Speaker A
So, in this video right here, I'll be showing you guys exactly how this store right here makes around 300k a month in revenue all by using C Dance 2 and Claw to actually automate its ad creators for the store. That's right, this entire
Learn how Andrew Yu uses Seedance 2 to fully automate AI UGC ads, boosting e-commerce revenue with hyperrealistic, high-converting AI-generated content.
Key Takeaways
- Seedance 2 enables full automation of AI-generated UGC ads that outperform traditional ads.
- AI ads can be customized in every aspect, including script, accent, and product presentation.
- Using AI for product research and validation reduces the risk of testing poor products.
- Hyperrealistic AI ads can generate significant revenue without viewers suspecting AI involvement.
- Proper AI workflows and tools like Arc Ads make creating high-converting ads accessible.
Summary
- Andrew Yu demonstrates how his store generates $300k monthly using Seedance 2 and AI to automate ad creation.
- The video explains replacing Sora 2 Pro with Seedance 2 for automating AI workflows to clone and improve competitors' winning ads.
- Andrew shares AI-driven automations for product research, validation, and ad creative generation used by his students.
- He highlights the advantages of AI UGC ads over traditional UGC, including full control over accents, scripts, and product showcasing.
- The AI ads are hyperrealistic and high-converting, often indistinguishable from human-created ads by viewers.
- Andrew addresses misconceptions about AI ads being low quality or 'AI slop' often seen on social media.
- He explains how AI UGC creators can be perfected for authority and messaging to improve conversion rates.
- Seedance 2 combines realism, lip-syncing, and motion to create effective AI ads.
- The video includes a walkthrough of Andrew’s workflow on Arc Ads platform, showing how to automate ad creatives.
- Andrew offers one-on-one coaching and access to these AI automations for viewers interested in replicating his success.
Full Transcript — Download SRT & Markdown
Speaker A
store's ad creators are currently all automated. All 100 plus batches of ads that I have right now are all using AI.
Speaker A
Now, I know that Sora 2 Pro is actually going to be going away soon, so I want to show you exactly how I would use C Dance 2 to go ahead and actually replace it. So, today you'll be able to go ahead
Speaker A
and see my exact AI workflows that allow me to clone my competitors' winning ads, and we'll even use AI as a creative strategist to come up with new iterations of those ads automatically.
Speaker A
And even as a special credit, I'll even show you guys exactly how to make all of this into an entire interactive software program that you can actually use with UIs and all that fancy stuff, and the entire code base is actually going to be
Speaker A
rewritten or actually written with AI as well. So, by the end of today's video, you should have no reason why you can't be making AI ads for your stores. And oftentimes, these ads are actually better than these UGC creators just
Speaker A
because you have the power of AI to replace, to change up, or to even improve on anything that you don't like.
Speaker A
And this is also the same exact method that I currently use with my students as well. Students that are pulling in 20k in a single day, 16k in a single day, or even around 17k in a single day, or
Speaker A
around 128k in a single month. Now, if you scroll down right here, here are the automations that are currently helping my students actually find the products, validate the products, and even make the ad creatives for their products. We have
Speaker A
an AI product research automation right here as you can see on your screen that actually finds products for you. Now, I know that not every single product is going to be a winning product, that's why it even goes through
Speaker A
the AI product validator automation to make sure that basically tells you, "Hey, these products are good or these products are bad based on the parameters that we give you as well." So, we make sure you're not testing bad products.
