Heads In Beds Show - Here's The Things That AI CAN'T Do To Help Your Marketing (Yet)

Episode Date: June 24, 2026

In this episode Conrad and Paul dive into some things that AI cannot do (yet) and explain how and when to use your AI tooling to get better marketing results.Enjoy!⭐️ Links & Show NotesPaul Manzey Conrad O'ConnellConrad's Book: Mastering Vacation Rental MarketingConrad's Course: Mastering Vacation Rental Marketing 101🔗 Connect With BuildUp BookingsWebsiteBook A Call With Us🚀 About BuildUp BookingsBuildUp Bookings is a team of creative, problem solvers made to drive you more traffic, direct bookings and results for your accommodations brand. Reach out to us for help on search, social and email marketing for your vacation rental brand.

Welcome to the Heads of Med Show presented by Build Up Bookings. We teach you how to get more vacational properties, earn more revenue per property, master marketing, and increase your occupancy. Take your vacation rental marketing game to the next level by listening in. I'm your co-host Conrad. And I'm your co-host, Paul. All right, Paul, we're back after a short hiatus, as it were, because I was on vacation. What's going on?

What's happening in your world? It's summertime, and AI is the farthest thing from my mind. So it's probably a good thing that we were talking about. this. So how are you doing, sir? How was your vacation? It was pretty good. Yeah, it was actually kind of a unique thing. It was my family came and then we had like a few days with some team members that I'd not met in person before. So we had like a client success, you know, like sort of mini retreat within the vacation, which was cool because like, I mean, I've worked with some people for almost three years and I never met them before, you know, in the actual flesh. So that was a fun

opportunity. And I, I enjoyed that. So yeah, all good stuff. A lot of things that we need to work on, which is always revealing when you, you know, kind of zoom out for a second and you don't talk about what's this one problem with this one person or this one client and you zoom out and think like what are the things we're trying to do better? And it's a little overwhelming at times to come up with all these ideas. And the hard work is always, all right, you know, how do we then make it a reality? But it was good. And yeah, looking forward to a lot of improvements, if you will, over the next few months, which that's all we can do, right? Just keep getting a little bit better as time goes on and we'll get those things in good shape. But it's funny. You joke about

AI, what it promises, because I think that this is probably a topic that not many people would do right now, so we thought we'd have some fun with it. We're all things that AI can't do and started, this idea came from a rant that I started with when we were talking before we record about meta ads. So maybe we'll just, and this is kind of a grab bag of different things. And you have some high level thoughts. Then I have like some, you know, very specific things. But, you know, we go through a lot of these ideas relatively quickly. But really simple example, like we want to use AI to help manage our advertising campaigns. I think personally, I found Claude and these AI platforms to be a little more quote unquote useful for Google because what it does really well is things that we can't do.

very effectively or efficiently. So for example, like AI has no problem reviewing 7,000 different search terms and then measuring that against the landing page and seeing if those search terms should be matched or not. That's awesome. So I love using Claude and tools like that for that purpose. But if we flip over to the meta ads piece for a second, one thing that there is an official meta ads MCP connector for your LLM model of choice, whether it be Claude or Gemini or chat GPT, which is great. But it can't upload videos. So it's a good example of like if someone's telling you right now that you can run your meta ads through an MCP, And I'm like, well, it can be a, it can help you.

It can be a little assistant. It can be a little researcher. You and I have talked at length about this, all the things that it can be a value ad. But if it can't upload videos into the dashboard and I was sharing with you earlier, that's 60% of our best forming ads are videos, then you're missing a key part of the process. We still have to go in there, upload the video. And then once we have the video and the hash and all that, then we can rebuild the ads through MCP, which is great.

