Heads In Beds Show - How We're Using AI Today (Mid 2024 Update)
Episode Date: July 17, 2024In this episode Conrad and Paul talk about HOW they're using AI today right now in the middle of 2024 and how Enjoy!⭐️ Links & Show NotesPaul Manzey Conrad O'ConnellConrad's Book: M...astering Vacation Rental MarketingConrad's Course: Mastering Vacation Rental Marketing 101GeminiChatGPTCustom GPT For Ads AI Overviews: About last weekAdCreative.ai🔗 Connect With BuildUp BookingsWebsiteFacebook PageInstagramTwitter🚀 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.
Transcript
Discussion (0)
Welcome to the Heads in Bed Show where we teach you how to get more properties, earn
more revenue per property, and increase your occupancy.
I'm your co-host Conrad.
And I'm your co-host Paul.
All right, Paul.
Good morning.
How's it going?
Good afternoon.
Sorry.
Good afternoon.
Yeah, this is Monday's a morning, Thursday's the end.
Yeah.
We're giving away all the secret sauce behind the scenes now.
It's a beautiful day.
It's that summer feel.
We're in the heart of summer when we're publishing, but it's, it's, we're just getting
gently rolling into summer on my side of things. So that's, that's always a fun time. And I know it's a fun time for the people we're talking to here. And I think, you know, it becomes a
busy time for some people. So hopefully some of the things we're talking about, the AI are going
to just meld right into that summer conversation.
But how are you doing today, sir?
Yeah, pretty good.
We were joking about before I hit record some new technology on my side with a new camera
setup and stuff like that.
I think I've got it kind of tuned in.
But you know what it is, though?
It feels good to learn new things.
You know, ultimately, I feel like I don't mind that feeling of being a little uncomfortable
and not understanding, wait, I got to do this and stuff like that.
I will say my desire for learning new things
has gone down over the years, just because I have less time.
Like I used to think there was this time in my life
when I would like learning new things
because I could fiddle with something for four hours
or five hours and kind of learn it.
And there wasn't really any like,
nothing was waiting for me on the other side.
I don't feel that way nowadays.
So now I get like a little anxiety
when I'm learning new things
of like, I should be doing something else.
But I guess the way I look at it is like, we're not gonna get better if we don't learn new things. when I'm learning new things of, I should be doing something else. But I guess the way I look at it is,
we're not going to get better if we don't learn new things.
So I'm learning new things.
We'll see with video.
And we'll see how this all turns out.
But consider this a beta episode from a video perspective.
And if I learn something, I'll pass it along,
and we can figure out better ways of going down the road.
But for right now, I feel like that feeling of when I first
started doing SEO or when I first started doing Google ads, of I kind of watched a few tutorials online, I kind of know, you know, little bits and pieces, but I feel like a little baby giraffe, which is not too dissimilar, honestly, from the AI stuff to, you know, like, there's, you know, there's that feeling initially of like, what do I do with this thing? How do I use it? You know, that kind of stuff. And I think that's, it's, I don't know, I like that feeling. Sometimes if it's that feeling constantly, it's a bad thing. But I like that feeling in my stomach a little bit almost.
I don't know. I think early on in some of the AI,
I mean, just the chat GPT releases and stuff like that.
That was my thing is like this looks cool and everybody's got these cool prompts
and everybody's using it for these cool things.
I couldn't think of things to use it for like that.
I couldn't think of prompts that were going to be beneficial to me.
I mean, outside of the, ooh, write this content for me, ooh, write this.
I don't really like when other people write for me.
That's, I don't know, editing was always something that was difficult for me because I was kind
of a tough editor when I was editing some of the
content for blog writers and stuff like that. So that was, it's like, I think now looking, you know,
a year beyond that and looking at some of the tools that we are using, now I can't think of
things that I would use instead of these, but that's half the battle with AI is knowing where
to use it. I mean, that seems like half the battle with AI is knowing where to use it. I mean, that seems like
half the discussions on LinkedIn is knowing where to use it, knowing where not to use it, knowing
what's beneficial, knowing where it's maybe clashing with some of your other efforts there.
But we're kind of going down that rabbit hole, I guess. We can start more high level of,
how are you using the platforms now?
How are you getting beyond that fledgling baby trap feeling
and really using it to optimize what you're doing?
Yeah, I think as always, it's evolving,
but I thought this was a clever idea we came up with the other day
to do an update on how we're using it today.
Our little subject line here in the title will be something to the effect
of mid-2024 update of how we're using AI today because I guess that's kind of the way I'm thinking about it, right? It's like,
this is going to change, this is going to evolve and ultimately that change in that, you know,
evolve is going to be something we could just talk about regularly because how we use it today is
probably going to be a little bit different than how we're using it, you know, in the future. So
here's the first distinction I would say, which is that I've been a little bit more of, you know,
on the teams are forming here a little bit, maybe a little bit.