Speaker A
And once you find your winning product, we provide you with an AI ad cloner automation that will basically help you copy your competitors' winning ads and pretty much just clone it. And even so, if you want to make other iterations to
Speaker A
the winning ads, we even have an AI creative strategist workflow that will help you do that as well. As you can see right here, the workflow that will help you recreate winning iterations of your competitors' winning ads. So, if you
Speaker A
want to work with me one-on-one to get results like this and get access to these automations, go to the first link down below. Now, a lot of people are going to see results like that and assume that there's no way that AI ads
Speaker A
are going to go ahead and actually have high-converting ads to bring in like 300k revenue per month, even though I just showed you guys the proof that it does. The issue is because a lot of people get their idea and their
Speaker A
information on what AI ads are from actual AI slop. People are going to assume that our ads are going to look like AI slop because that's what they've been consuming, and that's what they think AI actually is. AI slop comes from Reels,
Speaker A
comes from TikToks, and a lot of times these Reels and TikToks are supposed to be AI slop because they're supposed to be for entertainment. They're not there to convince you whether or not these ads are AI or not, they're literally there
Speaker A
just for entertainment, for views, and for attention only. While the AI ads that we run are actually hyperrealistic, and they're also high-converting. Now, the funny thing is, a lot of people may have already watched AI ads and not even
Speaker A
realize that these ads are actually AI. And none of my comments have anyone ever mentioned, "Hey, is this AI?" Because we're not trying to convince them that these ads are AI or not. We're not. We're no longer trying to do that. We're trying
Speaker A
to convince them, "Is the video that they're watching showing the product enough to solve the problem that they're currently having in their life?" So, they're no longer questioning whether or not the video's AI, they're trying to question, "Hey, does this product solve
Speaker A
my problem?" And the cool thing about AI ads is that you can actually go ahead and control every single thing about an AI ad. So, for example, with a traditional UGC ad, right? For example, if the content creator has an accent,
Speaker A
well, guess what? You can't really change their accent. If the content creator doesn't show off the product enough, or show off the product in a way that you like it, guess what? You can't really change that after you get the
Speaker A
final video. Or if the UGC creator just goes off script, yes, you can cut off the points where they go off script, but it won't be as natural-sounding or as fluid as you would like it. So, a lot
Speaker A
of these things actually can't be fixed with a traditional UGC system. But, if you have a UGC creator that's AI, you can change whatever you want. Uh, they have an accent that you don't like?
Speaker A
Change it to the accent that makes more sense for your messaging. Uh, they don't really show off the product that much?
Speaker A
Guess what? You can go ahead and prompt to show off the product however you want the product to be showed off. And you'll never go off script because you are the ones giving them the script. So, everything could be perfected here, even
Speaker A
the UGC creator could be perfected. A lot of times, when it comes to traditional UGC marketing, the UGC creator sometimes doesn't speak from a point of authority. Like, for example, if you have a person that's kind of
Speaker A
chubby like me, and I'm trying to tell you, "Hey, yo, this workout is great," you're not going to believe me because I'm not coming from a sense of authority. Unless you say that the UGC person doesn't have the right messaging,
Speaker A
right? Like, the way they're speaking isn't actually, you know, converting for the customers that are watching your video, you can now actually go ahead and fix the messaging as well. I couldn't spell messaging here, but I'm sure you
Speaker A
guys know what I mean. So, all of that now is actually fixed with AI UGC, and it's all powered by powerful models like C Dance 2 that's able to combine all the realism, all the lip-syncing. Every single motion now with C Dance 2 is
Speaker A
actually really, really high-converting as well. And like I said, all the ads that I'm running, no one even bothers asking, "Is this AI or not?" because they don't even come across their head that these are AI ads. So, if you're
Speaker A
wondering how to actually access these AI tools and these AI models that create these super hyperrealistic ads, I do recommend using Arc Ads. There should be a link down below. As you guys know, C Dance 2.0 is actually not out in the
Speaker A
United States, and it's not available for use. But, on Arc Ads, you can actually use their model right here. As you can see, C Dance 2 is now live. So, all you want to do is click login right
Speaker A
here. And now I'm in right here, you can see my workflow right here. Now, I know this looks super complicated. I am telling you right now, it is not that complicated. I'll be going over step by step exactly how you can go ahead and
Speaker A
recreate this workflow for yourself, but this is the workflow that I use that helps me automate all my ad creatives for all of my stores. And really quickly, I can go through a result with you just so you can see that these ads
Speaker A
actually produce really good-looking ads. So, for example, I'm going to click play right here.
Speaker A
Okay, ignore my messy room, but I have to show you guys I found. It looks kind of weird, but look at these side wings.
