The video
Every post in this series starts with an ad in my feed. This one starts a step earlier: with a YouTube video that teaches people how to make those ads.
The title promises ₹1,00,000 from an AI side hustle. Zero skill, zero effort, zero investment. The plan: get AI to write ebooks, sell them for ₹599, run Facebook ads, collect.
I watched all fifteen minutes. The most honest thing in it turned out to be his own numbers.
As always, I'm not naming him. He even shows a search page full of videos selling the same dream. This is about the playbook, not the person.
What he actually did
Five ebooks, five storefront pages, five ad campaigns. The whole thing in six steps:
- Find a niche. Paste five popular "AI ebook" YouTube videos into an AI and ask what they have in common. Then run an AI deep-research report on Indian ebook buyers. Land on AI, career, freelancing, weight loss and spoken English.
- Write titles nobody can say no to. "Lose weight without exercise, without diet, just 15 minutes a day." "Master English in 15 minutes a day."
- Write the books with AI. Feed it his own YouTube script, plus free "private label rights" PDFs downloaded from the internet. About 20 minutes per book.
- Mock up a cover. Screenshot page one, ask AI to make it look like a real paperback.
- List them for ₹599 on a creator storefront that takes a cut of every sale.
- Run video ads. One campaign per book, ₹500 a day each.
Notice step one. To find out what sells, he asked AI to summarise videos made by people selling the same dream. Unsurprisingly, the top niche it found was making money.
The receipts
₹3,600 in, ₹1,797 out. He rounds it to ₹1,800 and calls the three sales "crazy".
Per sale, he paid about ₹700 in ads to sell a ₹599 book. That's before the storefront's cut and the tax on ad spend, so the real loss is bigger than the one on screen.
And ₹3,600 across five campaigns at ₹500 a day is about a day and a half of data. Three sales can't prove a niche works. His closing line says it does anyway.
The line he crossed
Early on, he says something decent: we don't have to scam our customers. Ten minutes later, he makes the ads.
The freelancing ad uses what he says is his own story. Fine. Then, translated from his Hindi: if you don't have an experience like this, you can just make up a story. So he does.
- The weight-loss ad: "I went from 80 kg to 64 kg in three months. No gym, no diet, no medicine." He tells us it's fiction. He also says he has no scientific knowledge of weight loss.
- The career ad: a ₹15,000 job offer, then a ₹1.5 lakh offer he "rejected because it was too low". Also fake, he says, but made to look real.
The person buying the weight-loss book isn't paying for 15 AI-written pages. They're paying for the 16 kilos. Those kilos never existed.
The rules are clear about this too:
- ASCI's influencer guidelines say a testimonial should be genuine and come from the person's own experience.
- Since April 2025, ASCI expects influencers talking health or finance to hold verifiable qualifications. A no-expertise weight-loss book fails that before anyone opens it.
- Meta now judges health ads by their claims. In August 2026, an ad was blocked for promising "easy weight loss". "No gym, no diet, 16 kg" is the same kind of promise.
But the rules aren't really my point. Even if nobody ever checked, it would still be lying.
Credit where it's due
He published the loss. Most "copy me" videos quietly skip that part, and I respect him for not skipping it.
Two ideas in here are worth stealing, too:
- The topic beats the niche. "Switch jobs with a 50% hike" sells. "Career" doesn't. People buy a specific outcome, not a category. That's true for an ebook, a carousel and a landing page.
- Test small, then kill or scale. One campaign per product, a fixed daily budget, cut what doesn't work. That's how you should spend money you can't afford to waste.
The problem was never the tools. AI can draft a decent PDF in 20 minutes. It just can't give you a story you haven't lived.
Things I won't do
The list grows with every review. Three new lines this time:
- I won't pretend to be a customer of something I'm promoting.
- I won't buy followers, comments or upvotes.
- I won't talk about my own work anywhere without saying it's mine.
- I won't quote a statistic I haven't checked.
- I won't sell anyone growth that depends on fooling the people they want to reach.
- New: I won't invent a story and tell it as mine.
- New: I won't sell advice in a field I'm not qualified in.
- New: I won't put a number in a headline that I didn't actually make.
How to spot it
You'll see these ads in your feed. Four tells:
- A transformation with nothing given up. No gym, no diet, no effort, big result. Real change always costs something.
- "All my secrets in this ebook." If the secret made them lakhs, it wouldn't be selling for ₹599.
- Proof you can't check. An offer letter flashed for two seconds, a weight with no face, a name with no link.
- A number in the title that the video never shows. Watch to the end. Here, ₹1 lakh became ₹1,800.
The honest version of this business exists, and it's slower. Write about what you actually know. Show what actually happened. Sell it to people who already trust you.
If your product needs a made-up story to sell, the problem isn't your ads.
Sources: ASCI on genuine testimonials (Money9) · ASCI health & finance update, April 2025 (Conventus Law) · Meta health ad policy, July 2026 (Adligator)