I've been leveraging ChatGPT more and I think you've been using Gemini more.
So that's the first maybe place where we're differentiated just a little bit is that I'm
kind of leaning more into the ChatGPT platform.
No specific reason why.
It's just kind of been the thing that I've maybe tried first and I've gotten a little
bit more comfortable with it.
I've done some of the same searches, if you will, or queries, if you will, in Gemini and
seen great outputs there too.
So I don't think I don't have a strong opinion at this point,
like I do between, you know, like we talked about this
before, Ahrefs versus SEMrush, like that sort of thing,
where it's like, ah, I really do like the tool
that I have there.
I don't feel that way at the moment
with ChatDBT versus Gemini.
It's just kind of like, I'm used to it.
It seems to be producing good output.
They seem to be maybe innovating a little bit faster
is kind of my general feel at this moment in time,
but hey, that can change at any moment. So that's kind of one differentiation point
that I have here. I have gone ahead and upgraded to like the plus, you know, the premium or
whatever you want to call it type subscription. So that is a little bit of a something that
I think I don't think I had that last time at memory serves. Maybe that's a little bit
more of a more modern or newer change that I made, you know, on my side of things is
having the more premium version of it. So that's, that's another thing that kind of
comes in the fold on my side of things using like the more premium version of it. So that's another thing that kind of comes to the fold on my side of things,
using the premium or the plus version of it.
But yeah, to be honest with you, at this point,
I don't think it's replacing a ton of searches for me.
There's maybe one or two things that it has replaced.
But I'm using it a lot more like an assistant,
or almost like a junior level employee,
or it's like, I know what this is,
and I need a lot more of it.
Can you help me?
That sort of thing.
So my use cases are very work-driven. I really haven't delved into it at all from a personal perspective. I know what this is, and I need a lot more of it. Can you help me? That sort of thing. So my use cases are very work-driven.
I really haven't delved into it at all
from a personal perspective.
I know we talked about that.
We have a bullet point here about travel planning,
and that hasn't been the case for me at least.
But yeah, my use case so far has been some of those tasks
that we'll go through here in a second.
So maybe go back to that first bit about the tool
that you're using, because I think
you've leaned a little bit more into Gemini.
What's your reaction to that versus ChatTPT?
Do you have a preference? If so, why is your preference Gemini? Yeah, I mean, I think
initially some of the Gemini items were
Really it was it was kind of trying to start to
Replace and and then replicate some of those searches see what the searches that I had done in Google or
Kind of producing on the Gemini side of things
I think I mean we also have to think of there was barred in
there for a little while too, which I think was definitely a
bar that behind system like when it was chat GPT and barred,
chat GPT was way ahead, like miles ahead. So I think it was a
an item there of okay, once we've kind of enhanced the speed, speed is key.
I think we've both acknowledged that Gemini returns faster results.
Even I will say that right now I'm on one of the tests for 4.0 of JetGBT.
That's definitely speeding things up, but I still think from where the resources appear to be pulling.
And I think that's maybe the other side I prefer of Google
or of the Gemini side is that we are referencing back
to some type of search result still.
I know that ChatgyPT has gone in
and there is more of a citation of where you're getting that.
That's always been my concern with the LLMs is how, you know, where is it actually coming from? With content
creation, is it all just plagiarized? Are we all just plagiarizing each other now
with what by using chat GPT? But yeah, I think on the Google side of things, being
able to do some a little more image generation, the multimodal, the video
generation, and they've pulled back certainly because of some of the,
we'll say nuanced searches or nuanced images
that were being created and stuff like that.
I don't know, we don't need to go down that rabbit hole.
But ultimately, Google feeding the entire internet
as they see it into the system has allowed it to
have a larger model to learn from.
So I think that's maybe improving some of the results.
I also think there's some brevity there that is probably okay.
It's encouraging more interaction to get more information.
So you don't see Gemini going on these long paragraph, paragraph, paragraph, paragraph
responses.
It's usually pretty quick and hitting.
It's going to try to cite you back to something if it can, but there are some little nuances.
Overall preference, it's not a heavy one.
I mean, if I had to use ChatGPT, if I had the paid version, if I was using the paid
version consistently, I'm sure I'd get a lot more out of the GPTs over there because there
is some custom functionality
that's just not available on Gemini right now. I think it'll be there down the road, but
that's kind of where we stand right now. Well, we'll think of the companies themselves
for a second, right? How OpenAI got started was not from documenting and crawling and indexing
the web. It was a very different frame of mind where it's like, obviously, Google has got a 20 year head start and we're not almost probably close to 30 years on indexing
capturing information constantly, right? Like nonstop. That's at the essence of it. What
is Google probably the best out in the world, which is organizing the world's information.
