Ep 71: The Fallacy of Automating Everything
Watch the YouTube video version above or listen to the podcast below!
Episode Summary
Dave and Alex break down why “automate everything” is mostly hype: AI runs into real limits—quality, context, maintenance, and the messy reality of organizations. In SEO and marketing, it can scale output, but often scales low-quality work unless humans verify and refine it.
They discuss the “double-check tax,” sharing examples where AI created extra work through errors or formatting struggles, plus the risks of junior teams trusting confident-but-wrong outputs. The upside: AI is great for low-stakes, reversible tasks (like meta descriptions), brainstorming, and beating the blank page—especially when you apply a “two-door vs. one-door” decision lens. They close with enterprise reality: automation won’t fix alignment, approvals, politics, or shifting priorities—it’s a tool, not autopilot.
Ep 71: The Fallacy of Automating Everything Podcast and Video Transcript
Dave Dougherty: Hello, welcome to the latest episode of Enterprising Minds. You're stuck in with Dave and Alex. Today we're going to talk about the fallacy of Automate Everything and why that's a fallacy and, we just keep running into brick walls whenever we try to, be tech forward and, optimize things.
And yeah. So Alex, you seem pretty animated in the pre-recording call about a certain situation. So I'll kick it over to you to
Alex Pokorny: Yeah.
Dave Dougherty: talk about your example.
Concerns About AI and Automation
Alex Pokorny: I also have a pet theory that chat GBT is getting dumb. But that's a, that might be a side point here. We'll get to that in a moment. Small
Dave Dougherty: You mean
Alex Pokorny: I have concern.
Dave Dougherty: based off of the internet seems to be an idiot.
What?
Alex Pokorny: I know. Something that's starting to release advertising and suddenly need money is potentially pulling back on their model and making it dumber because it'd be cheaper. If I had a little magic eight ball, it'd say like
point to Yes. Yeah. Yeah.
Challenges with Automation in SEO
Alex Pokorny: So they automate everything. So this keeps showing up on LinkedIn especially if you follow any of the SEO news, but honestly, digital marketing and nearly everything else, it's oh, you can use AI for that.
Dave Dougherty: right.
Alex Pokorny: slap, this big giant bandaid. It used to be the big red button. You just slap the easy button and your website gets launched. Dear God, I seriously still want a button that somehow runs a script or something to actually launch a site. Because it's so funny when you talk to management directors, you name it, and they're like, launch it.
And it's you realize there's no button that does this. It's just a ton of steps.
Dave Dougherty: Right.
Alex Pokorny: Eventually it just there, there is no oh man, look,
Dave Dougherty: Yeah. Or my favorite for years was the just SEO it like,
Alex Pokorny: Sprinkle that fairy dust on it,
Dave Dougherty: yeah, no,
You, your simple little, just go do it. Is a six week project.
Alex Pokorny: Do you have any idea what you're asking for?
Dave Dougherty: Yeah.
Alex Pokorny: Yeah. So automate everything. The, I've run into some issues with that…There we go. One thing I was dealing with was the automation is trying to expand upon how much data can you really put inside any of these systems, and then where do you start hitting those limits? And you hit those limits super fast. And that's the one frustrating thing
Dave Dougherty: Yeah.
Alex Pokorny: the whole thing is that all of these ideas of, oh, we can automate the heck out of SEO, you can automate poor quality crap at scale.
You really can't do good work at scale and these tools help maybe expand you and push you on an extra 10% extra 20%, but they are not doing 200%. That's the thing that like, seems to be very lost in any of these times where you get these articles, especially with LinkedIn one search are like so boastful of this idea that you can somehow, do this massive increase of, productivity. Explosion of content, websites, apps, who knows, whatever.
Dave Dougherty: Right.
Alex Pokorny: The quality disappears really fast. Still one of these things of it's, maybe it's just a scale problem. Maybe that's just it, where we start, keep running into that. But I've been having conversations recently around AI usage and we keep getting this idea of what about entry level workers, entering into the market, they're using AI tools and the AI tools may be misleading them or giving them poor results.
How would they know any better? And that, that, that kind of is it like you can use these tools great, but you still have to be smarter than them. You still have to be able to judge them,
Every time they start making mistakes, you have to be there to correct it and keep going, and back and forth. If you have back and forth, you're not scaling. And that, that it starts to fall down right there. The
Dave Dougherty: right.
Alex Pokorny: of it I'll hit real quick is basically notebook LM can take a lot of content into it and then you can use Gemini to search it and you can use Notebook, lm. There's other ways to also do that.
Artifacts with Claude. Other ways to basically upload data, then ask it questions about the data.
Dave Dougherty: Right.
Alex Pokorny: to that all over the place, and the limits don't hit. They're not like, oh man, I'll never hit this sort of thing. And oh, only if I'm doing some crazy, wizard Re that somebody gonna never run into this problem.