Speaker A
If you're a side sleeper or wake up with a stiff neck, this center dip literally cradles your head and... All right, there you go. Obviously, there are some flaws with the video. I'm not saying it's a perfect video, but numb
Speaker A
of people are not going to notice it like you do. And number two, a lot of times you can either re-prompt or regenerate it if this is not a perfect version of the video. So, if you ever wanted to kind of do a little bit of
Speaker A
edits, you definitely could go ahead cuz look at this entire workflow, there's a bunch of things that I did in order to create that entire video. Now, first things first, if you want to scroll all the way to the beginning, we have these
Speaker A
things called nodes. If you're completely new to AI, or completely new to workflows, I get that this looks super messy, but once you understand it, trust me, it is not that messy. So, like I said, first things first, you have
Speaker A
these boxes right here. These boxes are called nodes. And in order to actually go ahead and make a node, you right-click and just choose the node that you want. Right here should be pretty self-explanatory, but for example, let's just say you want to go
Speaker A
ahead and create an image right here. So, for example, Nana Banana Pro, I can go ahead and right-click here, click image, and you can see you can go ahead and write a prompt to go ahead and create an image. Now, for this example
Speaker A
right here, I will show you how to do it step by step. As you can see right here, I have an image right here of my competitor. My competitor is called Melo. And all I got to do is right-click
Speaker A
here and click upload. I'm going to upload a picture of my competitor. And once I do that, all you want to do is connect this box right here to this box and kind of write your prompt in on what
Speaker A
you want to do with your competitor's image. As you can see right here, that's pretty much all I have to do in order to go ahead and recreate my competitor's ads. All I really did was upload a picture of my competitor's image right
Speaker A
here, and I clicked rebrand my product to Cloudora. And this node right here using Nana Banana Pro is going to go ahead and recreate that, and now my logo is on here, Cloud Cloudora. I never gave the image of my logo, I never said
Speaker A
anything about my products, but now it knows that it's going to go ahead and rebrand this entire image to my products. Now, once you get Nana Banana Pro with the product and your logo, you're going to want to split it off to
Speaker A
a bunch of different nodes. And the cool thing is, inside these actual nodes are LLMs themselves. So, imagine like ChatGPT or Gemini, where you go ahead and write in a prompt, and the answer for the prompt will actually show up in
Speaker A
the actual node itself. And it's cool because you can use it in a workflow now because whatever shows up in this actual LLM node, you connect it to something else as you can do right here. So, you don't have to do a bunch of back and
Speaker A
forth with your LLM in another tab and come back to this tab and back to another tab, you can actually integrate it within your workflow. Now, you're probably wondering, "Okay, why are we splitting off a bunch of different ways
Speaker A
right here?" And that's because we want to test a bunch of different formats. As you can see, this format right here is going to be a UGC style ad right here.
Speaker A
So, a ad UGC style video, very organic-looking. If you scroll all the way down right here, this ad is specifically for a podcast ad. As you can see, podcast studio. As going down right here, this ad is really
Speaker A
professional, looks like it was shot in the studio, like a very professional ad. As going down here, this format is like a little TikTok format right here. If you scroll all the way to here, it literally says like, "Make an ad that
Speaker A
seems like not an ad, but an actual TikTok trend video." And I'll show you guys a little bit of the video right here.
Speaker A
[music] Which is insane because number one, the music I did not put whatsoever. Number two, these captions I did not write. I did not tell that like I didn't do anything here. The LLM was the one that figured out how the video should be
Speaker A
like, and the dance moves that I never told the person exactly what to do, and it's all powered by C Dance 2. Like I'm telling you, this is a very powerful model. It knows exactly what you're trying to generate. So, you're probably
Speaker A
wondering, "Okay, Andrew, how did you figure out all of these formats, and how do you know these formats actually work for your store?" I don't know, and that's why I'm making these ads so I can actually go ahead and test them to see
Speaker A
what format works best. Or multiple formats actually might work best, and I'll scale all of good formats, and I'll kill off all the bad ones. Now, if you're probably wondering if you can actually go ahead and optimize this, you
Speaker A
definitely can. You can even make another LLM box right here and ask it what format would work best for this product for ads. And after that, you can actually split it off depending on what format it actually generates for you.
