ByteDance's Dreamina Seedance 2.5 launches globally on Dreamina, available in the U.S. as well. Dreamina now features dedicated product upgrades, making it the most optimized and cost-effective platform for running Seedance 2.5. Sign up for free now: https://bit.ly/jonlawedu7 or comment "Dreamina"
The highest converting ad format on the internet is a person talking to a camera, so in Part 4 of building the Arrow wireless earbuds brand entirely with AI, I'm giving the brand a spokesperson. In this video I create a realistic AI avatar in Dreamina, generate talking-head UGC style video ads with Seedance 2.5 using product reference images, add background styles from reference videos, and translate the finished ads into Japanese and Spanish with perfect lip sync. By the end we have a whole portfolio of spokesperson ads ready for Google, Facebook, and Instagram campaigns. Comment "Dreamina" below or use the link in the description to follow along and build your own brand.
CHAPTERS: 0:00 The Highest Converting Ad Format Online 0:19 In This Video: Giving Our AI Brand a Spokesperson 0:44 The Arrow AI Brand Build So Far (Parts 1-3 Recap) 1:09 Dreamina Sponsorship + How to Follow Along 1:29 Reviewing the Arrow Brand Assets 1:45 Creating an AI Avatar Reference Image in Dreamina (Image 5.0) 2:24 Using Dreamina AI Avatar Pro 2:38 Selecting an AI Voice and Adding the Ad Script 3:05 Why Seedance 2.5 Gives You More Control 3:32 Creating AI Spokesperson Ads with Seedance 2.5 3:40 Uploading Product Reference Images (Up to 50 References) 4:25 Using a Reference Video to Set the Background Style 5:04 AI Avatar Results: Lip Sync and Realism Check 5:47 Generating Videos with the Agent Canvas Chat 6:22 Seedance 2.5 AI UGC Ad Results 7:15 Composited Background Ad Version 7:48 Green Screen Spokesperson Version 8:23 Final Ad with the Arrow Logo in the Background 9:03 Translating AI Ads into Japanese and Spanish 9:48 Creating Ad Variations for Google, Facebook, and Instagram 10:22 Tips for Better AI Video Generations 10:57 Honest Limitations of AI Spokesperson Ads 11:31 What's Coming in Part 5: The Full Commercial Spot
------------------- Clips channel @jonlawclips
I do NOT sell any products of my own. Watch out for scams.
So the highest converting ad format on the internet right now, period, is a person talking
to a camera.
And that can be a problem if you're a one-person brand and you hate being on camera, or you
can't afford to pay creators to do it on your behalf.
So today, the brand that we are building in this six-part series completely with AI is
going to get a spokesperson.
And by the end of this video, we're going to try and create a whole range, a whole portfolio
of visual assets where our AI spokesperson who looks real, who sounds real in multiple
languages is speaking to the camera and informing audiences about the brand.
This is part four of our AI brand build.
It's going to be a fun one.
Let's go.
So catching you up if you're new here, I'm building a wireless earbuds brand called Aero
completely with AI.
In part one, we made our brands or logos and our brand elements.
In part two, we did a photo shoot to develop what the product looks like, as well as models
to use it.
In part three, we got into the world of video since cDance 2.5, one of the new AI video
models just launched.
And now for part four, we're giving our brand a spokesperson.
Now Dreamina has been an amazing partner on this series.
They are sponsoring this video as well.
And so if you want to support the channel and follow along, bring your own brand into
the world while you're watching me do it, then you can comment Dreamina under this video
or go to this link on the screen or in the description to follow along.
That in mind, let's jump into my computer and get into the build.
So just to recap so far here, just some of the assets we've created for Aero.
So everything from our models to our posters, our product images over here.
But what I want to start with for this video is creating our avatar.
And to do that, we want a reference image to pull from.
We'll add the voice and other elements later.
So I'm going to paste this prompt just in the dashboard Dreamina AI image here.
We're going to say portrait of a woman in her late 20s.
