Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
3 Tricks That Make ChatGPT Agents Actually Useful
30 Plays13 minutes ago
Get your 1-month trial of High Level - https://danchez.com/highlevel
In this AI marketing podcast episode, Dan Sanchez dives deep into the newly released ChatGPT agents and how marketers can actually make them useful today. While many early users find the agents unreliable, Dan shares three specific tips to make them work more consistently in real-world marketing workflows. From SOPs and login best practices to using agents to trigger automations, he breaks down how to go from hype to helpful. He also covers ChatGPT's new study mode, a privacy warning about shared chats being indexed by Google, and explores the creative potential of AI-generated video ads.
My Favorite AI Tools
- Tella - https://danchez.com/tella
- High Level - https://danchez.com/highlevel
- Zencastr - https://danchez.com/podtools
- n8n - https://danchez.com/n8n
Resources Mentioned
- Simon Sinek on Diary of a CEO Podcast - https://www.youtube.com/watch?v=I3WUiD8HYn8
- LinkedIn Poll on AI guiding crucial conversations - https://www.linkedin.com/posts/digitalmarketingdan_danchez-activity-7353015750447489025-aEie?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAJq6gwB_q7tbdFevQk2MATQbXLntrfkInA
- Google Veo 3 - https://deepmind.google/models/veo/
- Viral Ikea ad - https://x.com/Salmaaboukarr/status/1946530149299634618
Timestamps
- 00:00 - ChatGPT Agents Are Here: First impressions and skepticism
- 01:16 - What AI agents really are and why they underperform
- 03:18 - Tip #1: Use step-by-step SOPs to guide agents effectively
- 06:24 - Tip #2: Login credentials and how to safely give agents access
- 09:22 - Tip #3: Bypass limitations using automation triggers via form fills
- 10:43 - Shared ChatGPT links are now indexed by Google (privacy alert)
- 12:13 - New ChatGPT Study Mode and how it helps with mastery
- 16:13 - Poll: Is using ChatGPT for hard conversations inauthentic?
- 21:29 - Viral AI Ad of the Week: Ikea explosion box made with VO3
Recommended
Transcript
Introduction to ChatGPT Agents
00:00:01
Speaker
ChatGPT agents are finally rolled out across all the Pro and the Plus accounts, and everyone's got access to them now, at least if they're playing for Plus. So what do we do with them? A lot of people have started doing initial tests and are saying, they're okay, but they break, they're not reliable, they don't really get the job done. We're going to have to wait for a different iteration.
Challenges and Improvements in Agent Reliability
00:00:21
Speaker
And I'm here to tell you that after I've tested them, I've actually found a few ways to make them more reliable. reliable and actually more useful, something that you can actually utilize in your business or in your marketing right now.
00:00:34
Speaker
So I want to get into those, but welcome back to the AI driven marketer, specifically bot bros, where I usually cover AI news with my brother, Travis Sanchez, but he is out this week. I'll let him tell you the story next week.
Current AI Trends and Limitations
00:00:46
Speaker
But today I'm going to be covering three pieces of news, including that one about agents and a few other tidbits that marketers need to know to get the help from the hype so that you can actually move forward with what's working now in AI for your job in marketing.
00:01:00
Speaker
So let's talk about those agents. I've seen it all over the place on Twitter, on YouTube, and a lot of people are saying it's broken. It doesn't work. But I knew it. I knew it when it came out. I knew it before I even got it that there would probably be some limitations around it.
00:01:14
Speaker
So
Agent Capabilities and Task Handling
00:01:15
Speaker
let's review real quick. What is an agent? An agent is an AI that can act autonomously. You can give it a broad objective and it will act independently with the tools that it has and the knowledge that it can gain and the personal information that you've given it in order to accomplish the task. It's like working with a high-end executive assistant that could just get it done. You could be like, ugh,
00:01:36
Speaker
Like, I need to go on a vacation. Can you make a, vacation find an opening in my calendar or make a vacation happen for me? Book it. I don't want to think about anything. You know what I like, just make it happen. And it goes and, you know, books the whole thing for you. Just, it comes back. It's like, Hey, you're going to Cabo. You like that.
00:01:50
Speaker
And I booked a excursion for you. You're going to be staying at this hotel on the water. It's pristine. I reviewed it. It's a good thing. All your meals are taken care of and you're going to be good to go. Here's your plane flight. In two weeks, you're going to Cabo, you know, whatever it might be. It just takes care of it for you. Now that's kind of the dream, but agents aren't quite at that level yet.