That's like the original mission statement. If I had to bet, I would say like, who's going
to get better at that over the next five years? It's probably going to be Google over open
AI as far as crawling information, getting it quickly, verifying what's legitimate, what's not, et cetera,
et cetera. I think they're going to probably do a better job at that. It doesn't mean they're
going to make a better large language model tool to interact with. That's not a guarantee by any
stretch of the imagination. But as far as who's going to be able to get more accurate, quicker
indexing information, I think that's the case. I mean, think of how bad
ChatGPT was for some time where it's cut off. Date was like two years ago, right? So you'd ask about things that happened
in the even six months ago and it had no idea, you know,
it was completely unaware.
Now I think they've improved that significantly
as any product should gets better over time.
But I think you're bringing up a good point,
which is that one of these companies is a expert,
if you will, at creating, you know,
annexing information in Google.
And so therefore their product is always gonna be
a little bit stronger there.
I agree on your points a hundred percent here. Gemini does feel a little faster to me,
just using it day to day and like kind of interacting with that seems a little bit quicker,
no doubt about it. And it seems to be a little bit more respondent. When you click that button,
not only is the text print faster, it's like that initial kind of process of requesting,
it seems a little bit quicker. So yeah, I think where we're going, you know, going back to that
thread here in a second, you know, is that my kind of take on it today, like I think where we're going, you know, going back to that thread here in a second,
you know, is that my kind of take on it today, like the tools that we're using that
I think the meaningful question is what's where's your behavior changing? And I think for me, it's those types of things where it's like things that I was doing manually that I was having to type out,
like, I'll give you a good example. I don't know if you've done this, like 15 headline copies,
right? For, you know, Google responsive search ads, like, and now do that on chat, you'd be
to so much faster and quicker.
And even some copy or some smaller messaging
that I might have previously assigned to a junior copywriter
or a copywriter on my team, I can now do in ChatGBT
pretty quickly and much more efficiently.
And then I can just simply paste the text from one platform
to another.
I know there's a lot of people that believe,
and I can't disagree with the sentiment,
that that shouldn't be the case.
It should be that these A.I. tools are natively put
into whatever app or platform that you're using.
I can't disagree with that kind of thesis.
Why are we hopping between two different tabs?
We don't really do that frequently,
but it's kind of a new model, right?
I can't really compare it to anything else for the most part.
If anything, I think having the A.I. like the system level
seems like the most logical thing to me.
Microsoft will probably get there first, right?
But having a keyboard shortcut, or I could even envision like a button, you know,
like a button that's like looking at your screen. I think Microsoft has already hinted that they're
going to have some kind of screen watching function and it's going to be able to recall
what you were doing and things like that. I think that's probably more closer to what we
might end up with is that open AI would be like embedded into the Microsoft operating system,
or could my open AI or some version of that be embedded into like the Mac OS OS operating system where it's just like you're hitting a button or hitting a keystroke
or something like that and you're giving it a prompt and then it's doing something for you
based on that input or whatever it sees on the screen or that sort of thing. So that's my take
on it is that what I'm using it for today doesn't feel like necessarily what it's going to be
forever because it just is a little clunky, I would argue, even though it does save time,
it's a little clunky like the actual interaction between those interfaces. So I don't know your thoughts on that. But,
you know, go go to Iron Man. Like, I think that's where the taking into that next level,
and who's who's Tony Stark's little boy? Oh, boy. Oh, boy. This is
Jarvis. There we go. Thank you. Thank you. So yeah, I mean,
there we go. Thank you. Thank you. So yeah, I mean, someone's gonna fix. But I think like, that's, that's kind of what you have the long term. I don't know that that feels like the not so
futuristic idea or ideal for what AI should be able to do. I don't think anybody uses it on a
daily basis right now. I mean, you think of like some of the products like a Google Glass, like something
that like Google Glass as a as a wearable technology seemed like,
okay, let's build in some of this into maybe an augmented
reality or AI, generally, you know, that that type of thing. I
feel like we just never crossed the chasm. Even with Alexa and
Google Home and stuff like that. We haven't crossed the chasm into it really being a part of,
yeah, behavioral changes in your day-to-day life.
The idea of knowing your travel preferences based
on what you're seeing and based on what you're experiencing.
And that's probably going a little too deep even,
but it almost needs to be that ingrained.
I mean, we think about Google being ingrained
in everything we do thanks to the G Suite,
Gmail and all these other things and YouTube
and Google Drive and everything in the cloud
of Google's infrastructure can run our lives.
That is not the case generally with AI.
If it's really gonna have the impact where it is helping us to manage run our lives more efficiently and do things like that. I don't know, maybe that again, maybe that is too intrusive. Maybe that's taking technology too far. And I think that's, that's the conversation we're going to have more frequently about AI is how much do you want AI to run your life and like, maybe control and
understand your life? Yeah. Well, I think, you know, the comparison that I would play off of for our
listener is like, imagine there's a future where the AI is actually choosing the vacation rental
for the, you know, guests on the other side. It's not the guests going through and looking at 12
different options. It's the, it's the user or the potential guests on the other side, articulating
to the AI, Hey, go, go to every single website out there,
which you can do pretty much instantaneously.