It's no, literally you're gonna hit that with a small site. Depending on the number of pages and things you're trying to handle, no, you're gonna hit it right away. And if you're trying to support a lot of different activities, which commonly is SEO or digital marketing, we are all over the place.
We're
Dave Dougherty: Right.
Alex Pokorny: Just doing one thing for one reason and never gonna do anything else. Instead it's, we're doing this part for right now because it makes sense. And then as we move on to the next one, or the fire drills or the priority shifts, the quarterly changes. The boss who suddenly says something and then suddenly now I guess my job changed again this week. That kind of stuff like that happens all the time.
So I don't really think that the tools are quite at a place where they're supporting this, but also this idea, the fundamental idea of you can automate everything and then just walk away. Is the exact same fallacy that we ran against with the four hour work week, the outsource everything in your life and have someone else do it and somehow you'll still make profit, but that doesn't work.
All these things take and work. The idea of building out a ton of apps, great, you can build out apps, but what happens when iOS updates? I
Dave Dougherty: Right.
Alex Pokorny: have to go back and. Maintain all those apps now, and now you just built up a bunch of tech debt and what happens when competitors enter the marketplace?
All these things, they're dynamic. Like hard to have a business last a hundred years because stuff changes.
it's the same idea of you do that on any scale, you're gonna struggle at certain point when stuff changes and starts to break down. I don't know, I just, the concept of just that you can automate SEO or. Paid search to a pretty large degree too. PME is great, but there's still limitations there as well. Almost these systems, limitations all over the place. So this unlimited, this fusion energy that we somehow found with ChatGPT is I False statement.
Dave Dougherty: Or it's oh, we're headed towards the minority report. It's no. Do you understand the systems that would have to be in place for that? And like infinite levels of processing,
Alex Pokorny: Yeah. Yeah. There, there is. Yeah. I've hit on this before in prior episodes, like there's a very. Physical aspect of all these internet related things.
Are servers, there is bandwidth, there are fiber cables that do have limitations to them. There is a latency problem.
There, there's processing power issues and yes, those hard drives wear out and you have to have somebody there at the server taking them out, replacing them, fixing them, literally slotting in a new one and
Dave Dougherty: right.
Alex Pokorny: running again.
And that's maybe that's a robot someday, but. You're still paying power bills 'cause it's still plugged in. I mean there's there's all these
of just like, how does this actually function? There's very physical aspects of all this stuff behind your monitor and that your monitor's displaying that you run into.
And then you start realizing, oh yeah, there's a limit here. There's a limit here too.
Dave Dougherty: And that's why when AI first came out, I was like, man. First, once you get past the excitement of like your mind being blown that the computer's doing this particular thing, right?
Alex Pokorny: Yeah.
Dave Dougherty: Then all of a sudden you find yourself going, okay, but wait, I remember being through the marketing automation.
Have I ever seen anybody set up marketing automation properly? No.
Alex Pokorny: no.
Dave Dougherty: No. And i've been going through old episodes, so I was reminded of this, with one of the first ones that where we were talking, I said, honestly, it reminds me of when like TV dinners or frozen dinners came out. For some people that makes sense, right?
It You don't care about food for whatever reason.
Alex Pokorny: You're just trying to eat.
Dave Dougherty: You're just, yeah, you're just trying to get something in and you pop it in the microwave and it's fine.
Alex Pokorny: exactly.
Dave Dougherty: is it advisable, for doctors or general
Alex Pokorny: No
Dave Dougherty: life enjoyment? No.
Alex Pokorny: enough sodium to kill a squirrel. Yep. Should you be eating this? Probably not.
Dave Dougherty: No. No.
Alex Pokorny: But it's fine. You can still do it. It's just, yeah, that's the that's the top is it's fine.
I keep thinking like the metaphor and maybe the, like a Super Bowl bouncy ball. I don't know, I gotta find like a better metaphor than that.
But there's this idea of the expectations in the hype and this
Dave Dougherty: Yeah.
Alex Pokorny: goes off into the air eventually it like, smacks the ground and each time you're like.
Frustrations with AI Tools
Alex Pokorny: yeah, that was really cool, but literally I asked Chatt the other day what time it was and it Like even when I tried to search the internet, it still couldn't tell me the time it's actually still worth something trying.
I was amazed because it's been going on for weeks and it. It still can't tell the time, like even going out to a live website, which has the time on it, it can't give me the information, which is weird. You think you'd be able to like, there's
that would be a smack on the ground and then it goes off in the air again of it also is able to write this really long research report, which is great,
Dave Dougherty: What is time and do we actually need it?
Alex Pokorny: Oh, there we go. Research report on that. No more of like market analysis kind of stuff. And it's okay those numbers look good. And then you start looking at them and you're like, some of those were good, some of those were not good. And then the ball smacks the ground again. It's like this and it keeps bouncing until it pedals out and just stops and rolls along the ground because you're starting to realize the reality of the situation versus all the hype and the hype keeps bumping it up again and trying to get it another bounce.