00:02:11
Speaker
So people are giving it these broad objectives like, Hey, book me a date night and pick out some clothes for me. and give it maybe a few other lines of descriptions, but not a lot. I mean, that's a pretty broad task.
00:02:22
Speaker
And then of course it's thinking about it and it is struggling with it Kind of like a human would. I mean, imagine if you got a new executive assistant that didn't know much about you, that person, even if they're pretty competent, would probably struggle to get that done too. They're like, uh,
00:02:35
Speaker
Like, well, I'm going to have to guess their size. Like I've seen them in a few pairs of clothes because we just met. So I don't know exactly what their style is yet or what they like or don't like, you know, like there's a lot to think through. There's a lot of ambiguity.
00:02:48
Speaker
And even though our chat GPT accounts are certainly getting to know us through the memory feature and through the the fact that I can remember a lot of the past chats, it's still not at the point where it can like think through all the details and get it done for you.
00:03:00
Speaker
So here's three tricks or three tips that I kind of want to offer that actually get it done and actually get it working, I'd say fairly reliably. The first one is they need step-by-step directions. Yes, chat GPT agents. If you give it a ah s SOP, a standard operating procedure of what to do, it's more likely to actually follow it through and get it done. They're very logical and it doesn't have to be complicated.
00:03:24
Speaker
I try to get mine to take a podcast transcript because it can't podcaster. I'm always trying to figure that out and make that this process whole smoother. I give it a podcast transcript, tell it to turn it into a post and then turn that post and then actually schedule social media posts for me by actually logging into my high level account plug for high level.
Integrating Agents with Business Tools
00:03:42
Speaker
Of course, if you haven't signed up for that, you should go to danchez.com slash high level because it's the best CRM marketing automation tool on the planet.
00:03:49
Speaker
Of course, this this is a sponsored statement, but I actually do use it and have been using it for years. so So plug there, but I wanted to log in to actually post through high level. Why? Because wanted to see if I could create a login for it and it could log in and it followed the prompt step by step. Got it done. Now I ran into a few other interesting things as I was trying to execute this process.
00:04:12
Speaker
I couldn't just vaguely say go post as me because it's going to try to log in. And if it doesn't have the credentials, well, I don't really want to give it access to my credentials quite yet for LinkedIn or whatever. it's just going to need a lot of information. So I did take the time to write step-by-step directions and its ability to hit those step-by-step directions really good. And if it ever went off the step-by-step directions, I realized it wasn't its fault. I was unclear in my directions.
00:04:36
Speaker
I gave it a vague statement like, oh, go check my email. And it would go and try to log into Gmail. i'm like, oh no, I need to tell it, go check my email through the Gmail integrator tool that you have access to.
00:04:47
Speaker
And then it would do it every time. So oftentimes we just need to create SOPs for it to follow and it could do quite well. It's a lot easier than setting up an automation tool and make or NAN because those take a lot of time. Those are actually fairly complicated and hard to set up.
00:05:02
Speaker
ChatGPT agent is literally just a bulleted list. Do these things in order, a numbered list. One, two, three, four, do these things. And it is logical, it thinks through it, it's like, oh, and I need to do the next thing. Now I need to do the next thing.
00:05:13
Speaker
And it follows through just like a genius intern would. In fact, that that metaphor we've been using since we've started with ChatGPT is still a really helpful metaphor.
Enhancing Agent Functionality
00:05:21
Speaker
I find that the best training I've ever had for AI is not in my technical background, but is in my internship background. I used to run two different internships, and I had to train...
00:05:29
Speaker
students how to do fairly complicated marketing tasks like write SEO blog posts that would actually not just post but rank on Google. And we were very successful at that. But I had to train a number of them how to do it, how to revise it, how to do it in a systematic way so that I know I could train a freshman on how to write blog posts that rank. And we we did that.
00:05:47
Speaker
Yeah, he's very similar. You have tried and true processes. You just need to train them in the processes. Don't do the whole thing. Try little chunks at a time. It's I find that it's best to experiment and be like, oh, can it do this? I'll test the little thing at a time and then build it and make it longer and longer and more complex and more rich and have it analyze some things along the way.
00:06:03
Speaker
That's where chat GPT agents are really going to shine. It's early days now. They'll get better and more sophisticated as time goes. so So that was tip one, give it step-by-step directions or an SOP or something to follow.
