Do not index
Do not index
Surya Upadhyayula — CEO, Launchionaire. Runs a Meta ads agency with a wide variety of clients.
10 / 10
ad scripts approved first try → zero revisions, straight to recording
I just recently wrote 10 ad scripts for my client and she just said that all the 10 ad scripts are really great. We can go ahead into the recording phase for all the 10 ad scripts. Amazing.
The Challenge: Running an ads agency where every revision eats your margin
Surya is the CEO of Launchionaire — a Meta ads agency. His business is launching ad campaigns for a wide variety of clients. Different industries. Different tones. Different audiences. Different voices.
The workflow looks simple on paper: client brief → write ad copy → send for approval → record → launch. But anyone who runs an agency knows the real bottleneck.
It's not writing SPEED. Any decent copywriter can write ten ads in a day if they have to. The bottleneck is TONE.
Every client has a voice. A brand. A personality. And if the scripts you send back don't sound like THEM — they sound like a generic agency copywriter — you're going into revisions. Round one. Round two. Round three. Margin evaporating every round.
Surya needed a way to scale content output WITHOUT losing the tonal fingerprint of each individual client. That's a hard problem. Most AI tools can write fast. None can write in ten different client voices and nail each one.
The Solution: Train Poppy on each client's voice once. Then scale forever.
Here's Surya's exact workflow in his own words: "One of the things that Poppy helped me to do this faster is create their copy for my clients in their tone of voice very fast."
The Poppy canvas lets him lock a client's voice in ONE place. He drops in reference scripts, past winning ads, brand guidelines, tone examples — anything that defines how the client sounds.
Then every new campaign brief slots into that pre-loaded context. Poppy writes in THAT client's voice. Not Surya's. Not the agency's. Not a generic ChatGPT voice.
This is why it works: Poppy isn't stateless like ChatGPT. The canvas HOLDS the context. Client A's canvas. Client B's canvas. Client C's canvas. Each one a separate universe with its own tone locked in.
So when a brief comes in, Surya doesn't re-explain tone every time. He doesn't paste in examples every prompt. The canvas already knows. And Poppy drafts first-try-ready scripts in minutes.
The Results: 10 scripts. 10 approvals. Zero revisions. Straight to the recording studio.
Most recent test: one client, one campaign, ten ad scripts needed.
Surya wrote all ten in Poppy using the client's locked-in voice. Sent them over.
The reply: "All the 10 ad scripts are really great. We can go ahead into the recording phase for all the 10 ad scripts."
Ten out of ten. Zero revisions. No round two. No round three. Straight from script to camera in one send.
The agency math here is brutal. Before Poppy: maybe 3-4 approved out of 10, rest bounced to revisions. Now: 10/10. That's not a 2× improvement. That's 10× the throughput at 0× the revision load.
And Surya's not looking back. His own words: "I only see myself using Poppy a lot, going far, going forward."
If you're on the fence, give it a try. This might help you get your stuff done faster.
Written by