And we keep going through these bounces, but I think we're starting to get to the point where the ball's gonna roll against the ground. 'cause you realize like Cool word processing software, it's so great. Limits.
Dave Dougherty: And you and I had talked about how it was like a blank page killer, and how,
Alex Pokorny: Yeah.
Dave Dougherty: yeah. So this is a horribly mundane example,
Alex Pokorny: do Tell Dave, do tell,
Dave Dougherty: Yeah. Yeah. My wife bought new carpet
Alex Pokorny: Okay.
Dave Dougherty: and because of that I have to clean my office and move everything out of it. But I am a creative individual, so I have.
I have piles,
Alex Pokorny: Fair enough.
Dave Dougherty: so I'm going through the said piles, right? Of all the different post-it notes of all the different ideas and random thoughts and whatever, or sometimes if I'm in a big. Project, I'll write a bunch of notes of oh, I discovered this. If you do that, watch out for this. Thinking, okay, I'll turn around and I'll write that blog post sometime and then I forget to, right?
So I have these stacks of ideas and so today I went through and, there was one that was like three little, it was like mini legal pad pages. So like this size,
Alex Pokorny: This
Dave Dougherty: right?
Alex Pokorny: Sure.
Dave Dougherty: Yeah.
Alex Pokorny: page. Yep.
Dave Dougherty: So I had three of those and it was just a bunch of random, a random notes, and I read through it. I'm like, oh, I vaguely remember that project, but yeah let's do it.
So I just put, I put it into Gemini and I say, Hey. I have a bunch of these random notes from a project I was doing. Let help me turn it into a blog post or organize it for me, right? Don't write the whole thing for me, but organize it and
Alex Pokorny: Yeah,
Dave Dougherty: Gimme some feedback, right? And it's just man, was that frustrating?
Because what it came back with was just like random stuff. And then some of the SEO things that it was saying was not true. So then, of course I'm like calling it out on it, and then eventually it was just the, I'm gonna, I'm gonna rewrite, I'm gonna rewrite this.
Alex Pokorny: It's become frustrating enough that I now care enough just to redo it.
Dave Dougherty: Yeah. It's okay, cool. I went from three pages to, okay, now I have the outline. At least like I have the section te, headers, but I went and just rewrote all of it.
Alex Pokorny: Yeah.
Yeah.
Dave Dougherty: so on the one hand, cool, I have three fewer pages in my office,
Alex Pokorny: Sure.
Dave Dougherty: but to your point of the, bouncing it down I have spent an enormous amount of time building out custom GPTs. Automating absolutely where I can. So here's the tone of voice, here's my preferred style. Only cite these sources. Don't ever use a forum or a chat, thing as a source.
Any of that kind of stuff. So that I don't have to prompt it all the time. But even with that. It still came back with this garbage and it's just man, okay. But part of it too is I'm picky 'cause I am have a creative writing degree and I like writing. And so I have a particular idea of how I want to come across when I write.
Would somebody else have that level of nitpickiness? Probably not. the first draft would've been me. Okay. But if you go to claim that you have any kind of search experience the real SEOs would be able to call you out over a couple of turns of phrases where it's yeah,
Alex Pokorny: Yeah.
Dave Dougherty: Especially it was about international SEO, which is even fewer SEOs, just no, I want to get this one right.
Alex Pokorny: Yeah.
Dave Dougherty: I want to get this one right.
Alex Pokorny: Yeah.In my mind kind of a basic idea for auditing some analytics and kind of like some standard stuff for a couple different platforms.
Dave Dougherty: Right.
Alex Pokorny: And I had thrown into Chat chief, I was like, give me like a generic audit that I can react to of these three platforms. So I can basically just copy and paste the bullet notes of which ones I like and then expand from there.
Dave Dougherty: Yep.
Alex Pokorny: Seemed like I could set it off to do that while I do something else. I go back and check on it all the bullet points start with lowercase and they're all, like, weird. Now I can't copy and paste it. I was like, okay, just, can you capital letters at the beginning of each bullet point please.
Dave Dougherty: Right.
Alex Pokorny: So it capitalized everything. I was like, no for, ah, so here's an example.
Bullet points like this.
Dave Dougherty: Right.
Alex Pokorny: bullet point, then a capital letter. Then lower case sentence case. Please
Capitalization. So it gave me the entire paragraphs above, as well as the bullet points below single word started with a capital letter. So I kept
Dave Dougherty: Well,
Alex Pokorny: I was like, stop.
Give me AP styling. Try
Dave Dougherty: right,
Alex Pokorny: It just repeated it again and I was like, know what? Forget it. And I was like, man, at this point I've wasted more time than it would've taken me to write out the silly bullet points because I even gave it now examples of proper bullet points. And like AP styling, I even checked whether or not, that's proper styling, which of course, yes it is.
Start with cap letter. I mean it's like dumb. Things like that. I was like, great.
