What common problems do we keep talking about?
That's helpful.
Dave Dougherty: I think the other thing too, at least within that research and development side of things is if you take away. User data from things like SimilarWeb or Spark Toro or old keyword research kind of things to then be like, here's my audience. This is what I know generally from these sources about said audience.
Alex Pokorny: Yep.
Dave Dougherty: Help me understand how I can best. Speak to target with, refine what I'm already doing. For me, it's always refining Yeah. Or punch holes in this theory for me where, what I, what might I be blind to with the way that I'm, discussing this. Those have been massively helpful.
Now granted, those are all things that I love to do.
Alex Pokorny: Yeah.
Dave Dougherty: Of course I'm gonna go, I'm gonna go that direction with it. On the actual writing of the post. Creating headers that are optimized, towards the intents or writing a meta description that, Google's just gonna rewrite, but, you get
Alex Pokorny: Gotta
Dave Dougherty: yelled at by the, you have 14 errors 'cause you don't have a meta description.
So okay. Those all go in there. Getting all that done. Those are massively useful. And also coming up with, just give me five ideas for this title. Because I come up with one and that's the working title, but then I know I'm anchored on that title ' cause it's the first one I came up with.
So just seeing what some other things could do. I don't know that I've ever taken a title verbatim. From the ai, it's always, oh, I like parts of this one. I like parts of that one. But then if I put them together, it doesn't make sense. So let me rewrite it. There's always some editing and rewriting
Alex Pokorny: Yeah,
Dave Dougherty: to the point of the SEO automation stuff that we talked about.
The Human Touch in AI-Generated Content
Dave Dougherty: That's the stuff that I've always been worried about with ai, where it's the prey and spray.
Alex Pokorny: The programmatic
Dave Dougherty: Idea.
Alex Pokorny: throwing out
Dave Dougherty: Yeah. Where it's just okay, if you're a narcissist, if you're a narcissist and you don't care, or you're an anarchist and you don't care and you're just gonna create a bunch of content because it'll benefit you, then yeah, this'll be really powerful for you to do a lot of crappy stuff.
Personally, that's not me. I don't wanna be known for that. I don't even want any inkling. Out there that's what I would be doing.
Alex Pokorny: Right.
Dave Dougherty: So yeah. Anyway, back to the blog post. One thing that's also really been beneficial is just. During the editing process, right? Like during the creative writing degree that I had, one of my favorite things was you would always have to sit down and it would be your turn to get feedback from everybody in the room.
So you got an entire classrooms worth of bias and perception and internalization of whatever it is you created.
Alex Pokorny: Yeah.
Dave Dougherty: Then they tell it back to you. And there are a lot of times where there were a bunch of perspectives and I'm like, oh, I don't have that lived experience. So I never would've thought of that.
That's interesting, right? Or Oh, dang, I didn't mean to piss you off. I'm sorry.
Alex Pokorny: It's good to hear
Publishes. Yeah.
Dave Dougherty: yeah. I need to rethink this. There, there's a lot of that I've been playing around with too between the ais and honestly, that's one of the reasons I've been leaning more towards Claude recently because it's just better at that particular use case than Gemini or Chate.
Yeah, I think that's my workflow through writing. I think that makes a lot of sense just 'cause I do a lot of that. But I know you have like custom GBTs for like landing pages and things like that. Would you add anything or change anything for blog posts or any kind of webpage writing?
Alex Pokorny: Yeah, what you're saying is spot on. It's true that there's the. For those who are scared of writing, it is the blank page killer. It, you have a blank page, you're using it to give you something to react to,
but it's never good enough to be published. If you make a custom GPT, or gem from Gemini or whatever, any other system
Dave Dougherty: Right.
Alex Pokorny: you can preload it with, your tone, prior example articles, which really helps,
Dave Dougherty: Yeah.
Alex Pokorny: brand standards. elements that you always want to have in there. For instance, like your OpenGraph takes, if you ask to write a blog post, it'll write a blog post, but it won't give you the OpenGraph takes until you ask for the OpenGraph takes. But your GPT could just have that tossed in always and be added in. Those kind of pieces are important to get that initial copy and piece down. I think one abuse of AI tools is ideation
Dave Dougherty: Yeah.
Alex Pokorny: these are. Still, to be honest, mad Libs kind of tools
bits of logic and reasoning tied into them. So it's the most probable phrase that it will continue with.
Dave Dougherty: Right.