But, you know, good example of like, if someone's telling you right now, oh, yeah, you can run your meta ads completely through AI. that's just not the case. What you can do more so is like assist your as management through the, you know, connectors and MCP and things like that. But, you know, it's just one kind of thing that started this whole thing. So that'll be my first one. But whatever, your kind of thoughts either on that, or do you want to take the next one and just kind of go rapid fire here? I think you hit it. It's it. It's one of those things. It's a basic straightforward, seemingly straightforward implementation, but very impactful there. So, I mean, I think that that's number two, reliably pushing changes

into ad accounts. I think reliably doing anything is something that this is where the reason that I struggle to let AI take control is I have to answer to whatever changes, whatever, you know, hallucinations, whatever small mistakes may make. So it is. I think that's something that, you know, you were talking about depending on the platform, depending on the connector, depending on what you're doing, you don't know, unless you're physically hitting that button. And again, I'm a, I'm a physically hitting that button kind of guy. I like to know that it's happening. I'm not saying that I can't trust the AI,

but I have had it not work the way I wanted to. I have had a budget not change. I have had things not stay, remain on or not turn off. So those are things that I think, maybe that might be a personal bias, but I think there's some automation points that it's not, as reliable as you wanted to be.

Even, you know, I mean, if I think about like technically, you know, developing like car jobs and doing stuff like that years ago, I felt more comfortable that those jobs were going on every night. I don't have the same confidence in AI and maybe that will develop. But I think that that's reliably pushing changes into anything, much less, but at accounts where we have money on the table, you have to have that confidence that nothing's going to go wrong. And that is, I, I'm close, but not quite there.

It's trust but verify still for me. So what do you think about that? Yeah, actually, that's funny. You said that that was going in my head, banging around when you started that block there. Yeah, trust would verify. I think it's great. You know, like the idea of, you know, some of these things I have done, for example, like pausing ads and things like that.

I've absolutely done that through MCPs and connectors. Not fair as a connector that I've used quite a bit recently. I'm actually going to operate to their paid plans soon because I'm finding that the free version was doing good job. So that connector's been great. But yeah, I do take the extra 15 seconds, you know, the first handful of times I did it to be like, that actually paused just to make sure, did it do what I expected it to do? And so far, it's been great. You know, I haven't really had any trouble with it. But there's things that, you know, to your point that it can't do. So just be aware. And like you said, we are accountable for running the ads. And so if we trust the AI to,

quote, unquote, do part of our job for us, either A, better be reviewing its output or it's, you know, sort of notes or logs and saying, yep, that's working as it should. Or be ready to face the consequences. If it does make some kind of mistake or problem, you know, as the ad manager or someone, you know, doing that for a company, for a client, you just have to look at them in the eye and say, oh, yeah, I didn't look at it after the AI did it. And if you're comfortable doing that, great. But if not, you may have a little bit of a tougher go of it. So, yeah, there's things that can't do, though. So I'll take the next one here. You know, for example, just one example that I had in my notes here, fixing a broken URL. Like, sometimes that is a problem in the ad platform. Sometimes

the problem on the website. So then it becomes an exercise and, okay, why did this URL change? who changed it? Do I want to just change it back to what it was? Did that URL become invalid? Did the client stop managing properties in that community? Like, there's so many different reasons of what sounds simple in the surface. Hey, fix a broken URL can be so problematic. And I've actually had AI suggest that. Like if there's a URL that was broken and one of my ad accounts, Google, this is back on the Google side. And the AI suggested, Claude in particular suggested we'll send it to the homepage until we figure out what this URL is. And I'm like, that's like band-aiding the problem. Like, yes, that does solve it. You know, I'm putting the

scenario quotes for listener. This does solve the problem. Sure, the homepage is not broken. fine, but I'm like, let's figure out what happened here. In long short, it was just an accidental change that was made by someone in the back end. On the client side, they added some text content to the page, which is totally cool, but then they accidentally changed the URL when they did that because it like renamed the page title or whatever the case maybe or the page title and then that swapped in the URL field in WordPress. So a quick change back to the way the URL was, I'd log in WordPress, go back change. It took me all of two minutes and change to fix this problem, but just something that AI didn't

quite get right, you know, and as a result, that human in the seat, you know, kind of thing, or human in the loop is kind of the terminology that I've heard people use again and again is so critical and so important. And I was going to bring this out at some point today. So I'll just bring it out now. I love that. I saw this tweet the other day. I need to credit it properly. But I totally forget who it was. But it was like everyone is claiming these 10x productivity gains or 10x, you know, getting the work done gains or, you know, Jensen from a video talking about an engineer ready to be burning five million tokens a month or whatever, which is funny because it's just talking up his own book and what he sells. But whatever, that's fine. But the tweet was to the effect of, okay, if you're a 10,

X engineer, show me 10x the revenue, show me 10x the user, show me 10x the profit, show me 10x the active number of people using our platform. And the AI platforms can point to that, right? But if you look at a lot of these different things that we're seeing, you know, touted as these amazing advancements, there's kind of a little gap right now, isn't there? A little bit of hype, you know, machine that we're kind of in the middle of right now. And I think what sounds simple, you know, in this example here, a broken URL, how do we fix that? It still takes like, we're not quite there yet where it's like working perfectly. And as a result, you got to do a little of that human in the loop digging around and kind of figure out the core problem.