Like 10 minutes on a five minute task,
Dave Dougherty: Yeah. I was walking. Around during, so during the workday I will take a break and I'll just go for a walk,
Alex Pokorny: Sure.
Dave Dougherty: right? Because who knows what happened or,
Alex Pokorny: Yeah.
Dave Dougherty: Clear my head or
Alex Pokorny: Smart.
Dave Dougherty: Whatever. And I was thinking about this where it's okay, we're outsourcing so much stuff to ai, or at least those of us who are using it, frequently.
Okay. I feel like we don't give ourselves enough credit for the amount of crap that is in our head that we just don't have to think about. For example, your AP formatting, right?
Alex Pokorny: Yeah,
Dave Dougherty: For the last 15 years. Been writing in AP formatting, I don't have to think about it. I just do it. When I do an anchor text, I know exactly how I want that formatted.
I know exactly how, you know what phrases I'm gonna highlight for that. It's just done.
So then to have to sit down and go, okay, what do I know to be true? What do I prefer? Makes it really long, right? Because then you're like, oh crap. All this is in my head. And I don't usually think about it
Alex Pokorny: Yeah.
You're
Dave Dougherty: now granted I'm not in, I'm not smart enough to talk about the nature of intelligence, but it is at least, interesting to think of just all the stuff that's there that you don't even think about while you're doing your job because you have the experience or it's just innate in your habits.
'cause I know a lot of the coworkers that I talk to, the thing I'm getting frustrated. About us. I'm like, Hey, have you tried AI for this particular thing? I feel like you could do that faster that way, and then just go, no. And everything that they come back with, what they've tried for is so tactical.
So tactical that it misses 90% of the value that I think, AI could give you. Because it's I just want to create a PowerPoint. You should probably do that then, because it's gonna be formatted correctly. It's gonna, you're gonna be presenting, so it should represent you like, but coming up with the outline for the presentation.
Yeah, sure. That's totally an AI thing,
Alex Pokorny: sure.
Given some feedback on some ideas summarizing things down, making your sentences more, memorable,
You
Dave Dougherty: Yeah.
Alex Pokorny: Yeah.
Using AI for Creative and Personal Projects
Dave Dougherty: gonna say the biggest benefit I've gotten out of AI has been with my creative writing stuff, with the strategy stuff, like all the things that I would have gone to friends to say, Hey, give me your thoughts on this. And then all of 'em would be like, Ugh, do I have to, it's it's Friday, man.
I want to.
Alex Pokorny: See that's how you deal with the loneliness epidemic. Like you just talk to the bots instead. Because it's funny because I, it was, oh, probably over a year ago I talked about this, that I probably helped my marriage basically using AI to teach me about interior design and asked it a million questions about things.
'cause I knew nothing. I'm
Dave Dougherty: All right.
Alex Pokorny: Renovate a ba basement and it's big enough space with enough constraints and all the rest of those I don't know anything like, so
Dave Dougherty: Yeah.
Alex Pokorny: a million questions of this or that or what other options are there and everything else in between.
And it goes on for weeks that I'm like going back and forth on
Dave Dougherty: Right.
Alex Pokorny: ideas, and stuff like that, My wife who didn't really have any interest in the project could be bored to tears me talk about different ideas that I have for how we can make the space better, et cetera.
And it's oh my gosh. It is useful as a sounding board. It is useful for some of those basic things, but yeah, it's kind limits across all of it.
There's also I mean there's that psychopathic aspect of it where it always is trying to agree with you and say that your idea is the greatest idea ever and all that kind of crap.
Dave Dougherty: Right.
Alex Pokorny: which is potentially an output of human trainers up voting those responses more than others because they like that.
Weird on humanity, but it also creates inherent problems with it too, because you'd have friends who would be potentially more critical or have a larger view who would say yeah, that's a great idea, but keep in mind your boss is like this.
And it's yeah, that's true. This project would never get. The lowest, like I could spend hours on this thing and he's just gonna say no to it, which would be helpful instead of spending hours on it, which AI could help you spend hours on it you bring up those other pieces. First context, and it never has enough context.
Dave Dougherty: Then what was the
AI Decision-Making: One Door vs. Two Door Problems
Dave Dougherty: I think Jeff Bezos had the idea of this, is this a one door or two door problem? And this was on the. Marketing AI Institute podcast.
Alex Pokorny: Okay.
Dave Dougherty: stuck with me this week where it's how do you use, how do you use ai, for things? And it's it seems like AI is perfect for those two door problems.
And the two door problems, if I'm remembering correctly, are like, if we leave through that door, we can come back in and choose another one. The one door problem is you leave through the door, you make that decision, and there's no going back. So if it's a one door thing and you have to be right, don't use ai.
Alex Pokorny: Yeah,
Dave Dougherty: if you can put it out you could do a blog post and then, oh, you know what, there's some errors. Let's fix that. Put that back out there. That's more of that two door problem. But yeah. Does that resonate with you at all?