Speaker A
But, like I said, all these formats right here are pretty much just industry standard formats for majority of products. If you're selling a product, this should pretty much cover majority of products right here. So, first things first, once you have the node of your
Speaker A
product with your logo, you're going to split it off to an LLM. Now, this LLM right here, I put a very, very simple prompt. I said, "I need the perfect prompt for C-DANCE 2." And the second sentence I put, "I'm making an ad for my
Speaker A
product. It needs to be a UGC-style video, very organic-looking. It cannot look like an ad. And here's my product." And obviously, I put the links to my competitor. Now, I'm not going to read everything right here, but this is
Speaker A
pretty much what the LLM said to come up with a really good video for your product right here. And like I said, this is literally all written by AI. It did the research for me. It knows exactly what kind of avatar needs this
Speaker A
product based off of my competitor's landing page. And yeah, it pretty much did all the research for me without doing any of the research. It writes a script for you. It does everything for you. Once you have this node right here,
Speaker A
you're actually going to go ahead and connect it to another node called C-DANCE 2. Pretty easy to make. Right click, go ahead and put on video models right here, and put C-DANCE 2.0. And once you connect it, you're going to go
Speaker A
ahead and use that as a prompt right here. So, as you can see, we don't have to write a prompt anymore. All we got to do is actually connect the node from down here and make it into an actual
Speaker A
prompt. And here's how the video looks like off of the prompt by itself. Okay, ignore my messy room, but I have to show you guys this pillow I found. It looks kind of weird, but look at these side
Speaker A
wings. If you're a side sleep That is actually insane. Like, this video looks super well. Like I said, there are some errors like this is obviously not the hand of somebody that's sleeping on their back, but you can always go ahead
Speaker A
and regenerate it. And majority of times, I'm not even going to lie, most people will not even notice this. So, that's completely fine. Now, as you notice, it only lasts 10 seconds, which is completely fine. Because what we're
Speaker A
going to go ahead and do is do another node right here called extract frame. If you really want to know how to make that, you do right click right here, and go ahead and do tools right here, and
Speaker A
then go ahead and put extract frame. There's a bunch of different tools you can actually use and do for yourself, and actually customize your uh your workflow if you like. For example, replace actor, video editor, a bunch of
Speaker A
things you can do. Add captions, literally anything that you want AI can probably do nowadays. But, for us, all we really want to do is make this video longer. Because by the end of this video, I don't think it actually, you
Speaker A
know, does too well explaining what the product does. Literally cradles your head. And that's pretty much it. Like, they only said it cradles your head.
Speaker A
Obviously, that's not a really high-converting ad. Now, the next thing you want to do is extract the last frame, and then you want to go ahead and change the camera angle. And like I said again, if you want to create that node,
Speaker A
you want to right click here, click tools right here, and click camera angle. I actually switched the camera angle. So, now, instead of a front side, it's going to show the side angle. And the reason why we did that is because
Speaker A
obviously, we're trying to show off that the pillow actually works for side sleepers, and we're trying to show off that it actually cradles the side of your head. And we have another LLM right here. I don't want to figure out what's
Speaker A
next on this ad. I'm going to let AI do its job entirely. And all I'm going to do right here is put another prompt on how this neck pillow helps with neck pain using C-DANCE 2. Like I said, these
Speaker A
prompts can be super, super simple. Just make sure it makes sense for your product, and that's pretty much it. And once you have it like that, all you need to do is connect the camera angle to the LLM, connect the LLM to another C-DANCE
Speaker A
2 prompt, and connect the actual camera angle to the prompt again. And here is what it spat out.
Speaker A
And those weird side wings, they're a game-changer for side sleepers. [music] Your shoulder tucks right in, taking away all the pressure and keeping your spine perfectly aligned. There you go.
Speaker A
Now, you're probably wondering, "Oh, yeah, the voice doesn't sound too congruent here." There actually is a fix for that as well. All you really have to do is right click right here, click tools, and go ahead and click change
Speaker A
voice to make the voice more congruent. But, moving on right here, as you can see, the next thing you want to do is create more content because this UGC ad, or UGC ads in common, are usually around 30 seconds to a minute long. As you can
Speaker A
see right here, same thing, right click, extract frames. So, grab the last frame from the last photo, or the last video.