Or in Google's case, they've already indexed all this information,
so it's in a database table somewhere.
Yeah, find me every single listing for vacation rentals on both Airbnb
and direct booking websites in North Myrtle Beach.
I want these specific parameters, three bedrooms, two baths.
I really want to spend less than 2,000 for the week if I could.
I'll go a little bit over it if it's something awesome, That's kind of what I'm looking for. I want it to
be within walking distance of X, Y, and Z. Basically, you could see a future where the
interaction for the user is a little bit less. Now, obviously, that's not very visual. So
I think that type of fictional scenario that I'm creating here is still something like,
okay, great. I've returned results to you and it's six listings or 10 listings that
fit your very specific search criteria. Hey, do you want to go ahead and book them? But it's bad. It's
interesting to me how bad text interfaces are something. So my wife and I are going to a
concert next week and we wanted to find a hotel close to the concert venue. And like you go on
Google and search like hotels near PNC arena or whatever. And it shows you stuff that's like,
you know, 30, 40, 50 minutes away. You know, it's not really that accurate. Whereas like,
I feel like the AI, if I prompted her properly and said,
only show me hotels within 20 minute drive of this particular stadium.
I feel like it's like that's that was the one thing my wife cared about was like
clean, comfortable. We weren't it's one night.
You know, the luxury there is not there.
We even looked on, you know, air, it'd be out of curiosity.
And we saw some interesting stuff, but it's like one night
just doesn't make much sense. Hotel's the way to go.
But yeah, it's like this is the one thing we care about, like nice, of some level of quality. And it
has to be as close as possible to the stadium. Of course, the
prices are very high and some are sold out because we waited
too long. But that's not the AI's fault. That's our fault.
But that actually does bring up an interesting point that I had
in our notes here. This was an idea that David Sachs had in the
Holland podcast about his like chat app that he's building
glue, where the AI would be promptless.
So in other words, right now it's like,
I think this is what you were getting at a minute ago.
I have to go into the AI and I have to ask it, tell it,
instruct it, detail what I want it to do.
And then it's like decent right now already,
and it's gonna only get better
at doing what I tell it to do.
Cool, but that's probably not how we want it.
The whole Jarvis bit, I think, is this idea of two things.
Number one, it's always listening,
or it's always context aware of what's going on. And that's kind of what, you know, Saks was talking about with
glue. It's like, it's looking at every message that's being sent. It's not inserting itself,
hopefully, you know, unless you need it. But then at any moment, you can, you know, go at, you know,
Paul or at, you know, Paul's AI agent and say like, Hey, this client, they their ads turned off, is
that the case? And then it would go to Google ads, look in the account, you know, would know some of
that information, and then return it to you. So I could see that kind of being the next thing that we cross over into where there's this AI
function that's, again, like I said, I think it could be lower system level. Microsoft would build
it, Mac would build it, Apple would build it, excuse me, that sort of thing. And then beyond
that, it would also have a layer where it's almost like it knows when to help you and it knows when
to shut up and just leave you alone. And of course, the inevitable thing is going to be,
you're trying to do something, the AI
is going to override you or tell you to put that file somewhere else because it's based
on the file name or something.
It's wrong and you're mad at it.
It's like, ah, I didn't do that right.
It's the classic clippy memes back from 2000 whatever era of Word.
So I think that's where we're going.
This is just my personal thoughts on where I think it could go.
I think travel planning seems like an obvious use case for the listener and something to think
about. How much information are you giving on your pages to where the AI could potentially benefit
you down the road if you can articulate things? Could written text and written information be a
little bit more valuable in the world of AI that can crawl that information and read it, but then
summarize it for someone that cares about it? Knowing that it's two miles from this hospital
may not matter for 99% of guests, but the 1% of guests who wants to stay at a home
near XYZ hospital, maybe, oh shoot, perfect.
Like if I didn't include that, it might not have been there.
So I could kind of make a case for some of those types
of things maybe down the road.
We're obviously still feeling this out.
We don't exactly know.
And in the meantime, we can go through some things
that we're doing today.
I know that's next in our outline,
but yeah, any thoughts there on those ideas?
What was exciting about one of those first GPT plugins or add-ons was that it was Expedia was
one of those first apps that kind of got loaded in there. And again, I really checked to see whether
or not that's being- Well, if you looked at the interface, the interface is bad, right? Like,
you're doing a search and then it's just returning. It's like returning a bunch of text, right? Like,
we don't shop for the issues with text.
You know, we shop visually, you know,
photography is a huge part of it,
reviews are a huge part of it, so I'm sorry to interrupt,
but yeah, the interface is great.