But am I glad I had the AI there checking to because will I click on every single URL. You know, could I click on every single URL across every ad they're running every single day? No, I couldn't do that. I don't have time for that. Can the AI do that? Absolutely. So, you know, my analogy, again, of the Iron Man suit, like the Iron Man suit can allow Tony start to do things that he can't do. That's the Iron Man suit.

It is awesome. But like he's still in control of the Iron Man suit, right? So that's kind of the, he's still kind of checking on things, make sure it's working properly and makes him go faster. But it's not this, not this perfect, you know, solution to everything. So I digress. But what are your thoughts? I think it's, all these ad changes, I'm going to jump down to the jump down one and go to approving, you know, its own budget or spend changes. I mean, you've used automation for budget changes and kind of budget automation before.

And I think that's, that's something where you're talking about that gap a little bit overall. And I do. I think that that's where we started to create new things when AI came out initially. Now, I'm not going to say we've run out of. things to do, but we're creating problems to solve. It feels like we're doing the, the fun stuff. We're creating our bottlenecks and then we're fixing them. But we're not doing anything new. We're kind of Toy Story-Fiving it right now with AI. That's fine. It's good. Hold on. I'll let explain that real quick. I feel like I don't understand that reference as much as I

want to understand it. Toys, like the franchises every week. As we are not creative people. It's Toy Story 1 through 5. It's it's all these movies. We can't get out of our own. We can't write anything new and original. And I do. I think that that's something that we're not really creating new stuff. We're maybe putting a little lipstick on a pig with AI and making it maybe more fun, more functional, gamified, more, things like that. But I think that's where we're kind of reaching that gap of we don't know how to take AI to that next. I think maybe AI is the only thing.

And that's a whole other meta, scary thought. We can get into that at some point. But I do. I think that when we talk about approving its own budget or spend changes, I don't feel comfortable at that. Did I feel more comfortable with an automated bid management system? For some reason, yes. Even though ultimately it's the same thing, you're writing the same type of rules. You're seemingly having slightly more control, but not really because it's all automations behind the scenes. We're just putting AI as a decision maker behind that as well. But ultimately, aren't we doing that with automation? I don't know. It's, it's, it's,

of those where it's I think we are in that weird straddling the line of what is AI and what isn't AI and what is just the algorithms that we previously built and how is that impacting the way we're making some of these decisions you know anything relating to money budgets anything like that it's scary I don't want them making the call I don't want them to have the authority because ultimately they don't have the AI and I'm calling them they doesn't have the accountability. They're not the ones who the AI is not the one who's going to get on the call. The AI is not the one who's going to have to justify why to spend adjustment was made. So yeah, anything relating to money and AI, I'm struggling to give it full control there.

I mean, I guess that's where where are your lines of decision making where you have to make the call or you're feeling more comfortable? You mean, you've got a team beneath you as well. That team also has the authority, accountability, responsibility to make those decisions. decisions. Where is that decision tree for you? I think that's why some of the software that we've used for a long time has become so valuable. So like optimizers are a budget-pacing software and I do trust that system. Now again, could you log in directly the interface and double-check? Absolutely. But, you know, I think it's fine to basically say here are the parameters. Actually, if I could do an analogy, I think the listener would connect with this. It's very similar to how like pricing software

works. Like you can give a pricing software no sponsors, but like someone that can price labs or whatever. You can say don't go under a $220 night and don't go over $500 a night. But like play within that window and try to find out, you know, what my optimal rate might be for this property and then give it some kind of some free reign. That's my understanding. Having looked at some of these tools in the past, understand at a high level how they work, although I'm not a revenue manager, don't pretend to be on the internet. That's my understanding of how those things, you know, can actually operate, which is, which is great. So I think it's the same kind of thing, right? Like, I think it's okay to let a software kind of help you and give you some options to make some changes and things like

that. But, you know, it has something that you probably want to keep eyes on and double check and say like, okay, the software said this, you know, did that actually end up being true, you know, or was there some mistake? Or can I sometimes go against the recommendations of the software and go back? You and I could point to one. This wasn't in our outline that we frequently have eye-twitching, you know, laptop throwing moments on, which is Google Ads recommendations. Some of the recommendations in Google Ads are good and you should follow them.

