GTM Sauce

Flamme · B2C · couples app

Give AI a 10-page brief with product details, winning scripts and a master prompt

Team posts ~200 pieces a day from it; winners are fed back in

Workedorganic-short-videougc-creators

What they did

Each playbook is a ~10-page PDF the team feeds to AI. Section A: product and brand context (name, category, what it is in human terms, core features, target audience, psychology). Section B: winning script references (own and others' winning content in the category, with 'what works, what to avoid, do not repeat them') plus screenshots of top slideshows. Section C: master prompt with a system role ('You are a viral TikTok question slideshow copy engine for a couples app called Flamme'), context and output requirements (title slide, set categories, a Nano Banana image prompt). Takes the founder a full working day to build; he tests the output before handing it over. GPT writes the image prompt; Nano Banana Pro makes realistic vertical images. Winners are added by hand to the master prompt (AI edits hallucinated), creating new playbooks.

What happened

Source of the team's winning formats and output of ~200 pieces a day.

In their words

An Nayal - Full App Growth Guide 2026 - How to Make Viral Content to Get Users & Make Your App Retentive

Play from 34:25
33:58

JosephSo this one would be somewhere around, you know, 0.2 to 0.5. And then, would you have Here's another reaction one. So was there some kind of formula for for writing the hooks to this? Cuz I feel like the the actual demo is pretty straightforward. It's just, you know, showing the features of the app, but then the first 2, 3 seconds seems like it's the hardest part to figure out how to make that engaging.

34:25

An NayalWe are operational people, we are not creative people, so what I did is what you would do to operationalize the entire thing, I created playbooks. Playbook would have three big parts. Playbook is this PDF document which my team would feed into AI and get an output. What the playbook would have was three parts. The first part would be around the section A would be product and brand context. It would define the product name, category, what is it in human terms, core features, et cetera.

34:52

An NayalThen the second part would be winning script references, where we would put every single winning script we could find, we would put it as a context reference that, "Hey, these are formats which have worked in this category. This is what we want you to take from them, do not repeat them." Section C would be the master prompt, where we broke down a system role: You are a viral TikTok question slideshow copy engine for a couples app called Flamme, your role is to do this, this is the context, this is our output requirement.

35:20

An NayalAnd that is how we discovered our winners. We would keep slamming the playbook until we would have something that resonates, and then we would dig down deeper into it.

An Nayal… I would say. One is the initial prompt we set in place. We try to give it as much context as we can. I know people who go to it and say that, you know, "I want to put out a slideshow for TikTok, help me generate something." That is really shallow. I think AI is this tool which works really well the more you give it. The better context it has, the better it knows what is winning, the better it's able to give you something that wins, as well.

42:23

An NayalThis is the first thing we do well. We provide a lot of context initially. Like to create each of this playbook, it takes me one working day. It takes me a whole day to put out this entire playbook. I have to collect the right formats, I have to collect the right screenshots, I have to make sure that they are aligned with whatever product is, I have to make sure that I curate the master prompt to firstly, give an output that has potential to go viral, but also give an output for what we need, which means it should have the title slide, it should have these categories, it should have a Nano Banana prompt as well.

42:49

An NayalOnce I do all of this, then I put that into AI, and then I test the prompt to see what is the output coming out. If it comes out well, then I pass it on to the team, if it doesn't come out well, then I go all over the process again. This is one of the key steps, which is setting it right in the first go.

Get tactics like this every Monday

New tactics from the week's founder interviews, each linked to where it was said.

More from this episode

Tags: tiktok, ai-content, prompt-playbook, master-prompt, winning-scripts, nano-banana