Speaker A
So, you actually go ahead and recreate or extend the video even more. Again, like I said, I had a very simple prompt right here. And once the LLM came up with its own prompt right here, we're going to connect it to the last one,
Speaker A
connect it to the extract frame again. And then it's pretty much what pops up.
Speaker A
Cloudia pillow and the sleep mask. It's the perfect combo for getting a good night's sleep. All right, there you go.
Speaker A
And then to put it all together, you're going to actually going to go ahead and make a node called stitch right here, and pretty much just combine all the videos that you made. As you can see, drag all the videos that you made, and
Speaker A
boom, now you have one congruent UGC ad. And that was literally all could be automated with AI. Now, let's just say you want different iterations of your UGC video, maybe a different actor, a different avatar, uh different setting,
Speaker A
or just overall a different vibe to the video. But, you still want the same exact script, and you still want the same exact things to your video, you actually could do that by just splitting it off again. As you can see, the
Speaker A
original LLM node that makes it into a UGC ad, all you got to do is split it off again, and make it C-DANCE 2 video again. So, for example, this is the next version, or next iteration.
Speaker A
Okay, ignore my messy room, but I have to show you guys this pillow I found. It looks kind of weird, but look at these side wings. If you're a side sleeper or wake up with a stiff neck, this center
Speaker A
dip literally cradles your head. As you can see, that's completely different from the original UGC creator that we actually had, which is right here. Even though they came from the same exact prompt, it's going to be different iterations of the ad. And same thing
Speaker A
here, I don't want to go into too much, you guys already know the deal. But, for example, you want to extract the last frame right here, put another LLM node right here, and pretty much go ahead and extend the video based on what the
Speaker A
actual AI tells you. So, not only are you getting a different kind of script as well, you're getting a different iteration of the ad completely overall.
Speaker A
It actually might go into different pain points, different messaging, and stuff like that. So, as you can see right here, once you connect the LLM bot as well as extract frame node, this is the result that comes from it.
Speaker A
But, here's the real secret. It's actually an anti-aging hack. Because of these side cutouts, my face doesn't get squished into the fabric all night, which means I'm no longer waking up with those annoying sleep wrinkles. There you go. Like I said, it went to a different
Speaker A
angle and a different pain point. So, that means they go ahead and actually AB test all your different UGC ads, and have different iterations of the exact same prompt. As you can see right here, we're going to extend it one more time.
Speaker A
So, extract frame again of this girl, another LLM bot right here. And I pretty much just put add a CTA to UGC video.
Speaker A
The prompt is for C-DANCE 2.0 only. And as you can see, you click start.
Speaker A
So, now you can kiss sleep wrinkles goodbye. Click the link below to get yours before they sell out.
Speaker A
There you go. That is pretty much it. And then the last thing is just a stitch bot, or a stitch node, where you put all of it together. So, let's go ahead and give it a quick watch really quickly.
Speaker A
Okay, ignore my messy room, but I have to show you guys this pillow I found. It looks kind of weird, but look at these side wings. If you're a side sleeper or wake up with a stiff neck, this center
Speaker A
dip literally cradles your head. I actually noticed a difference in my neck pain. But, here's the real secret.
Speaker A
It's actually an anti-aging hack. Because of these side cutouts, my face doesn't get squished into the fabric all night. Now, it's pretty good. Like I said, there are some things that are off with it. Obviously, it's going to be the
Speaker A
subtitles. You can actually prompt it to go ahead and remove the subtitles, and then add it later by clicking a node right here, and choosing subtitles under tools right here. But, like I said, there are going to be some things that
Speaker A
are wrong with it, but nobody's going to care, and nobody's going to notice. Trust me, no one actually cares about these small things. But, if you're OCD like me, you definitely can change it.