Some we found are bad and really don't give you good performance. And the unfortunate truth of Google Ads recommendations is that if you sent someone brand new into Google Ads who had never run Google Ads before and just had them click apply in every recommendation, their account would actually be a disaster, which it's surprising me that Google allows that. I guess maybe they get away with it because a small advertiser and not spending a big budget can just kind of let a lot of the dollars be wasted as long as some stuff is converting it kind of like all kind of fleshes out in the wash or whatever the case may be. But if you're a large advertiser, I don't even large, I mean, if you're spending, I don't know, two, three, four, five thousand dollars a month or more, which is pretty much any small business once it gets going a little bit and is relying on advertising to drive business, you know, to their to the website, draft traffic and bookings their website, you got to go put eyes on that thing. And you got to understand what it's suggesting and why and all this kind of stuff.

And I took over management of an account just, well, I signed the contract yesterday. I'm kind of starting it on Monday. And this person is like pretty savvy. It's a very seasonal business. It's kind of like adjacent to vacation rentals, but it's not vacation rentals. But they would just have a call where the Google rep every two or three weeks. They would listen to the Google reps kind of thoughts and recommendations, P.U. Some of his own judgment.

And then he just kind of clicked apply and everything. And most everything the Google rep suggested. And there's like some good in that account. It's not like a total train wreck. But there's a lot of bad too. So I'm like, well, we're going to spend 30, 40 minutes on doing a all the bad stuff that I think is not actually harming you. And here's why I think that, by the way. And then we're going to go make these changes. So it's like, you know, judgment experience.

I don't know. It'd be interesting actually to, you know, we could do this, go through and look at the Google ads recommendations tab and then ask a lot about each one and be like, when should I use this? Why shouldn't I use this? It'd be funny to kind of see that. That'd be good, that be good little test there. Absolutely. Maybe I'll grab the next two and bunch them together because then I think you can take kind of the next bucket as we go into it. So I've not seen anything here. The only thing that I can see is like there's not really way to bulk update like listings across OTAs. The one thing that Claude is really skilled at that is kind of like adjacent to this, I would argue, is let's say you were going to add something to the first line of every single one of your

property descriptions. And you had to log into your PMS of choice, whatever that may be, streamlined track, owner as, doesn't matter, whatever it is host way. Cloud can actually use your computer, which is kind of compelling in computer mode. And it could, in theory, like if you trained it or showed it what to do, it could open every single URL or open every single, you know, text description, add a line, click save, that sort of thing. So it's like, it's kind of, again, one of those things where you want to double check it. But it also fails a lot.

Like the moment there's authentication, it's going to fail, the moment that it's going to go tweak, you know, have to have to like click on multiple things. Sometimes it'll fail if you haven't given it the perfect construction. So, you know, those kind of things of like, again, OTA platform is PMS editing. It's like not going to do that perfectly in my experience or it's just going to be so clunky. Almost to the fact where sometimes I'm watching cloud computer use and I'm like, dude, it would have been faster for me to do this myself. It's like it's so, it's so frenetic, just clicking around looking and trying to do things that it doesn't. doesn't actually save you as much time as you think. Same thing with something like a Google My Business profile or posting content to Facebook, Instagram, you know, all that kind of stuff.

It doesn't really do that that well in my experience. Even if you gave it the creative and said, go post this stuff, you know, on Instagram, it's going to have problems probably, you know, logging in, using the scheduler, you know, et cetera, et cetera, et cetera. And yes, I've tried this kind of stuff and it's just kind of clunky right now. There are some, I'm noticing more and more I get ads for this kind of stuff nonstop too. There's more and more MCPs coming out every day, though. So all this stuff could be, you know, moot point here in six months or whatever the maybe where you can then update Google My Business, Instagram, et cetera, through an official MCP connector.