And then we're describing what she's wearing.
I'll add hyper-realistic on the end here as well.
I'm going to keep this on image 5.0.
And let's generate our reference image for our avatar.
Here we go.
We have our model.
And we can see very realistic, high resolution as well.
Let's go ahead and download these images.
Now with that in mind, what we're going to do is jump back to the home page
and then switch from AI image to AI avatar.
We're going to keep this on avatar pro.
And then I'm going to upload this photo of our spokesperson.
And then for voice right here, we can select
from voices that are within the Dreamina platform.
So let's go young adult.
We'll go English.
I think Valley Girl is most appropriate.
So we'll stick with that.
And then let's enter this as the script.
These are Aero, 40 hours of battery, noise canceling that works.
And they cost less than the big names.
I haven't taken mine off in a week.
So we're not yet going to actually insert the product images for now.
We just want to get a feel for the avatar and what that's going to look like.
And let's let that prompt run.
So that's one way to do it.
And I wanted to illustrate the AI avatar model,
but what is even more relevant to the times
considering the very recent launch of Dreamina SpeedDance 2.5
is just going through this flow with more control over the content,
more references, adding in the actual product to the videos
and some other advantages.
And for that, we're going to switch from AI avatar to AI video
and do this directly with Dreamina SpeedDance 2.5.
So what I'm going to do,
same process of uploading two reference images
because Dreamina SpeedDance can have up to 50 now.
So we have our two reference photos of the model.
And then I have some other references here of the product I'm adding
so we can see product shots like these.
And then I'm going to select right here our spokesperson.
And then here, I'll go ahead and select the product.
So what I'm saying is,
hey, I want this kind of a medium shot with our spokesperson.
I've pasted in the script
and then we're referencing the product.
So we're just going to feed this straight into Dreamina SpeedDance 2.5.
We're going to keep it on 16 by 9.
We have Omni Reference selected.
Let's make this more like go 13 seconds and let's let it run.
Now something cool we can also do is upload a reference video
of the style that we want for our spokesperson to use.
So let's say I want a kind of gym vibe for this.
What I can do is search for what that gym looks like.
So we can go for something like this where we can see here in front of the camera
kind of looking straight onto it.
Let's download that.
And then now I can upload this as another reference right here.
And then I can say use the background style of
and then reference this video in the gym.
So let's let that run as well and then compare the difference
that that reference makes to the Dreamina SpeedDance 2.5 model.
Our avatar generation finished.
Let's give this a watch for a few seconds.
These are Aero.
40 hours of battery, noise cancelling that actually works
and they cost less than the big names.
So super high res, the lip sync is perfect and the hands are good as well,
which for those in the know was not always the case.
Now, again, this was our base take without the products
and that's not going to be in there.
But that's just the base quality of our spokesperson and avatar.
I would want to speed up the audio a little bit.
That's something we can do.
But overall, super awesome first pass for the brand.
And now we just want those other, of course, product and brand elements in there.
One more thing before we look at more of those final generations
with our spokesperson is again, if we hit forward slash in the Octo canvas here,
then we can just chat with our Octo agent to do the same thing.
So, for example, I can take this image.
I can open a timeline view here and I can say something like,
I want to animate our female gym model with our Aero earbuds,
generate a video with ZDance 2.5.
So if you prefer this kind of canvas kind of bulk generation flow
with more of an assistant feel, then you can do that as well.
But that said, let's go check out some of those other finished apps.
And here's one of those.
Let's open it up larger.
These are Aero, 40 hours of battery, noise canceling that actually works,
and they cost less than the big nines.
I have not taken mine off in a week.
So super awesome there.
I mean, if we just zoom in, hands are perfect.
That's the actual product.
So it's consistent there.
High res.
What we do want, and this is just a first pass,
are brand elements in the background or something more engaging.
That's what some of these future ads we just created are for.
But again, extremely high quality.
This is something that you would probably sell to some extent
that you could plug into an ad platform.
