I have no clients. No case studies. No proven process. And I am more confident about this business than anything I have built in the last five years.
Here is why.
The shift nobody in advertising wants to admit
Every single ad that has ever existed in human history has featured a real human being. A face. A voice. A person you recognize or are supposed to trust.
That model is about to be completely dismantled.
The next wave of performance advertising is AI avatars that look, sound, and feel human — running hundreds of variations simultaneously, at a fraction of the cost, tested and iterated in real time. Different hooks, different emotional angles, different scenarios, all generated at scale without a single person on set.
The brands that figure this out first will have an enormous competitive advantage. The agencies that can deliver it will print money.
That window is open right now. In 18 months it will not be.
"Making the avatar look real is the easy part. The hard part is the same thing that has always been the hard part in advertising — the hook, the psychology, the scroll-stopping moment."
Why most people building in this space will fail
Here is what almost everyone gets wrong about AI UGC: they think it is a technology problem.
Get the AI to look real enough and the rest takes care of itself. That thinking is exactly why most AI-generated ads fail to convert.
Making the avatar look real is the easy part. The hard part is the same thing that has always been the hard part in advertising — the hook, the psychology, the scroll-stopping moment in the first two seconds that makes someone actually watch instead of skip.
That is not a technology problem. That is a content problem.
And it is exactly why someone with a decade of viral content creation has a structural edge here that no tech-first person can replicate overnight. I have spent ten years studying what makes people stop scrolling. Hundreds of millions of views. Dozens of brand deals. I know the difference between a video that looks good and a video that converts. That background is not a footnote to this business. It is the entire foundation.
What I am building before I pitch anyone
Knowing the craft is not enough. Before I approach a single brand, I am solving two problems first.
Quality control. I am not going to sell something I do not believe in. If I pitch a business knowing the output is not good enough to actually make them money, that is not a service — that is a scam. The standard has to be high enough that I would stake my name on every deliverable before it goes out.
Pipeline. The goal is a workflow where I can generate 100, 200, even 500 different video variations in a single day. Quality without pipeline means I can never scale. Pipeline without quality means I scale fast and destroy my reputation doing it. I need both locked in before I go to market.
"The first client who pays me and makes five or ten times their investment back changes everything. That result becomes my case study. And proof in this industry is worth more than any pitch deck ever written."
The only question that actually matters
The honest fear underneath all of this is not about cold pitching. I am not scared of outreach.
What I genuinely do not know yet is whether this converts.
I can produce something that looks like a real human on camera. But does it actually make a brand more money? Does the audience trust it the same way? Does the hook land? Those questions do not have definitive answers yet — for me or for anyone in this space. We are all figuring it out in real time.
That is both the risk and the opportunity.
The first client who pays me and makes five or ten times their investment back changes everything. That result becomes my case study. That case study becomes my proof. And proof in this industry is worth more than any pitch deck ever written.
That first thousand dollars is not about the money. It is about answering the only question that matters: does this actually work?
Everything else I am building is preparation for finding out.
The AI UGC industry is in the same place YouTube was in 2012. Most people are still watching from the sidelines trying to figure out if it is real. A small group is already building. The gap between those two groups is going to be enormous in 18 months.
I am documenting every step of this build in public — the wins, the failures, the first client, the first real result.
If you are building something in AI right now — or thinking about it — what is the one thing stopping you from starting? Drop it in the comments. I read every single one.
Originally published on Mr. Lin Uncut on Substack
Top comments (0)