But just check what the MCP can and can do. Go to the permissions tab. If there's a connector and look at what it can't do. There'll be a list of all the things that can't do there. And there might be 40 things that can do, but there might be 90 things that can't do. Right. So just kind of going back to that thing of like understand what you're getting yourself into. And don't just kind of buy into the nonsense, right, of like, you know, people that

read this stuff online about you can do everything through this interface. It's going to help you, but it's not going to do all those things across those platforms directly. So yeah, you want to take it. kind of a few out of the next bucket and some thoughts on some other pieces here. Yeah, I mean, we've talked about the creative side of things and using AI for images and things like that. I think that there's a lot that you can do with AI when it comes to images. But I think ultimately what we have to root everything in this, you make it rooted in actual photos of the properties. I think the fact that AI can generate a beach house doesn't mean

it's one of your beach houses. It doesn't mean it's, it's, it's, it's, actually your property. It doesn't mean it's your anything. So I think that that's very, we do. We have to ground everything we're doing in actual reality because without that, how can you give that guest, that potential traveler, an actual understanding what experience they're going to get if none of your images are actually of your property? So I don't know.

I think that that's, we've talked about stock imagery versus, you know, AI photos. and it's different. Like a stock image has been taken. It is, it does, you know, someone took a picture of something that exists. That is not the case with AI. And I do, I think that that's something that can be very misleading at a very low level and very deceptive and just wrong. I think we border on illegal with some liberties that people take with images. You know, I think even adding people to.

We talked about videos, adding people into videos or images or stills or things like that. It's not authentic. And it's not truly giving people the experience that they might be getting there. And it is. I've created some really cool videos. But when I'm gone back and looked at them, they're not good. I'm not happy with them because they're adding things in. I'm taking 20, 30 still images.

and just having it create a nice video. And it's still hallucinating things from the images that aren't there. So I think anything relating to creative, and you can go into just the written content as well. It's something that we don't have that full control because we don't know the ultimate sources that are being used to fill out that ultimate creative. But wherever you run into specific areas

where you've been concerned about creative or you've seen creative that just doesn't match what the expectation should be. Oh, yeah. I mean, like, it's so funny when you go to publish a meta ad now. And they're kind of like, you know, I spent anywhere from like 30 minutes to an hour and a half working with the client and my designer, you know, Pedro, on getting the ad exactly how they want fonts, colors, you know, which property do we put

in the background? What's kind of our message or hook, et cetera, et cetera? And then the last page before you go to publish a meta ad now is like, do you want a bunch of AI generated slop images, you know, to test against this one? And I'm like, no. We worked really hard on the one that he gave you. That's why I spent so much time with my designer. and spent so much time with the client approving it.

What's so frustrating about meta, though, is that, like, if you just lose your attention for a tenth of a second, they check in some automated box of, like, I've had it happen multiple times now where they checked in some automated box on a video that we made or an image that we did for the client or whatever, where they're like, we're going to go translate this into another language. And I don't know if I want that or not because, like, it may not be perfectly accurate. So, and most of our clients are speaking English anyway. So what's the, you know, what's the upside of that? I don't know if there is much of an upside to that. Or, again, like, we're going to use this AI generated image of a B.

house that's not actually your house. And I'm kind of surprised, you know, too. I do this every once in a while. I go to Airbnb. I just put in no dates and I just click on popular destination. So I go to the Airbnb page for Destin, Florida or Myrtle Beach, or I did one for Deep Creek Lake a little while ago. And I just click search with no dates added. I just kind of see what's in the top 10.

And every once while, I see like a very obviously clear, clearly AI generated image showing in the top 10 or 15 results on Airbnb. And I feel like they're probably going to stamp that out. Some of the ones that I see don't seem to last very long. So I wonder if they're just like testing them and then they don't do well and they kind of fall out of the mix a little bit. But I mean, it's already happened, I'm sure, right? Someone has booked a property with an AI generated image, right, that they thought was

this particular place or this particular listing. And I don't mean edited. Again, the analogy I always do is like a model would never walk on stage without makeup on, you know, him or his or her body, right? Like, that's just how that would go. I'm not talking about makeup or cleaning up a little cloud in the background. That's not what I'm talking about. I'm talking about like a truly AI generated image that doesn't look anything like

is not anything source material of what your actual property is. and then people write that as an ad, I think it's just asking for either a horrible review. Best case scenario is a horrible review. I think worst case scenario was you get banned from the platform, you get removed because it's like we can't even trust that what I'm looking at is real. So yeah, I'm sure for other industries and spaces, being able to do AI generated images is awesome. I've seen people talk about like e-commerce.