But we do have more work there to do.
That does give you an idea of the quality of the model.
And if you haven't seen AI-generated ads before, this is where they're at.
It is really quite something to witness.
Here's the next generation here.
We can see we've layered in or composited a background.
So it looks more realistic.
This is getting much closer to the exact kind of ads
you see out there in the world.
These are Aero.
40 hours of battery, noise canceling that actually works,
and they cost less than the big names.
I have not taken mine off in a week.
I like the little layered sound of them opening there at the start.
That's pretty cool.
Let's check out the other ones.
So here's our version with a green screen.
So if we wanted to composite a background, then we could.
Or if we wanted to feed this back into 3-Minute Feed-Ins 2.5
to make it easier for the model to just go in and layer
different backgrounds or movement.
These are Aero.
40 hours of battery, noise canceling that actually works,
and they cost less than the big names.
I have not taken mine off in a week.
Oh, I like that.
I like when she looks down.
Super realistic.
Looks good.
And then this is the final little generation I created here.
You can see our logo in the background,
consistent with the original generated images.
This is my favorite of the bunch.
Number one, it shows that you can repurpose.
That's going to be critical for ads.
You want to take one ad.
You want to make different variations of it for all your platforms
or for social posts.
And then this makes it feel much more like, you know, you're in a studio.
You're in the office of the company.
So we can see Aero in the background.
Let's give this final generation a look.
These are Aero.
40 hours of battery, noise canceling that actually works,
and they cost less than the big names.
I have not taken mine off in a week.
There we go.
So that is.
Now, quickly, something I did forget to mention here in Showcase
is the ability to translate content in these ads into other languages,
which is a massive use case.
So what I did is I took this exact same video that we just made.
And here, you can hear it in Japanese.
Same thing in Spanish.
Perfect lip sync.
Accent is on point.
And so what this means is you can take your one ad,
and then you can just now open it up to 5, 10, 15, 20 plus new markets,
and they're adding new languages all the time.
So super powerful use case there.
Just wanted to shout that out and showcase these.
Our AI spokesperson for the Aero brand.
We've created a few video ads at this point
that we can use for various campaigns.
We might want to layer some B-roll of the products in over it.
We could do that with the timeline feature over here.
And so if we combine just what we've created here
in the past few minutes with all the various product shots,
that's probably 5, 10, 15 different ad variations
that we can put out there on Google Ads, Facebook, Instagram,
on social accounts.
And that is what is coming up in the near future.
We have one more video where we're going to continue
building out some of these assets,
and then we'll get into the actual launch of the Aero brand.
And like I always say here, right, I want you to understand the process
and see how you can create the kind of layer
different assets into these generations.
So I hope that that's what you're noticing,
how to go through these flows, what's important,
like context, starting with a reference image,
asking for logos to be placed in the background.
It's those steps that will take your generation
of media assets with AI to the next level.
And there you go, Aero is coming more to life by the day.
So overall take here, yes, the English version
is by far the strongest.
So if you want to go develop across 20, 30 languages,
you are going to have trouble doing that.
I hope that's a point they can improve in the future.
Number two, close crops are rough in terms of realism.
I think things which are a little more wide angle
are going to work better.
And again, that's something future video models
are going to improve on.
And then finally, if there's a lot of hands or motion
in the frame, also just something you want to generate
multiple takes of and pick the best.
So don't assume by default that it's perfect.
And so there we go, Aero has a spokesperson,
the brand is coming to life.
And the next episode is going to be a fun one.
We're going to stitch a lot of these assets
to be made together into a final kind of commercial spot.
So that's something like a real TV spot
or a flagship ad that you could run for the brand.
And it will build on a lot of the base work
we've done over the past few episodes.
So it's going to be a fun one.
Again, you can comment three minute under this video
to get the link to do it yourself
or just go to the description.
And that in mind, I will see y'all back for part five
in the very near future.
I appreciate you being here
and I will see you in the next one.
Cheers.