Alex Pokorny: I was thinking about that.
AI in Legal and Marketing Contexts
Alex Pokorny: it was maybe now two years back, but it was a lawyer who had put together some case documentation with AI and
Dave Dougherty: Yeah.
Alex Pokorny: it made up that didn't exist. And it went before the judge. And then the judge realized that these are false. False judgements and false comparisons.
So that was the thing that definitely you can't take back and just, submit another version of it. And yeah, in that kind of case, you really shouldn't be using it for that.
Using AI Tools for Content Creation
Alex Pokorny: is, yeah, I think that's a good point because it goes back to the idea of the blank page killer or even some of the automation stuff I've been doing recently.
So Screaming Frog has, it's a crawler tool set on a website. It crawls through all the links, grabs data as it does that, like page titles can even grab the whole body copy, whatever. You can also include different API connections to it. So Google search console data, Google Analytics data or AI systems. And it'll basically run whatever prompts you have put into it page that it crawls through. So you could have a selection page that just let it go wild on the website. It's expensive and it's really slow, unlike those other services which are free and a lot faster. So it. It could take days instead of hours, and it might cost you 60, 80 bucks for 5,000 pages depending on what kind of, how many prompts you're doing.
And the complexity though of them, which is a weird discretionary income that or discretionary budget that you now need for crawling tools. But. The data that it runs is great data, or depending on how well you crafted those prompts about each one of these pages, especially if you're trying to do something at scale find a page that doesn't mention this particular product line or doesn't have a link to this page, or does it have some other kind of quality to it?
You could put in that as a prompt and now you could do 5,000 pages. You absolutely could, it would be okay data, but I would never consider it a hundred percent confident. I would still not be confident, and I would, even if it was something simple like that, of does this phrase exist on this page? Which it should be able to match things. There's other ways to do it too, that'd be more cheap and verifiable, but I would still give it a 70%, 80% confidence level that I got all of them on. That was 5,000 pages. With that data, I would consider it perfect. And if it, I asked something else give me SEO recommendations per page. would take a look at those recommendations, but I would probably deny most of them and I would take only a small portion of them forward, but I wouldn't ever push it through the door in your analogy, whatever. It just came back. I wouldn't do
Dave Dougherty: Do.
Alex Pokorny: I would have to review it first.
Dave Dougherty: Here's another question for you. Would you. Do you find it more annoying to create something new or to double check somebody else's work?
Alex Pokorny: Depends on what it is, like
Challenges and Trust Issues with AI
Dave Dougherty: that's my experience with ai. You are either creating something new
Alex Pokorny: Yeah.
Dave Dougherty: or you're spending all of your time double checking what it's done.
Alex Pokorny: That's a really good point. Yeah, that is. Yeah, that is it. Yeah. Either it is doing something absolutely brand new or you're just going back and forth like crazy, trying to get the stupid thing to do what you're actually hoping do.
Yeah. Huh.
Dave Dougherty: So what's more annoying?
Alex Pokorny: I'm trying to flip this to a positive too, of what would I rather look forward to working on it? It's yeah, I can see why. I also have a hesitation sometimes of opening up an AI tool versus not
of that back and forth aspect that it seems almost inevitable with almost every question.
Sometimes it's a very helpful starting point.
Dave Dougherty: Right.
Alex Pokorny: But then I still always have trust issues if it's brand new,
I dunno, I was asking Claude the other day of pizza dough recipe. I need a pizza dough recipe for two eight inch pizzas and one 12 inch, and I've got an hour and a half, so it can't be long. Proof of ferment kind of process. Can't be overnight or anything. What's a pizza dough recipe that fits all of those qualifications? Very hard thing to Google search
because I have this, output quantity that needs to be nailed.
Dave Dougherty: Right.
Alex Pokorny: I have constraints on top of it. And honestly, Google's not really good with negativing out or, bullying search was, you put minuses in of things and it would, you cut. Cut the stuff outta your results and optimize your results a little bit better.
Kind of advanced searching. It's not great at that. And recipes are really poor at that. In fact it mismatches all the time with recipes. So it gave me a recipe back and it was good, but I had questions. And that was the problem was that. This like trust would verify thing, just like hits over and over again with whenever you get an output. So it was something, I had nothing, it gave me something. So this was a brand new case,
But my trust level wasn't that high and it worked out okay.
It's pretty forgiving recipe to begin with, to be honest.
Dave Dougherty: right.
Alex Pokorny: but it wasn't great either.
Dave Dougherty: 'Cause good things like good food, take time.
Alex Pokorny: Yeah. Yeah. Plus, I don't know, just some of the things that I did were like a little odd
during the process. I've done a lot of pizza doughs recently. So it's not like I haven't, done once in the past. I just have never really saved a recipe that I really enjoyed, and I don't know, it was okay, but
The problem.
What about you?
Dave Dougherty: I find it really obnoxious having to double check work.
Alex Pokorny: Yeah.