They're selling watches and they're like, oh, we take a picture of the watch and then we can put it on like different people or you can put at different settings. And I'm like, that's awesome. Like that's cool, right? Because again, what you're selling is the watch. You're not selling a thing behind it. You know, but yeah, that's fine. But with what we're doing, like, it has to be rooted in.

reality. So I just, yeah, I just can't get vibe with that. I think it is. It's retail versus experiential, truly. I think it's retail you can get away with making the product look nicer by not impacting the ultimate view of the product. I mean, ultimately the experience, everything around the rental is still the rental. So I do. I think that that's, that's something where if you're, if you're painting an inaccurate picture or it not even because it's not intentional but AI cannot take that picture and not influence it in a way. I think that there's something there that I don't know. I think we actually are shortchanged because AI has put, I mean, not AI, Google has put some incredible tools into kind of the image generation on the ad side of

things to be able to enhance your product pictures except it's not something that we're going to do. because ultimately that's more deceptive than helpful for us. Yeah, yeah, I couldn't agree more. Well, let's see. Anything else? Maybe we can kind of build off this last bucket and then come to a close here. You know, just to what, again, there's more things that we could do. But last thing, analytics.

Like, again, do I like a lot the idea of downloading analytics data, you know, having it look at things like referral sources, having it look at things like, you know, hey, is there, you know, tell me traffic trends that are occurring this month versus last month or traffic trends that are changing year every year and all that kind of stuff. Love it. I think it's great to download that data or there are some connectors again or ways that you can add in Google, things like Google Analytics data and things like that to your LM system of choice. Love it. But can it know if it's accurate or not? Like no. Like you may be double counting it may not know that. You could have the pixel installed on two different websites and it may

not be able to really see that without, you know, doing a deeper dive that it may not do by default. It can't really probably, again, unless you have this connected, which very, very few people would, it wouldn't be able to match up like direct bookings in the website with direct bookings on, you know, from your PMS platform. So that's something you kind of have to do manually or at least go gather the data manually and have AI assist you. It certainly isn't going to be able to fix some of the attribution holes that we've lost over the last, you know, year or two with iOS privacy rules, cookie loss data, all that kind of stuff. There's just nothing there for them to look at. So, you know, it can look for things like, and you flag this actually first, the Paul

AI found this first, you know, why are you getting a bunch of traffic from China and Singapore when none of your guests are coming from there. That kind of thing that maybe AI would eventually grab, but you just caught it more manually. But those kind of things like helping you kind of researching, hey, I'm looking at this analytics data. I'm trying to understand kind of what's going on where people are coming from,

all that kind of stuff. It's great, but there's just things that it doesn't know. One more example on this, then we can move on. But I was looking at a client example the other day, and he was using analytics data to build a city targeting list from, you know, so hey, we want to run our ads to these X number of locations because this is where we're getting the most traffic and the most bookings from.

And I had to remind him that iOS data in particular is, anonymized to a relatively large region. So I said, you know, for example, sometimes I go to access my phone. It'll say my connection on iOS is coming from Atlanta, Georgia. For reference, I'm about six hours by car from Atlanta, Georgia. So if someone was targeting me thinking that I'm physically located in Atlanta, Georgia, they'd be very mistaken. Again, I'm six hours away with the way that, you know, my phone location is so general. It's not, it doesn't know exactly where I'm located in North Myrtle Beach. So, you know, just kind of those problems are ones that I see quite a bit. I mean, we're not saying that AI is not a beneficial tool that's helping us.