Speaker A
It's not a big deal. Now, let's actually move on to the next format that we have right here, which is going to be podcast ads. All right, for the podcast ad, we have another LLM node right here. And
Speaker A
pretty much all I'm doing is pretty much asking it to make a perfect prompt for C-DANCE 2 for an ad for the product. And it needs to be a podcast studio. And that's pretty much it. And once you have
Speaker A
this entire prompt right here, you connect it to the C-DANCE 2 video model node, and we go ahead and watch the results right here. Yeah, exactly. And honestly, ever since I switched to this alignment pillow, my mornings are
Speaker A
totally different. The side sleeper wings actually support my neck. It's a game-changer. Yeah. there you go. And you're probably wondering, "Well, why would I want to run podcast ads?" I think it's pretty obvious, but a lot of people are drawn
Speaker A
to podcast ads simply because the person on the screen just looks like an authority figure, and just looks like they know exactly what they're talking about. For some reason, it just has that effect on people. I'm not going to argue
Speaker A
with it because uh it even works on me a lot of times. So, I'm pretty much this is why a format is really good for every single product that I test, and actually crushes for majority of products, especially if it helps a really
Speaker A
deep-rooted problem in people's lives. Anyways, enough about that. The next thing you want to do is actually extend it, just exactly what I did with the UGC ad. So, same thing, extract the last frame right here, so the video has
Speaker A
context on what the person looks like, and what's actually going on in the video. And then another LLM bot right here. Same exact thing, except that now you can actually go ahead and write how this pillow helps with the neck support
Speaker A
in the first 10 seconds. Same exact thing, go ahead and actually put this into another C-DANCE video. These are the results right here. It's a game-changer. It features breathable fabric that incorporates advanced cooling technology.
Speaker A
[music] This helps regulate temperature and prevent night sweats. All right, there you go. That is the actual result of the second clip right here. And then if you want to go ahead and change things up, you definitely can as well. As you can
Speaker A
see right here, I actually generated an image of a night mask right here, uploaded image of a night mask, and then another LLM bot right here. I wanted to go ahead and say finish this podcast ad by showing my upsell. So, if you wanted
Speaker A
to sell two items in one, you also could now because of workflows. And extract the last frame as usual, and this is going to be the last clip that we have right here. But, to get the absolute most out of your sleep, you have to pair
Speaker A
the cooling pillow with our new 3D contoured sleep mask. When you combine the cooling comfort of the pillow with the total blackout of the mask, it's the ultimate recipe for deep, uninterrupted sleep. There you go. And this is going
Speaker A
to be the stitch node that puts everything together, so you don't have to put everything into CapCut, or your own video editor. You kind of just stitch it all together like this. Just click down on the top right, and this is
Speaker A
the video right here. But, to get the absolute most out of your sleep, you have to pair the cooling pillow with our new 3D contoured sleep mask. When you combine the cooling comfort of the pillow with the total blackout of the
Speaker A
mask, it's the ultimate recipe for deep, uninterrupted sleep. It's a game-changer. It features All right, obviously, I didn't say that correctly right there. But, if I wanted to change it, I definitely could by clicking onto this and moving this video
Speaker A
to the bottom. Now, I think you guys get the entire idea by now. But, the next one we're going to have is going to be more of a studio professional ad, and the prompt is in a LLM node. We connect
Speaker A
it to a C-DANCE 2 video generator right here. And this is going to be the actual generation of it.
Speaker A
[music] All right, as you can see, a super professional, more like TV-esque, more like, you know, super highly branded company-esque, like Nike or Amazon or something crazy, a billion-dollar company budget, except that, you know, it was just created with AI using Arc
Speaker A
Ads. Like I said, if you want to use Arc Ads, make sure to use the link down below. But, the next thing you want to do is obviously extract frame as usual, same thing, LLM node, and boom, you
Speaker A
actually go ahead and combine these two for the final result, which looks something like this.
Speaker A
[music] [music] [music] [music] All right, so right there, that what, like 20-second commercial, it could have been filmed in some crazy podcast studio or some videography studio, but it was actually all made with AI, and it makes this product look super, super premium.
Speaker A
And all we needed to do was write what, two sentences in order to go ahead and create that. So, it's actually crazy.
Speaker A
Now, I have a few more formats down here that I'm not going to show up too, too much. I already showed you guys the music trend right here. Really quickly, I'll show you guys again.
Speaker A
[music] All right, I don't know what music to use, I don't want to get copyrighted, but that's pretty much it right there.