Dave Dougherty: I would much rather create things, hence my enjoyment of music and writing and cooking.
Alex Pokorny: Instead of editing. Yeah.
Dave Dougherty: 'Cause editing is where the real work happens,
Alex Pokorny: Yeah.
Dave Dougherty: right?
That's where I like walking through things like, help me think through this. What are some options for X, Y, and Z, right?
I have this general idea, should I pursue it, should I not? Or go do a basic search, like deep research thing. Help me be directionally correct.
Alex Pokorny: sure.
Dave Dougherty: I don't, I don't need to be a thousand percent correct for this thought process, but, let's be, on the way there, so that's where I've used it a lot.
The. The funny thing is the biggest use for the Chrome browser Gemini that I've found
Alex Pokorny: Yeah.
Dave Dougherty: as I've been like, I don't really ever use that.
Alex Pokorny: Same. I have, I saw it. I was like, oh, that's so cool. And then I have yet to
Dave Dougherty: Not used it. Yeah. So I started doing some content migration between websites.
Alex Pokorny: Yeah,
Dave Dougherty: and like a lot of things you're focused on creating the content and getting it out on time so you know, certain things fall through the cracks. I find myself needing a bunch of meta descriptions,
Alex Pokorny: Sure.
Dave Dougherty: so I share the tab with it.
It can read the article that's on the thing, and I just type in meta description and it pumps out three versions and I copy paste it. Done. It's been great.
Alex Pokorny: Yeah.
Dave Dougherty: Could I have now to your point, in your earlier example, can I do that via Screaming Frog all at once across the entire site? Yeah. Is it worth me stopping what I'm doing
Alex Pokorny: No.
Dave Dougherty: to then, go do that at scale?
No, I can just do it as I'm doing it as part of the job and just get it done. Yeah, so it's it's those small little things where it's okay, no, it's just I wasn't anticipating this. I don't have it. It's a throwaway thing anyway, we know Google's gonna redo it anyway. But then at the very least,
Alex Pokorny: Man, is a perfect example of usage,
Dave Dougherty: yeah, where it's just, it is kind, it doesn't really have any kind of big impact, but at least your reporting won't yell at you for having, 12,000 errors. Right?
And
Alex Pokorny: don't wanna be yelled at.
Dave Dougherty: exactly
Yeah. Now will I win any awards for leveraging? Gemini in that way. No. No. Do I care about that? No,
Alex Pokorny: yeah,
Dave Dougherty: but it's been a really useful thing, right? Because then you know, it's God, I just got done writing 1200 words. Now I have to come up with a perfect two sentence summary.
Alex Pokorny: That less than 30% of. People who ever see that surface result will ever even see
Dave Dougherty: right.
Alex Pokorny: overwrites many anyways.
Dave Dougherty: Yeah. Yeah. So it's that to me is that's the perfect thing. 'cause it's a throwaway thing, but then it's, it still has to be done and
Alex Pokorny: You.
Dave Dougherty: it's nice to have in case, something else comes up. But yeah. Yeah, I think I've been leveraging, let's switch topics a little bit.
Go back to what you were talking about with chats getting dumber.
Alex Pokorny: Yeah.
Dave Dougherty: So I have been finding, so I've been working on some creative writing that I've been doing and has been a main goal of mine for 2025, and I got 85% of the way there. So I'm carrying it into 26 to, finish stuff. And every time that I would work on this project, I would open up Chachi, pt, Gemini, both of which I pay for.
Alex Pokorny: Okay.
Dave Dougherty: And I'd open up Claude, because I've heard from a lot of people that Claude is better with writing,
Alex Pokorny: Yeah. Yeah.
Dave Dougherty: but I don't pay for Claude. So I'm using the free version just to test it out, right?
Alex Pokorny: Sure.
Dave Dougherty: And so I would copy paste. The prompt to each one of 'em, and then I would see what it would come up with, and then whichever one had the best response, I would continue the conversation with.
And since I've been doing this, at the beginning of 2025, it was totally chat GPT.
Alex Pokorny: Okay.
Dave Dougherty: we got to the middle of 25 and it completely switched to Gemini. Gemini was way better as soon as 2.5 came out.
Alex Pokorny: I was about say, I was like, that was the big release.
Dave Dougherty: Yep,
Alex Pokorny: I'm with
Dave Dougherty: that it got so much better.
Alex Pokorny: Sure.
Dave Dougherty: Now I find myself using the free version of Claude
Alex Pokorny: Interesting.
Dave Dougherty: because it has better feedback.
It understands the context and the intent behind the writing. Better than Google. Better than chat PT because like the thing that I'm finding really frustrating about. But particularly with writing is if you have two drafts of something or I really like this and I like that. So combine them, right?
It will just verbatim stitch 'em together. It's no. That's a crap experience.
Like you need to synthesize it and rework it and do that.
Alex Pokorny: more steps with that. Yeah.