00:06:15
Speaker
You don't have to spell it out to the nth degree, but the more clarity you give it, the better it performs. Tip two, They can log in as you you, if you give them the credentials, they will use them and log in as you. This is kind of scary because and like if it logs into your social media accounts, like it it can go off the rails, it can do things. There are potential ways that bad people can hijack your agent as it's trying to figure things out and take over it. So you want to be careful of it and kind of limit the scope.
00:06:42
Speaker
That's why I would never give it credentials for my LinkedIn account or my other social media accounts. Instead, I gave it access again to high level, for it to log in there where it had, and I even, even in its high level login, all it had access to is the social media posting tool. And all it could do is post.
00:06:55
Speaker
That's it. And schedule posts, right? I want, I was okay, I was okay with it posting, but I didn't want to go change and comment and i and do anything else in my social media file. That's dangerous. That, and if the the link got hacked, then like a hacker could only do so much with it because it's had limited reach or scope of what it could do.
00:07:13
Speaker
So I gave it its own email and its own login. It went and logged in with that. Funny thing is, is it had to use two factor authentication, but that's where all these connectors are really cool because can give it its own email. It would have to send a two factor authentication to my Gmail account because this email I created for just forwards all the emails to my, my, my real email.
00:07:32
Speaker
And so it logged into my Gmail, went and found the authentication code after it tried to log in and then logged in. I'm like, Ooh, this thing's pretty good. Like it can get around stuff. If you just say, Hey, go here, do that authentication code. We'll come to this Gmail. It'll be coming from this email with this six digit code and it can do that. It's very cool. And it was able to log in and do that for me.
00:07:52
Speaker
Now I want to run into tip three. So that's tip two. Tip two is Be careful what you give it. If you do give it login credentials, think about giving it login credentials that are not your account, but a account you made just for the AI, knowing if it goes off the wall, like it's in a, it's in a kind of contained place. And if the credentials leak out onto the internet, somehow it's not, you're not in trouble, right? It's, it's, it's its own contained account. It's an account just for your AI assistant, just for chat GPT.
00:08:21
Speaker
I think we limit a lot of the risks that way and we can manage it a lot better. If it does get compromised, we can shut that one down, spin up a new one. Easy peasy. So that's tip two. Tip three is sometimes these things have limitations. They're purposefully baked into it to protect you. But sometimes, yeah, you you kind of want to get around them.
00:08:37
Speaker
I wanted it to post as me. I wanted it to go and just, I wanted i did a test. I was like, hey, don't post to LinkedIn. Post to my Facebook page, which nobody ever sees or nobody knows about really. And I just want you to post the line, a bot posted this for me. And I want you to post it and tell me when it's done and then I'll go check it on the Facebook page. And it's like, no, we can't do that. It always checked to see like, hey, before I push publish, I need you to look at this high level page to make sure it's all good before I hit go.
00:09:01
Speaker
And it always checks. I'm like, I don't want you to check. I want to, I want this to be done. I, I trust you go for it. I wanted to see how far I could go, but it would refuse to do it. So I'm like, okay, here's, ah here's where the third tip comes in.
00:09:12
Speaker
You can actually trigger automations by having it fill out forms. And then those forms can do whatever you want it to do. So maybe i put the social post into a form and that form triggers an automation that then posts to whatever social media posts I wanted it would be a way to get around that thing.
00:09:28
Speaker
And it got me to thinking like these forms can trigger all kinds of things. It's actually an interesting way to build an agent that triggers normal automations or other agents by form fills because a form can trigger all kinds of automations, whether you're hitting it with Zapier or even like an Asana project can be kicked off with the form fill, right?
00:09:47
Speaker
There's Asana forms and you fill it out and then it gets plugged into Asana. That's kind of an interesting thing because you can have an a chat GPT agent do research, compile a list, do some stuff, and then fill out a form for you that gets all the information that it found into the right next step and triggers whatever workflow comes next.
00:10:03
Speaker
That's kind of an interesting application that opens up a whole bunch of possibilities that I'm looking forward to getting into.
Future of ChatGPT Agents
00:10:08
Speaker
These are just kind of my initial findings with agents that I'm like, these things are actually fairly useful now. And of course, hopefully they continue to do iterative improvements over the year. If it gets as good as 4.0 has gotten, which 4.0 sucked this time last year in comparison to what we have now, which is that one ah model from ChatGPT. So hopefully they make little improvements to it throughout the way and gets a little better, a little faster, a little better, a little faster throughout the way. And then these things will freaking rock.
