The problem I kept running into
In agency work I ran search, social and programmatic — Meta, Google, DV360. I got good at the levers: structure, audiences, bidding, tracking. And I watched the same thing happen over and over.
A campaign would underperform. We'd rebuild the structure, tighten the targeting, fix the pixel. Small gains. Then someone would swap the creative and the numbers would move more in a week than three months of optimisation had managed.
For a large brand that's an annoyance. For a small business it's fatal — they can't afford a production budget, so they run whatever they have, and no amount of media buying rescues a weak asset.
Why AI video
AI tools closed that gap. I can now write, generate and cut a cinematic video ad in a day, at a cost that makes sense for a shop in Pune or a D2C brand doing its first ₹1L month. Same day, I can put it into a properly structured campaign and watch what it does.
That loop — make it, run it, read the data, remake it — is the entire advantage. Not the AI. The fact that both halves sit with one person, so a creative test takes hours instead of weeks.
Where I've worked
My agency years covered categories with very different economics — which is where most of the useful lessons came from.
FMCG
High volume, thin margins, brand and performance pulling against each other.
Banking
Heavy compliance, long consideration, lead quality over lead count.
Jewellery
High ticket retail where the creative has to carry the entire sell.
Pet nutrition
D2C, repeat purchase, where LTV matters far more than first-order ROAS.
Gardening
Seasonal demand and a niche audience that broad targeting handles badly.
Small business
Cafés, local retail and creators — small budgets where every rupee is visible.
Channels & platforms
How I work now
The creative
Scripts, hooks, and finished video built with AI tools. Cinematic product ads without a camera, a studio or a crew — which is what makes testing affordable at small budgets.
The media
Campaign structure, audience strategy, tracking, and the discipline to cut what isn't working early. Five years of doing this across categories where the margins left no room for guessing.
What I actually believe
Most accounts I look at are not broken in exotic ways. They're running too many things at once, spending on creative that stopped working weeks ago, and measuring ROAS before costs — which is why I built the break-even calculator.
I'd rather tell you your unit economics don't support paid ads yet than take your money for six months. That's usually a shorter conversation, and it's a better one.