Dave Dougherty: Yeah. Whereas, so if I do that, GPT will stitch 'em together. Gemini will understand that, and then we'll put something together. Claude will totally rewrite it and then be like, here's why I did that.
It's I like that so much better.
Alex Pokorny: yeah, I gotta test something. Maybe this upcoming week I'll test it of I've been using chat GPD and paying for it for a little while,
and it has a fair amount of information stored on. History of projects
stuff like that, which it's been absolutely awful with it's ability to review prior chats.
Like it
Dave Dougherty: Yeah.
Alex Pokorny: no information, no knowledge. Been really getting frustrated with that because there's a lot of topics that like get revisited and it's okay, from that conversation, like now we're at this point let's keep going on it. It's like
Dave Dougherty: Right.
Alex Pokorny: no knowledge. It's i'm literally searching the words that the chat has been called in the bar, but maybe there's some way that I can import into Claude all the stuff that chat GT knows about me.
Maybe I can transfer the history because that has been one thing that has been keeping me with chat GBT usage the fact that it
long history, Their, when their thinking model came out, they, one of their it's 5.1, not 5.2. 5.2 is upgrade. But 5.1, it was a like a load balancing change.
So instead of having, you had a slower model, you had a turbo one, you had
These like different models and it was hard for people to understand which one should I be using and asking.
Dave Dougherty: Right.
Alex Pokorny: So they came up with this new one, which basically like tries to understand your query and then shifts it towards one of these other kind of submodels to try to answer it either quickly or correctly or more in depth.
Had less options now, less ability to choose. want the advanced one. When you want the advanced one. Sometimes they would give you the wrong one and there was a bunch of anger online and then they fixed. So things,
Dave Dougherty: Yeah, I shut off that auto select immediately. I said, no, I don't want you making decisions for me.
Alex Pokorny: Yeah I think they're back to it, pushing it again they've been pushing it down and down to maybe cheapen their server costs or fix
Dave Dougherty: Probably.
Alex Pokorny: or something like that. But I don't know. It also concerns me because I use their API, which is actually expensive nowadays. Like I'm starting to wonder if they increase their cost with that too.
Finally increased my speed, which this was a funny thing. Got an email from OpenAI and they're like, Hey you've been using our A PIA lot, so you're now like level two and we've increased the rate that you're at, and we're not, basically, we're not slowing you down as much as we used to anymore.
I didn't know that was a thing, so I'm now in tier two or whatever. It's, it was like the box of mac and cheese that now says, now made with real cheese.
Dave Dougherty: All right.
Alex Pokorny: The only thing that starts is what was it made with before?
We're not slowing you down as much as we were. You're slowing me down.
Congrats. We're not doing it as much. Why? Why?
Dave Dougherty: Yeah.
Alex Pokorny: It was just one of these
Dave Dougherty: You have proven yourself worthy.
Alex Pokorny: I was like looking at it. I was like, Hey, cool. Cool. I guess
Dave Dougherty: Yeah.
Alex Pokorny: at the amount of time. I guess I lost. Okay.
Dave Dougherty: And it,
Alex Pokorny: finds you happy. No, I found me sad. It found me
Dave Dougherty: Yeah. You made me sad. Thank you,
Yeah, the and that's the thing too
Enterprise AI Adoption and Challenges
Dave Dougherty: . We talk about enterprise marketing a lot and in the context of enterprise marketing in these large companies. I don't know that ai, the AI productivity thing will ever hit as big as the leaders will hope for all of the reasons that we've talked about today, where a, you need a bunch of leaders totally aligned on a single direction,
Alex Pokorny: Yeah.
Dave Dougherty: so that's going against it.
Thing two is you need to be empowered to. Try things, break things and make decisions to move forward.
Alex Pokorny: Yep.
Dave Dougherty: You don't get that in big organizations. When you were talking about one of your scenarios, I was thinking to myself like, that reminds me of when we were working in a coworking space together and we could just lean over and look at each other and be like, Hey, you know what's dumb?
Alex Pokorny: Yeah.
Dave Dougherty: Let's fix that.
Okay.
Alex Pokorny: got fixed that way
Dave Dougherty: Yeah, exactly.
Alex Pokorny: is
Dave Dougherty: I have yet to have a project since then. Like that,
Alex Pokorny: yeah.
Dave Dougherty: Because it's oh, here's what, here's as much as we can do before we hit, it's red tape, or you have to go get this person's approval. And it's but I don't wanna talk to that person, like
Alex Pokorny: I know that was the funny, like la land, the magic land, the fantasy that ai, a lot of AI posts and comments live in
is this fantasy world where there are no limits to anything and everything is just possible. And it's that just simply isn't reality. we can run projects with as much AI as we possibly can, but we're still gonna run into all the other problems.
We have an RFP. Great. I've got an RFP, GPT. Great. The RFP has now completely changed. I guess that GP GT's worthless,
Dave Dougherty: right,
Alex Pokorny: right? So there's like this, repeatability and limitations that we always have because of things being so custom. Constantly changing.