No, no. And you shouldn't be leveraging it. However, there are clear areas where AI just doesn't measure up to a human touch or some kind of human touch. So ultimately, we're still in trust, but verify mode. And we're a year and a half into the AI experiment. So I do. I think that we should all continue to be of that mindset of, you know, use AI as a tool. make sure that the most important decisions of your business are still running through some

type of human element so that you're you're you're you're comfortable again someone is accountable someone's responsible yeah i can't be accountable that's just something that it's not there yet and i don't hope i hope it's not like that's something i that's the next level that's super ai and we're nope nope nope not there yet yeah you know it's so interesting two thoughts here before we come to a close because I'm up against the timewise, unfortunately, today. But one is we did an interview with Ashley Ching on Art of Hospitality, and she talked about staffing levels and service levels, and she gave examples of a hotel lobby.

I thought this was so great. Like, I've not heard anyone describe it in this way. It makes complete sense. If I'm checking into a Super 8 or a really, really budget low-end hotel, and an AI is there to help me, and I can chat with the AI and get into my room quicker, I think that's very appropriate. And actually, what I might expect out of a Super 8 motel experience. Because the alternative is there's one person up there.

There might be five people trying to check in. That person's overwhelmed. they're overwhelmed, underworked, etc. I know I'm getting like a very, you know, minimal, you know, experience here. I know what I'm kind of buying when I buy a Super 8 or a very cheap roadside motel experience. So when I get the AI, it's actually probably vastly superior to the human experience in that regard, right? On the flip side, you kind of compared to like if you walked in the lobby of a high end premium five-star hotel, you walk in the four seasons hotel in Boston, which I've done before to visit

my uncle. And I didn't stay there, but I was just visiting him there. And you walk in and there's 14 to 15 different stack members, both with check-in, there's someone opening the door for you. There's someone helping you if you had luggage, you know, there's someone that will bring that to your room, et cetera, et cetera. That's kind of what you want when you pay $1,000 tonight, right? Like if I got up to the front desk and it's like, welcome to the four seasons, you know,

I'm your AI assistant. We're helping you check in. You might actually be like, that's not what I expected, right? So there's a whole bit in there about expectations and price and value and all that kind of stuff and what you're paying for. And I think like as agency owners, you know, as consultants, whatever, you know, kind of role you and I have, I think it's almost like we got to pick a lane a little bit, right? It's either like use the AI and just kind of have it give you that very automated service,

and that's fine. But it's cheap. So that's, you know, you're kind of get what you're getting there. Maybe there's some pros and cons to what it does there. Or you have someone running the AI for you, you know, or you're working with this consultant and they're layering in all this judgment on top of it. And they're saying, here's what I think. And here's why they're building a more custom plan.

And I kind of think both are valid in a way. Like, if you're a small business, I think having AI help with your marketing is actually a huge advantage to be clear. That's why Paul and I are not saying that we don't like these AI platforms. We love them. Because I think doing it yourself, we're trying to like DIY your own marketing or having no help or feeling like, how do I even understand any of these numbers or metrics. and I will sit with you for hours and just explain things over and over again to you of like,

I don't understand what that means. What is a session? What's the difference between a session in user and analytics? They'll sit there and explain to you. You could go, I don't get that. Make it simpler for me. And it'll sit there and do that with you endlessly. That's awesome. I think that's great because I think a small business now using these tools. And I feel as a small business owner, I use AI to help me with things that I don't understand well. And I would have previously maybe gone to an accountant or gone to this or that. Not super mission, critical stuff, but just stuff that I'm just like, I don't really get this. Like, what am I looking at here? Or you know what? Read this contract. I had a client send me an NDA the other day. I'm like,

Before I sign this, give me in plain language, what am I agreeing to when I sign this? And I trust AI enough to do that because it's not super, super, super to high stakes. You know, I read through it myself as well. But it's just great to have that little assistant. But the moment you get to that business and things are scaling, I think you're going to demand more out of both AI and your partnerships that you have in marketing and the people on your team. And I do think the right person with the right AI tools, that's the superpower. That's the Tony Stark. That's the ultimate thing.

So that's what I'm really bullish on. But all good, man. Well, Paul, we'll have to save any future thoughts for next episode. I've got to put a ball in this one and head over to the next one as it were. Dear listener, leave a review. Go to your podcast, app of choice. Click five stars, iTunes, Spotify, listeners

to get the most downloads there. We super appreciate a review. It makes us keep going here and we'll be back next week with another episode. I hope you have an awesome rest of your day. Thank you very much.