Speaker A
Honestly, not much that needs to be changed with this workflow, unless you want to make different iterations, which is going to be pretty easy. And that's pretty much just taking this LLM node right here and making it into another
Speaker A
node. So, for example, video model C dance two, and just pretty much doing it again, and just connecting the prompt right here to this right here. There you go. And if you ever wanted to do static ads, you definitely could do that. You
Speaker A
just go down right here, we just have regular Nana Banana Pros. So, now we have two static ads including a bunch of video ads as well, so three video ads, all of different formats, actually four video ads because we have two iterations
Speaker A
of the UGC. And all of this can actually be automated in one thing, where all you got to do is upload your competitor's image and rebrand it to yourself, and then boom, once you click play, it'll basically be able to go ahead and
Speaker A
automate all of this, and all you really got to do is click run at the end of the day, just copy all of these videos down here, and honestly, that's pretty much it. By the way, if you ever had any
Speaker A
questions while watching my videos, I actually do have a place where you can actually communicate with me. I have a private Discord down below, as you can see, that I answer questions daily. I also have weekly trainings every single
Speaker A
Friday, so if you ever want to interact with me in a live call where I answer your questions, I am there every single Friday as well. I even have case studies showing you guys exactly what my store looks like, the entire website, the
Speaker A
video ads, where the product was found, my actual website as well, and even the ads manager, so make sure to check the Discord link down below. I'll even help you scale your ads in one of my Discord channels, and we have an active
Speaker A
community chat with like-minded people that are trying to figure out this drop shipping game, or a bunch of us that already figured it out and kind of just helping people. So, if you want to join the Discord chat, a bunch of us are
Speaker A
actually there right now, so make sure to go to the link down below, and it's called Prime. I'll see you guys inside.
Speaker A
All right, now that we've gone through the most, you know, popular formats, there's a few other formats you can actually do as well. Because we are using AI, the formats are actually unlimited. So, for example, if you are
Speaker A
selling a fashion product, you can actually go ahead and do a TikTok try-on video as a format. There's even creative ones like street interviews, or even product unboxing videos, those do really well as well. Even something crazy out
Speaker A
of the ordinary like an ASMR video as well about the product, that could work as well. And if you're wondering, okay, how am I getting these formats? Number one, you can actually ask AI what formats would work best with the LLM
Speaker A
node, like I said before on Arc Ads. Or number two, just kind of look at your competitors' videos and see what formats are working for them. You know, usually, whatever formats that they have most prominent of are usually the formats
Speaker A
that are working, so you kind of want to go ahead and just copy what's working already. Or if not, just go ahead and tell AI, "Hey, this is my product, these are my competitors, recommend me a few formats for me to try out." Because you
Speaker A
will never know until you actually try out these formats. But like I said, because you have AI now and Arc Ads, the limits are literally limitless because you can do any scenario you want. You can have Neil Armstrong on Mars with
Speaker A
Elon Musk, literally anything that you want. And like I said, if you want to use Arc Ads, make sure to use the link down below. So, now that you actually understand how these workflows work and how these nodes work and how these API
Speaker A
models work and everything like that, you actually want to go ahead and start automating the process. Now, everything that I mentioned so far has been pretty static, like the workflow I just showed you. Unless you change a couple of the
Speaker A
prompts, it's not going to work for every single product, right? So, go back to our workflow right here. If I actually went ahead and changed this product image right here to like a razor, right? It's not going to make
Speaker A
sense when I say rebrand my product to Cloudora. That makes no sense. So, you actually have to go ahead and make a lot of these prompts into dynamic prompts.
Speaker A
So, for example, the LLM should actually be something like, "Take a look at my product and recommend a brand name, and then rebrand my product." So, once you do this, it does not matter what kind of product I upload here. I could upload a
Speaker A
razor, I could upload something like a motorcycle pad, I could upload some home furnishings like a chair. It doesn't matter because once it gets this, the prompt is going to be, "Take a look at my product and recommend a brand name,
Speaker A
and then rebrand my product." And whatever it is, it's going to rebrand it to something that makes sense for the product. So, that means you want to make all of these nodes right here actually dynamic. Now, the second thing right
Speaker A
here, as you can see, I need a perfect prompt for C dance two. That's completely fine as well. But as you keep going to the right, it says right here, it needs to be a UGC video, very organic-looking, but at the end, it
Speaker A
says, "Here's my product." Now, obviously, that's very, very static, right? Like, our process wants to be automated. You don't want to go ahead and keep switching out every single prompt because that is not automation.