Alex Pokorny: not necessarily smart and it's not judging its own work. You can ask it about its own work and how did it do that? It doesn't actually know. Instead, it comes out its training data to come up with the best possible reply to then reply to you.
Can go on infinitely. understand a whole lot of what's going on.
It can have some prompt memory, or recent reply memory in the context window and the like, it's very simplistic. So where I've seen it gone wrong is when people have been querying these tools for, what should I write about?
Dave Dougherty: Right.
Alex Pokorny: do that without giving it the context of saying, these are our prior articles, this is the audience we're going after, this is the type of products we have, the type of service that I'm trying to push, or the opinion that I wanna make. Any of those kind of points, and you're
Dave Dougherty: Right.
Alex Pokorny: here's my website. What's something I should write about? It's gonna give you a dozen ideas that are probably not worth actually doing.
I've done that a few times even with our podcast thinking like, maybe, we've got a lot of transcripts out there.
What should enterprising minds talk about next? It'll give me like a dozen ideas. And they're terrible.
Dave Dougherty: Yeah.
Alex Pokorny: It's really bad. Like these are things that yes, they were bore to death an audience, correct.
Dave Dougherty: Right.
Alex Pokorny: It has something to do with AI and marketing and who knows what else. I
Some of them are just, I can ask 'em to be more professional, more fun, more whatever, and it still can't ever seem to nail a topic like
Dave Dougherty: right.
Alex Pokorny: surprisingly bad at that.
Dave Dougherty: YouTube has that too. And I don't know if you've jumped in and tried that. Suggested Yeah.
Alex Pokorny: What you should make a thing about next. Yeah.
Dave Dougherty: And it's always some sort of inflammatory thing or,
Alex Pokorny: thanks
Dave Dougherty: five product reviews for this, blah, blah, blah. And I'm like, no, that doesn't feel like us.
Alex Pokorny: No. It never does. And is the
Dave Dougherty: yeah.
Alex Pokorny: is you're always talking to a piece of software. It is a bot. It's a script that's running. And that's what you're responding to and you're getting back and it's not a human being who would
So many other questions and try to get a good understanding of who you are, the context of
On, and then give you advice.
It's that friend who you randomly text for a piece of advice and they give you something that you know, with no background, they're not gonna give you great advice.
Dave Dougherty: right.
Alex Pokorny: an individual who you've been talking to, for months and. Your significant other, your spouse, your brother, your sister, your close friend, you name it. And then you ask them for a piece of advice that is
To you off of all those experiences and their own experiences bundled up together,
Dave Dougherty: right.
Alex Pokorny: chat, CBD does not have experiences. It has a training set of data and most probable phrasing, and that's it. And even if
Dave Dougherty: That's why I don't.
Alex Pokorny: you can, you never will get a good response from it
Dave Dougherty: Yeah, and that's why I don't understand why people would get into this whole like AI relationship thing because like how do you know someone loves you if they're not yelling at you?
Alex Pokorny: That, that's a separate therapy question. Oh, no.
Dave Dougherty: Yeah.
Alex Pokorny: base level. And that's the funny thing is we mentioned this once before. Mooch has a nice article on it too, about how many romance novels were, have been fed into training data because it's available in cheap, text. Wiki, Wikipedia runs out at a certain point, but if you really want a whole lot of cheap data,
Gives it some interesting personality but if you're developing a book, basic character, yeah, it can do that.
But
Human being. The depth is never there.
Dave Dougherty: right.
Alex Pokorny: And that's always the problem with an AI response is the depth isn't there. You have to add your own personal touch to it, and then it becomes interesting. It becomes worthwhile consuming it versus, another listicle of 12 things to do with ai.
Thanks. 11 prompts to save my life. Yeah. Come on now.
Dave Dougherty: Yeah. Interesting. I feel like I've been saying this line a lot lately, but clearly you can tell we're still wrestling with these ideas. I think everybody is to some extent. As with any, yeah, and it's changing real quick. Like just from the beginning of the year till now oh my God, the amount of things that have been announced and have done, it's it's a lot.
It's a lot. Everybody, please let us know what you think. Does this resonate with you? Have you seen any kind of productivity gains? Are you feel like you're able to do more with ai or are you just, sliding through? I don't know. It'll be interesting to, it'll be interesting to hear from people.
Subscribe, share help everybody get in on the conversation. We love the feedback, when you're giving it to us. Keep that going. Keep trying ai, keep trying marketing. 'cause lord knows we gotta fight for it. Cute. Beastie Boys. And we'll see you in the next episode two weeks from now.
Take care.
Alex Pokorny: Cheers.