No, no, no, no, no, no.
So that's what GVT's got.
Well, Google is so tied in up, outside down.
I mean, inside out, upside down, and nots
about what they're doing with travel searches
because this is the one area where we did,
well, about a month ago, we started doing some of those,
checking to see what AI summaries
were for a trip planning process, kind of,
but those results just aren't great.
Like the Google is for, in our space,
for vacation rental searches,
it brings up meta search results. If you're going to do accommodations, anything like that,
it's going to start melding in some hotel and resort results in those vacation rental results.
If we just search for lodging and stuff like that, okay, now you start to see some other hotel results,
but there's no clear trip planning escape or path that
people can take on Google right now.
And AI summaries, that's the AI overviews is just something that I don't think they
can touch in this space for a long time because unless you're giving a trip planner, an online
travel agent of some kind, an AI travel agent. There's nothing better for Google
than to get those clicks through Google ads
or to do something like that or through the MinuteSearch
to get that money that way.
So I think this is one area where we are,
it feels like we are behind compared to every other industry
and how like the shopping, the e-commerce,
those feel like they're a little more or not tuned in,
but both of those are being represented.
Maybe it's because that's the clear path of, oh, I want to buy this product here, buy this
product there.
Travel because maybe it is a longer process or we can go down the pathway of why it's
happening, but their search presence, billions of dollars in ad spend have to be accounted
for there.
I don't know.
I feel like we need to come up to speed
because this seems like the low-hanging fruit
that is available of, yes,
we have a knowledge of travel preference.
We have all these things.
We have your search insights right here.
Let's use that to build you a better trip,
not just the stay, but the trip, the whole experience.
It can be more experiential,
but we're lagging in our space.
So yeah, I think that that's just something
that I do want to see more AI being built into Google Travel,
Google vacation, Google hotels, I don't care where it is,
Chad GPT, I don't want it to just be that add on,
I want it to be something else that we can do.
Yeah, no, I think that's a good example of where we don't
necessarily know where everything is going,
so we kind of have to make some logical adjustments.
And Google naturally wants to figure out
how they can make as much money as possible right from travel.
Travel is their bread and butter in many respects
when it comes to where their bread is buttered,
where a lot of earnings come from.
So there's something there.
Yeah, real quick, we'll put this in the show notes. But this May 2024 update, you're listening to this a little bit later. This was
published somewhat recently as of our recording of this. But yeah, Google just backtracking very
quickly on the AI summaries. We thought this was pretty funny. Read the title of that article.
I think the listener would like that. So the title of this particular article is about last week.
article is about last week. And it's just, this is the new, to be fair, this is Liz Reed, the new VP of search. I mean, this is a position that Jerry Dishler had kind of been the face of Google ads before
in Google search. This is Liz Reed's opportunity to kind of introduce her to the greater public. I don't
know. It's, have fun, Liz. There's, I think it was meant to be fun. I don't know. It's have fun, Liz.
There's I think it was meant to be fun. I think it was meant to be to kind of poke some fun at themselves to highlight that,
you know, people have been kind of goose in the system a little
bit as well. But what they're saying here is they're admitting
we we made a mistake, we overdid it. Even with the beta, getting the immediate beta
of probably 10x the number of searches they had in the entire eight month, 10 month, 12 month beta
before that. And then saying, yeah, it's not ready. And I think some of what I've seen, Maurice,
as BrightEdge does a lot of work with just kind of examining those SERP pages and to see like the in the initial week, I think 80%
of queries had AI overviews. Now we're down below 15 for below
15% for which you know, how many are actually generating AI
overviews and it's going to fluctuate day to day, you know,
as you learn but whoo.
I think it's super hard to track because you got to be signed in,
I believe.
So I wonder how they're even doing that.
Honestly, the tracking,
maybe they're like scraping screens
of people that are signed in,
which is a bit of a privacy concern.
But yeah, I mean, you know,
this is supposed to kill organic search.
And then it turns out like,
you know what's weird about it?
I guess this whole thing,
organic search turned out to be pretty good.
Like the way that organic search works,
showing the results based on their algorithm.
You know what I mean?
Like it's so funny to think that like Google kind of, I don Google kind of almost solved it, but Google had it working pretty tuned in,
like the end of 2022. And they've been endlessly messing with it since then and ended up with some
really strange odd results. I will say it a huge W with Google yesterday, and I'll give credit where
credit's due. The official support docs for the camera, I was debugging a little bit, a battery
issue that I was having, useless, right? Turn it off, turn it back on, you know, typical advice. The Reddit thread that
ranked very highly in Google for, hey, I'm having this battery issue, had the answer inside of it,
right? So like, I'll give Google some credit. Sometimes Reddit is actually the right answer.