00:10:32
Speaker
so So that's it on the news on chat GPT agents.
Privacy Concerns and Indexing of Shared Conversations
00:10:36
Speaker
They're all rolled out. They are helpful. Just use my three tips when building something with them or testing them for yourself. And the other news, chat GPT, this is a lot of chat GPT news today, has announced that a lot of their shared conversations, like when you go to a conversation you've had with che chat GPT and go to share it to share the link somebody with somebody, those are now indexed by Google.
00:10:56
Speaker
Yes, they're showing up in Google search results. This is important because some of us have had conversations about private things ish, and then we've shared them with other people that you didn't realize were now going to be on Google.
00:11:08
Speaker
So someone can search like someone could search Dan Sanchez and maybe like try to just hack at like different vulnerable topics. Maybe that I'm talking about like business insights, AI driven marketer analysis or whatever.
00:11:21
Speaker
Not that I'd be afraid to share that, but like there's probably could be things, all the things that I did have, I've taken the advice that I'm now going to give you is to go and into your settings under data controls, under shared links and and like remove those from being shared.
00:11:35
Speaker
Because some of you don't realize that you've shared maybe things that you only meant to share with one person, maybe a small group of people, but now it's going to be available for the whole wide world to find information. If you don't go and unshare them.
Educational Applications of ChatGPT
00:11:46
Speaker
So this is just a quick alert, a quick warning, like go back to your chat, GPD account, go review what you've shared before and turn them off because every marketer should be concerned about this. If you've been using that share this chat feature, because.
00:12:00
Speaker
yeah indexable so that's news two news uh the third news item i have is chat gpt study mode got launched this week and while it's not as like as impactful as chat gpt agents this is actually still really cool this feature was aimed of course at education and teaching students but i'm like aren't we all learners here aren't we all trying to learn new things all the time i know i am i have a whole bookshelf behind me of all the books that I'm trying to read and keep on top of.
00:12:27
Speaker
And honestly, I think we can all be honest about the fact that we've learned something in the past that we're like, oh, this is so good. Like I'm going to learn like this is the coolest thing ever. And then we we kind of learned it loosely, but we never really like mastered the topic.
00:12:41
Speaker
And maybe there's something like that for you. Maybe it's marketing positioning, something that's fairly abstract and fairly difficult, but we all know is like really important. Maybe it's copywriting. Maybe it's doing analysis on landing page conversions, whatever it might be. There's probably a subtopic of marketing that you're like, oh, this is so important. This is so good. And then we didn't actually like dive really deep into it.
00:13:00
Speaker
This is something that chat GPT study mode, which is kind of by the agents under the ah tool section. When you go to enter your prompt, there's like a little tool button underneath is right there. We can now use this to our advantage.
00:13:11
Speaker
I did it the other day just to test it. I was like, Hey, I want to learn more about Alex Hermosi's hundred million dollars offers specifically how he creates value or his value equation. Now I'm actually pretty familiar with it. I've worked with it a lot.
00:13:23
Speaker
It's a fantastic tool, but I was testing it to see like how it would teach me more. And it does a really good job. It uses something called the Socratic method where it leads by asking you questions instead of giving you the answer. It just asks you questions to pull it out of you. And that's a very helpful experience. it's like working with a personal tutor and can cause you to go deeper because once it tested me and understood that I actually knew and could explain in my own words what all the points of the value equation were that Alex Ramosi created,
00:13:51
Speaker
they started testing me by giving me scenarios to apply it to. Like, well, what would you do in this situation? Hey, based on previous conversations, you're working on this thing. How could you apply it here? How could you apply it here?
00:14:01
Speaker
How could you tighten it up here? And it would work with me on it and force me to really think and wrestle with the concept. This is really good because we could be doing this. for all the things that we know we need to dig deeper into, but really take the time to do.
00:14:12
Speaker
This way, we don't have to like, we can essentially react to the teaching. That's the cool part of a tutor is if we're teaching ourselves, we have to kind of like tutor ourselves, which is a lot harder. it takes a lot more thinking and effort to be able to like tutor ourselves in the program and really come up with our own exercises that we then complete. But to have someone pitch you ah the questions, we can only think about the topic and not think about how to...
00:14:33
Speaker
train ourselves in the topic. It's kind of like personal training. It's a lot harder to go to YouTube for physical exercises, even though you're like balancing out, like, do I need to be doing this one? How much is that one right for me versus a personal trainer? It's like, okay, now do this much.