Dave Dougherty: That's a good point actually, is like your GPTs or your custom gems or whatever, those now need to be included in your rebranding.
Alex Pokorny: Oh, yeah. And they gotta be updated. Like some of my
Dave Dougherty: Yeah,
Alex Pokorny: realizing they're also running old models.
I have to just go back to them just to edit them so that they start using the current models.
Dave Dougherty: Right.
Alex Pokorny: that's a very small thing. But also I tested some of them and realized new things aren't working and also new things have changed.
So yeah, brand standards
They were built on have been updated. So now I have
Dave Dougherty: your prompt probably has to change because it's a newer model. Yeah.
Alex Pokorny: The tech debt, basically, each one of these now creates maintenance administration
Dave Dougherty: Yep.
Alex Pokorny: People, coworkers ping me occasionally saying, Hey, I got this weird response. Can you help me out? Or, I'm trying to do this project, can you help me out to do it?
So it's saving time-ish. Yeah, maybe
Dave Dougherty: Right.
Alex Pokorny: time somewhere else. To be honest it's pretty hard to tell for productivity too. yeah, I don't, I. I do question some of that. I think it's still useful, I mean across the board, but
Dave Dougherty: Oh yeah, and I forget the episode where we talked about the actual adoption amount, but I remember being shocked at. I don't consider myself like a crazy advanced user, but according to the stats,
Alex Pokorny: I was gonna say,
Dave Dougherty: We are very advanced users.
Alex Pokorny: chat? GPT end of year number. Top what percent of users?
Dave Dougherty: I don't remember that one, but I do remember that I was one of, I was in the first 1% of all GPT users.
Alex Pokorny: Yeah. I think I was at 3%.
Dave Dougherty: Super early adopter. Yeah. So that, I was happy about that.
Alex Pokorny: yeah. So when 99% of users, which apparently they have millions are below you,
97%, yeah, we might be kinda on the edge and that is actually, that's the other funny thing too, is once you have these discussions with other individuals who are heavily within kind of the ai, like trying to move forward.
Dave Dougherty: Yeah.
Alex Pokorny: break stuff and move and learn environment and mentality. They all always have the same point where they always think oh man, I feel like I'm falling behind because
Dave Dougherty: right.
Alex Pokorny: out there who has a, a LinkedIn post that got, reposted a dozen times.
Dave Dougherty: Right.
Alex Pokorny: that's way cooler than what I'm doing or way more advanced than what I'm doing, and it's first a question of believability of anything that anyone posts online and you
should you're in a little tiny bubble, right?
Dave Dougherty: Yeah,
Alex Pokorny: Like the vast majority of the world gives no care to this and may hold the same role as you at a different company.
Dave Dougherty: right. I had an interesting discussion with a coworker of mine whose dad was a doctor, and when they started forcing. The software, like note taking and whatever else the software does. He refused to do it. He was just like, no, I'm not gonna do it. I don't want to do it. So then he worked as long as the company would allow him to not do electronic records, and it was just like that was his choice.
Could he have worked longer? Probably. Probably, yeah. And I was, I think of that story though, a lot when I'm talking to some coworkers about Hey, you could try ai, we could do this. We could try that. Like just do an experiment. And they're like, eh, I don't know. I know I have to email this person.
I'm gonna go email that person's like. All that sounds like really busy work. Like you sure you wanna be doing like only busy work, but then I look at the half of my day and I'm just like, damn.
Yeah.
Alex Pokorny: But
Dave Dougherty: Yeah.
Alex Pokorny: that model we mentioned before. What was that? The trash can model. I the phrase spaghetti model, but it's like the mess that is actually inside of every company of how things actually get done. And then how little of it is actually like effort towards building the product or selling the product. Like so much of it is just all the internal stuff,
Dave Dougherty: Yeah,
Alex Pokorny: man, I want an AI for that, for the other stuff.
Dave Dougherty: yeah. Who can do the internal office politics, ai, and,
Alex Pokorny: Yeah. Yeah, that does that. Thanks.
Dave Dougherty: yeah. God. Okay. That took a turn at the end, but I did enjoy the the conversation as always. Everybody, pathways newsletters coming out every week. Got a good schedule going for that. So go check that out. Share this.
Alex Pokorny: Yeah, I
Dave Dougherty: Okay.
Alex Pokorny: those. Check those out.
Dave Dougherty: Awesome. Thank you for that. Yeah. They've been, once I figured out what my week needs to look like to push those out on time, then they started getting better.
Alex Pokorny: That's.
Dave Dougherty: Yeah, go check those out subscribe, share these with a friend so that we can get more people involved. I do have to thank everybody who comes up to me in real life and. It compliments the podcast, it means a lot. And I've had some awesome conversations with people who have done that. So thank you and we will see you in two weeks in the next episode.
So take care.
Alex Pokorny: Cheers.