And, you know, user whatever, you know, of course, these funny random Reddit usernames,
John Doe 123 will just say that that's not really what it was. But he was the guy that helped me
have the issue. I would not have figured it out. Otherwise, long story
short, I had to take out the camera or excuse me, take out the battery, turn on the camera,
which makes no sense. Of course, it's not gonna work. So I didn't think to try that
who would think to try that? Wait 10 seconds, turn it off, put the battery back in and turn
it on. So that somehow resets it and then it works again. And I was like, again, never,
never, never. It wasn't in the official documentation. I was going crazy. Thought the battery was bad, thought the camera was bad and like, yeah, red saved me.
So shout out Google for promoting reddit. That one particular search was clutch. The other ones have been a disaster.
Yeah, that feels a little bit. I'm just that feels a little bit like blowing in the Nintendo cartridge. That's all.
It doesn't feel like that honestly.
But even my wife was like, cause I was getting, you know,
I was kind of fronting and frustrated.
I was kind of after work hours.
Cause I got up at the end of the day and she's like,
what are you doing?
And I'm like, I can't get this to work.
I'm really frustrated.
And of course she tries what I tried.
I'm like, I already tried that.
You know, of course this is how you fight with your wife.
So yeah.
But anyways, yeah, she did that at one point to it,
you know, and then she, you know, put the battery back in.
I was like, sure, that's going to do it, you know?
And then it didn't do it.
And then Reddit saved me. So anyways. I was like, sure, that's going to do it. And then it didn't do it. And then Reddit saved me.
So anyways, I got some examples, by the way,
going back to the more practical pieces of ChatGPT.
We put this in our outline.
So I'll put a link at the show notes to this meta GPT
marketing co-pilot, custom GPT.
I think there's something to this.
I know Brooke, on your side of things,
and the Vittoria side, is playing with uploading his book
and uploading his transcripts into ChatGPT.
And then you ask the Brook AI questions.
And he was like, damn, it's pretty accurate.
That kind of stuff is cool.
But yeah, if you want to write ad copy or working on generating unique,
it's so hard to say, beautiful beachfront house.
How many times can we say that before we just go a little bit baddie,
a little bit crazy?
The answer is, I don't know, some number less than the number of times
I've said it, I guess, at this point when I'm writing copy for clients
or I'm seeing copy for clients, but that one's
been good, because you can give it a prompt, you give it a
landing page, or you give it the text from the landing page. And
you say, basically, write me a bunch of ads based on this. And
of course, like, this is kind of my feeling to with with AI, just
like a really, really junior employee, they give you some
horrible ideas that are like, that's stupid. We're not gonna
say that. But in there is like some nuggets that you can either
take and refine, or you can take and modify and kind of make
better, or you just take and you go, huh, I would not have used
that but it actually turns out to be the case. Or it's really hard for me to come up with
a bunch of unique combinations that are exactly 30 characters. But you know who's pretty good
at that? AI. Like I can give it an idea and it can actually come back with 10, 20, 30
ideas that are exactly 30 characters or, you know, 29 or something like that, which makes
the flow a little bit nicer. So I like that. But you had a different tool that was kind
of interesting that you're playing with this ad creative.ai tool.
Could you walk through this one? I'm curious. The headline of
the website is number one most AI tool for advertising, tool
for AI advertising, which I thought was funny because how
many tools could there possibly be? But your thoughts on that
tool? How's that working?
This so this is this is kind of taking the writing the ad copy
because that's that's a big part of it is certainly being able to
scale the amount of ad copy, the's that's a big part of it is certainly being able to scale the
amount of ad copy, you know, the assets that we can do. But then it kind of takes it a step further where as you're actually rolling out some of these assets, it's kind of reviewing the metrics on the
back end, reporting on the back end, and optimizing the next ad assets and the current ad assets to match what is really performing well.
So kind of, you know, talking about using that, not just the, the creation of the asset, but
optimizing based on performance of the asset. So really rolling the AI into a couple of different areas there.
But it's still certainly early testing that we've done on it, but I really like, I like, you know, the opportunity it gives you kind of the,
the potential to, admittedly, I, we've,
and we've talked about it, I'm not an ad designer
by any means, but it does allow you to be a little more
creative and then not just leave the creativity on that
initial draft, but iterations two and three and AB testing
opportunities and stuff like that.
I think it gives the opportunity for display is always going to be that the mystery meet
to a certain extent or something that's going to just contribute to the funnel, but not
be that high driver.
What I've seen initially is that we are seeing incrementally better performance on display.
So it's again,
early on early results, but I'm excited about the possibility of let's let the system kind of
do what it does evaluate. I don't know. Like when I look at the reporting for
by dynamic asset in a Facebook ad, I don't know if it's the face, if it's the headline that's
really doing it, or I don't know if it's the primary text, I don't know if it's the description, the image, what it is,
but that AI should be able to break it down behind the scenes
and give me a much better idea
so that we can build better ads moving forward.