00:14:45
Speaker
Okay. Now do this exercise. No more harder. Good. Now do this one. It takes a lot of the thinking out of the teaching. Because otherwise, if we're teaching ourselves, we have to think about the teaching and the topic. In this case, we can just think about the topic, which is what we really want.
00:15:00
Speaker
And it's not bad to teach yourself and all that kind of stuff. I'm just saying that you can go a lot farther faster if you don't have to also think about the teaching. You could just think about the thing that you're trying to master, which is what a good tutor does for you or a personal trainer does for you if your goal is to work out and get stronger or lose weight or something.
00:15:17
Speaker
It takes the thinking out of it and just gets you to the doing. In this case, you're learning the mastery of whatever the thing you're doing. And of course, this applies to more than just marketing. You learn master anything, whether you want to become a better parent, a better spouse, a better, I don't know, you want to get better at a hobby you're going after, like learning something new, whatever it might be. This is a great way to learn it. But of course, it applies to marketing because marketing, we always have to learn more and more.
00:15:42
Speaker
It is a never ending loop of learning, which is one of the things I love about this profession is that there's always more to learn. You can never go deep enough. I've even taken some friends that had mastered very complex topics like chemistry and they find they never get bored of marketing because marketing's always got something new to learn.
00:16:01
Speaker
So moving on from there.
AI in Marketing and Advertising
00:16:04
Speaker
I want to talk about the poll of the week. I did a poll of probably two weeks ago. And the question for the poll was this is using chat GPT to navigate hard conversations inauthentic and why?
00:16:18
Speaker
The reason why I asked that question is because Simon Sinek had a great interview over at, I can't remember the name of the podcast. That's one of those big podcasts. I could picture the host, but I can't remember the name of it, but it'll come back to me.
00:16:30
Speaker
But anyway, he was having a conversation with AI saying like, oh, if ChatGPT is coaching you on your crucial conversations, that's inauthentic. Imagine if a husband came away from a spat with his wife and then had a conversation with ChatGPT to get clarity on the conversation and come back with the right answer.
00:16:47
Speaker
And then he gave his right answer and the wife turned to him and said, to chat GPT come up with this for you? And the husband had to admit, yes, how do you think that wife's going to fill? That was the point that Simon Sinek met. And I was like, you know what?
00:16:58
Speaker
As long as that husband actually believes it, agrees with it, realizes his points of fault and realizes like the story he's telling himself to see, but not just the story, but sees her story and actually comes to a ah resolution about it truly. Then I'm like,
00:17:13
Speaker
I feel like that's fair game. Like, that's good. Why would anybody be upset with that? I hope my wife does that for me. Like, and I hope I, people do this because I, that's why we go to trusted counsel, hopefully not just to agree with us and affirm our story, whatever it might be, but to help us through it And this isn't a marriage podcast, but oftentimes I find that crucial conversations, a lot of the best things in life are on the other side of crucial conversations. And we don't have them hard conversations all the time.
00:17:40
Speaker
with our boss, with our peers, and certainly with our and employees. Oftentimes we're holding back on things for lack of having that hard conversation. And this holds a ton of marketers back because we often don't work independently.
00:17:55
Speaker
And even if we do, We have clients where we have to have kind of hard conversations, right? This is an important life skill and it applies to marketers completely. i know it's not related to marketing and it's not related to AI, but ChatGPT specifically, and and a lot of the AI tools too tools can do this, but they can help coach you on having those hard conversations. So oftentimes I just go into ChatGPT, I'm like, hey, I'm running into this issue with this person.
00:18:20
Speaker
Here's kind of the the backstory and what I'm running into. What I want to have happen is something like this. Help me understand what's going on with the friction here. This is what I know about this person. You know me and my personality and my defaults.
00:18:34
Speaker
help me find a resume Help me try to find a resolution for this conversation or prep me for a hard conversation with them so we can actually hit the goal that I want here. ChatGPT will do a fantastic job of helping you.
00:18:47
Speaker
Now, warning, I will say, while it's a great tool at coaching you through crucial conversations, there is a dark side to AI that you have to be very careful of. It is a confirmation bias machine, and it will often lead that that bias. So if you're like, oh, this person's so bad, whatever, like it will just assume the person's bad, but you don't want it to. You want it to actually push back on you more than the person. Assume you're at fault and need to overcome some things.