Speaker A
So, what you need to do is also switch out these workflows to be a little bit more dynamic. So, same thing right here, you're going to say, "I need a perfect prompt for C dance two." And something like, "The image is my product." And
Speaker A
that pretty much makes this node super dynamic because whatever uploaded here or whatever image gets recreated here, it's going to use that image for context and not your competitor. Now, if you want to be extra savvy with it, you
Speaker A
definitely could because like I said, these things are actually all limitless. All you got to do is right click right here, do text inputs, put another prompt, and just put, "This is my competitor." And then you want to go
Speaker A
ahead and put your competitor's link right here. And once you do that, you can actually not only just link the prompts, but you actually go ahead and link the prompt and the photo to your LLM bot, so it gets all of the context.
Speaker A
And you want to do that for every single workflow node that you have here, so it could be super dynamic, and it doesn't only go ahead and talk about one specific thing. And that's how you should start automating the process by
Speaker A
changing all your workflows to be more dynamic instead of just making it very static about one single product. So, does that mean your entire workflow is now automated? No, not really. It does make it more dynamic, and it is more
Speaker A
automated in the sense that now you can put any product in the beginning and kind of change it out, and then the entire workflow will work towards just the first few nodes, but it is not automated by any means. So, for example,
Speaker A
if you go into Arc Ads right here, yes, all you have really got to do is put your product image right here, and also put your competitor's link right here, and all of these nodes, theoretically, if you change the prompts, should work
Speaker A
off of the original two nodes right here. But the issue is that you have to go ahead and click regen or generate for every single node. Generate, generate, generate, generate, which means it's not completely automated. Now, if you want
Speaker A
to be more automated, what you need to do is firstly go to Arc Ads and grab your API key. So, Arc Ads API, go to Arc Ads, go to settings, and click public API right here, and grab your API key.
Speaker A
Now, you do have to be on a very specific plan on Arc Ads in order to go ahead and get that, so make sure you upgrade your plan on Arc Ads before you do this. And once you do that, you put
Speaker A
that API key into Claude code. And once you put that API key into Claude code, the next thing you're going to go ahead and do is take a screenshot of your entire workflow right here. And once you take that screenshot, you're also going
Speaker A
to put that into Claude code as well. And you're going to tell Claude code exactly what you want your software to go ahead and start doing. So, for example, a very simple prompt could be in the terms of, "All I really want to
Speaker A
do is take some of my competitor's product as well as their landing page, and I want to recreate a bunch of different ads based off a bunch of different formats that would work for me." And once you do that, Claude code
Speaker A
might ask you some questions about UI and other miscellaneous stuff, but at the end of the day, it should go ahead and create an entire software for you just like that. And that's pretty much it when it comes to actually making your
Speaker A
AI workflow a little bit more automated. So, if you are interested in AI workflows and you want me to help you one-on-one, make sure to first link down below. I have students that hit $1.7 million in revenue before, 40K in a
Speaker A
single day, even 85K in a single day. And these students right now are currently using my product research automation that shows you a bunch of products on Facebook Ads Library. And obviously, you're not going to test bad products, so we even have an AI product
Speaker A
validator tool that helps you make sure you're not testing bad products right here. And even then, we have our AI ad clone automation as well as our AI creative strategy automation as well.
Speaker A
So, if you want these tools and you want my help one-on-one to help you scale to numbers like this on my screen right here, 40K in a single day, 85K in a single day, or $1.75 million in revenue,
Speaker A
make sure to hit the first link down below. And yeah, that's pretty much it.
Speaker A
If you guys want to go ahead and use these ads and actually launch them on Facebook, I should have a video right here, and I'll see you guys next time.
Speaker A
Peace out.
Topics:Seedance 2AI UGC adsautomated ad creationecommerce marketingAI ad cloningproduct research automationhigh converting adshyperrealistic AI adsAndrew YuArc Ads
![[정치포차Live] 인천광역시교육감 도성훈 후보_제9회 전국동시지방선거 — Transcript](https://i.ytimg.com/vi/zT0MUGjyGYs/maxresdefault.jpg)