And again, maybe turn display into a slightly higher converting channel
as opposed to a,
let's dig and dig and dig to find that value
in our attribution reporting there.
Yeah, we won't open that can of worms because we'll be here for two more hours and it won't
make any progress.
I had to dangle the worm at least.
Yeah, no, no, no.
Some fish that are smart, the big old fish, I'll swim by it and I won't bite on them because
I know there's a hook in there and I know it's going to hurt my jaw and I don't want
that.
No, but I do like that example quite a bit of like, that's the hard part at times,
right to come up with a bunch of different unique ideas. Again,
I was joking about a minute ago, but how often how often can I
say this or whatever? I think the other interesting piece of
that from a creative perspective is imagine you show 10
different messages to one person. Well, for person A, they
want to see the revenue messaging on the homeowner side
person D wants to see the guest care side of it or the you know,
property care side of it, right or whatever You know people have different motivations
So if I just keep showing you enough creative
Maybe I'll eventually find what you're actually into and then in theory like one would hope that I could get smarter and be like
Oh Paul cares about this issue. Let me tailor my website my copy my messaging etc to that issue
Which is then gonna make it a lot more productive
So yeah a little bit of speculation there in some respect
But we can see the direction where Google could go if they have a bunch of different creatives to build off of.
Whereas if they have one creative to work off of,
then it makes it a little bit harder, needless to say.
You're just giving them one card to play, so to speak,
which makes it a little bit trickier.
Right.
Awesome.
Well, let's write us out here, Paul.
We had this idea when we were putting together our outline
what we're not using AI for.
So last little bit about where we could go,
some things we're actually using.
We'll put some links in the show notes again
to some of these tools if you wanna check it out.
But how about this?
What are things that you're not using AI for?
I can lead us off here.
So I just put like general marketing copy.
So it's like the H1 of your website,
the homepage of your website,
the any piece that's, you know, important, basically.
I don't trust AI.
I mean, it's all important,
but I guess what I'm saying is AI produces very bland copy.
So if you want your home page to be like awesome vacation
rentals, then yeah, use Trap TVT and you'll
come out with something that'll be something similar
to that type of headline.
But if you want to say something meaningful,
if you want to have a unique point of view,
if you don't want something homogenized and bland,
then AI is not the way to go for you.
So I can't imagine a time when I would defer writing our monthly newsletter to AI, because
it's not a summary of things. It's like my thoughts about what's happening. It's an AI
doesn't know what I'm thinking. We're not there yet. Maybe we'll get there, but we're
not there yet. So I think there's so many things that AI just does a bad job of. I've
called out the LinkedIn people that use AI endlessly all the time too, because it's like,
it's not again, I try to post on LinkedIn when I have something to say, say something to share something meaningful, my
opinion, my take on it, etc. I don't want to take like, you know, Google ads can be a great way to
dynamically transform your business and get more traffic, which is kind of what like, chat GBT and
Gemini and the like spit out. So look, if you want the consensus, copy the consensus answer, go for
it. But there's so many things that I think if you're trying to say something meaningful, AI is the wrong way to go. And it's hard for me to explain it in one bucket,
but those are kind of the ideas that I have in my head of when I just would never reach for chat
to PTL. I think it is you can take it generally, but going into more of those personalized email,
it's another area where it seems like it's something yeah, let's just push out. I think
it's more often with workflows, like the cold email, certainly, but a workflow email, oh, we will just lead them right down the thing. No workflows already feel
kind of templatey and things like that. Adding AI and AI generated content into it. It's just one
of those things. It feels more impersonal. So you have to personalize the communication on that
email messaging overall, because they've given you they've given you a little bit. They've given you their personal email or their business email.
They've given you an opportunity to communicate with them, not necessarily market to them, but
communicate to them. So make sure you are doing the communicating because it will come in. And
that would be something I'd be interested in testing is having a larger subscriber list and kind of just A-B testing
it versus AI-generated content over a three-month period versus human-generated content and
what the unsubscribed rate or what the engagement rate, because I would hope that it would be
pretty equal, but I'm assuming, and that would be the, assume what we want, but that you
would.
You would have more engagement on that human written content. And you'd have
fewer unsubscribes coming from that from that AI generated or
from the human generated content. That's, that's just
another channel outside of just the general marketing company,
you need to be more personalized.
So I think I think the way to play off that is like, I want AI
doing my research, I think, right? I want to put in an URL
or an example is why I just actually tried it really quickly.