00:19:09
Speaker
It's always a better way to approach a crucial conversation that way. So I even tell ChatGPT, hey, find out where I'm wrong here and give me a feedback as if it's up to me to make this happen and not the other person.
00:19:20
Speaker
Much better approach because ChatGPT will give you a million and one reasons logically why you're always right and you're not always right. Oftentimes we're wrong. So good podcast to go listen to. I'll link to it in the show notes is the dark side of AI. It's something I did with Travis just a few months ago, because you need to understand this in order to truly master AI, because AI does have a dark side and it's leading a lot of people astray in person in their personal lives. So just be careful when you're having personal conversations with chat GPT. They're great, but
00:19:52
Speaker
They can have major consequences if you're not familiar with how it can lead you down the wrong path. So in the poll, the swing back to that poll, I asked people again, is using ChatGPT to navigate hard conversations and and inauthentic? Why? And of course, I have a lot of ai fanatics that follow me.
00:20:11
Speaker
A lot of AI fans out there. So of course, 61% said no, 20% said sometimes, 6% said mostly, and 9% said absolutely. And of course, a a measly 4% said, I'm not sure.
00:20:24
Speaker
So people have some pretty strong opinions about it. The comments to that one got really fun. I'll link to that in the show notes if you want to see what the dialogue around that was. But generally, most people are like, No, have the hard conversations or chat GPT first or work out the kinks before you go to the person.
00:20:40
Speaker
And as long as that person believes what they're saying, and they're not just saying it because it's a cheap hack, kind of like a three-year-old or more like a seven-year-old. have have a bunch of different four kids and they're spanned all the ages, but a seven-year-old will sometimes try to tell you what you want to hear, but you know, inside because they can't hide it in their face yet that it's not true. They're like,
00:21:03
Speaker
I'm sorry. You're like, are you really? They don't really believe it. But if you really believe what ChatGPG is giving you is like the path to resolution that I think it counts. I think that's that's the message we're walking away with.
00:21:15
Speaker
All right. Viral post of the week is actually a lot of viral posts because keep seeing it all over x over and over again, is this one AI video someone made of like, it's an it's essentially a commercial for Ikea, but it's fantastic made with AI. And I'm like, man, this is great. Like if we could do more like this, that's just a great idea. So just just could do search for it. I'll link to it and show it again.
00:21:37
Speaker
of Ikea blocks explosion. It's essentially a blank, a bare room, hardwood floor, nothing on the walls, and just one big box in the middle with the words Ikea on it with its logo, right? And then the box kind of opens slowly and then explodes and a whole furniture set comes in like it it's already assembled.
00:21:54
Speaker
But it's just kind of like this fantastic ad. It's and then a little over exaggeration, right? So you you got to assemble that stuff. But it just goes to show you that one visual image that would have been really difficult to create this, this, this video, if you were making it with CG or if you were trying to make it so like they were literally exploding out, it would have be really hard to make.
00:22:17
Speaker
It's probably a CG thing, right? It would have been really hard to make this. But with a $300 subscription to VO3, this ad is actually pretty easy. Of course, you're probably gonna multiple iterations to get it right. You might have to go into some video tool and Photoshop on or like, you know, map on the IKEA logo onto the box if it doesn't quite get the logo right.
00:22:37
Speaker
So there's probably some work here, but it turns it into what would normally take a $10,000 creative to come up with this fantastic video ad and is now done with maybe a few hundred dollars.
00:22:49
Speaker
That's just amazing that we can make ads like this. Now, if I were in the ad space, like running a lot of like Facebook ads specifically or video YouTube ads, I would be pounding this to come up with creative over and over and over again, because the sky's the limit.
00:23:03
Speaker
If you wanted to make it look like a Hollywood trailer in a 10 second video, you could like, there's nothing stopping even a 30 second video. If you can put multiple multiple multiple of these together, the sky's the limit now when it comes to AI video.
00:23:17
Speaker
And right now it's only good for like short form. And people are pushing the envelopes of course, of like real content, movie style things, even movie trailers. so We'll see where that goes over the next couple of years. so But right now it where it's at, it's making fantastic ad
Podcast Conclusion and Listener Engagement
00:23:31
Speaker
creative. If I were doing ad creative, I would be spending a I would definitely be giving my money to Google for VO3 and I'd be spending a ton of time there. It would be really worth it. So go check out that viral post. Link is in the comments. Thank you for joining me for the AI Marketer this week.