And Jim and I put in your Twitter URL. And it's like, tell me more about Paul, what he's into. And it's like, I think, right? I want to put in an URL or an example. So I just actually tried it really quickly. And Jim and I put in your Twitter URL. And it's like, tell me more about Paul,
what he's into. And it's like, I'm not able to really crawl that yet. Maybe they haven't figured
that out yet. But imagine that, right? Where I could put in the URL of like a guest or I can,
based on a guest, you know, past history or communication or chat with my team. You know,
imagine AI is transcribing every phone call, every text message, every message for a guest
that's arriving in one of your vacation rentals. And then you get it there and you go, I want to provide a meaningful gift to Paul, or Conrad,
as he comes on vacations with us here next week. Awesome. He's watching the Celtics,
send him a Celtics t-shirt and he'll love that for the finals here, which if you're
listening, hopefully they won. So I think that's the thing that I think we would want
to lean into. Yeah, to then click a button that says, oh yeah, go ahead and send that
out. That's where it's just going to be like,
I crawled your Twitter profile and it turns out
that you're a big fan of XYZ,
like it just becomes very creepy very quickly.
So I think there's a human touch element to that
that we're still in place for there.
I mentioned like general marketing copy,
you mentioned cold personalized or like emails,
like it's just really hard to like make that unique.
I would say for like new ideas.
So I kind of touched on this a second ago,
but we just finished a revamp
of our growth activation audit. And if you ask Google for, excuse me, chat GBT or Gemini
or something like that for advice on insert topic of choice here, how do I do an SEO audit? It gives
you like very generalized advice, but it's not that specific. It wouldn't tell you, it'll be like
review your site of broken links, but it doesn't tell you how to do it, right? There's such a wide
gap between what the AI returns and like the actual process, like step one, step two, step three,
step four, that two, step three,
step four, that sort of thing.
So I don't think we're really well suited
for that at the moment.
I think that honestly, you're a lot better off
taking some of these ideas that you have
and maybe helping, maybe getting a little bit
of scaffolding, excuse me, or a little bit of an outline
from ChatGBT or Gemini, but then actually filling in
the exact process yourself.
And again, treating it like a junior employee
or someone that's maybe helping you along the way, that's one way of doing it. And then you can kind of get the right process
in there. Yeah, I mean, I think ultimately, the way to kind of bring this home is to say like,
you know, chat GPT and the like are, they're limited, right? We don't have everything we want
at our fingertips. And ultimately, you know, the way that I'm looking at it and thinking about it
right now is, you know, like I've mentioned, kind of having that like lower level kind of person
helping me and I can ask you questions, it will respond to me, it will give me some different feedback and ideas,
and those are all very positive things.
So I'm looking forward to doing more with it, and I think what I said when I pitched
this episode idea to Paul, it was like, how do we go through and maybe do one of these
every few months or every six months, or whenever something notable happens or our workflows
change, give some different ideas, because these things are going to change pretty frequently,
and as they do, I think we'll try to leverage these tools
to the best of our ability, know what they're good at,
know what they're not good at, know their limitations,
and kind of build off that and go from there.
But yeah, I'm looking forward to it.
Any other parting thoughts here before we leave today
off of what tools we're using or how we're using AI here
in the middle of 2024?
I think probably if you just listen to the AI-related episodes that we've
done over the last year and a half, I think this is probably the one that we... Now we
know what we're doing with it. There's a little more confidence in dealing with the AI side
of things. I think that uncertainty is always going to be there, but I don't think there's
been any hesitation in diving in and learning and trying to use it more as much as we can.
It's just trying to use it in ways that are going to be beneficial and, and
knowing that there are going to be spaces where you don't want to use AI.
You still there, there is something to be said about using that human touch or
however you're using that or bringing on a person or doing anything like that.
So just knowing those places where you do want to use it, or you don't want to
use it and again, maybe hoping that there to use it or you don't want to use it. And again, maybe hoping that there are fewer places
that you don't want to use it moving forward
and that it is being ingrained more
and helping us do our jobs better
and more efficiently, hopefully too.
Yeah, right on.
Awesome, well, like I said,
we'll put some links in the show notes.
We'll have to wrap this one up for today, Paul,
but a good discussion.
One thing that doesn't take any sort of complicated AI or fancy large language learning models or anything new or fancy is leaving us a
review. So go to your podcast app of choice, iTunes, Spotify, listeners. You are where we get the most
downloads. Therefore, you're the most eligible. You're the most helpful person to leave us a
review, but any podcast app will take a review. We're not going to complain or whine or yammer
about it necessarily, but those are where we get the most downloads. So we super appreciate it.
Leave us five stars. Let us know what you thought of the episode. If you have questions, you can reach out to us on LinkedIn. Those are always where we get the most downloads. So we super appreciate it. Leave us five stars,
let us know what you thought of the episode.
If you have questions, you can reach out to us on LinkedIn.
Those are always linked up with the show notes.
You can send us a message there and you can email us
if you want to email us.
We're not really into that thing,
but if you are, that's cool with us too.
We appreciate you, dear listener,
for sticking with us today and going through all the way
to the end of the episode.
So I hope you have a wonderful day.
Fantastic.
Wednesday for your listening pleasure.
Thank you so much.
And we'll catch you next time on the Heads and Beds show.
Have an awesome day.