How to Create AI Generated Ads: The RAIO Energy Drink Case Study

A fake UGC photo, a studio hero shot, three footballers with reference sheets, photorealistic storyboards and 15 second film sequences — a complete beverage campaign for a fictional can.

RAIO ENERGY does not exist. There is no can on any shelf, no factory, no brand team. And yet by the end of this canvas there is an amateur iPhone photo of the can on someone's kitchen counter, a studio hero shot, a cast of three street footballers with reference sheets, storyboards, and a set of 15 second cinematic film sequences of a golden hour street match in a Brazilian hillside neighborhood — the complete asset stack of a beverage marketing campaign, built end to end with AI on one board.

The RAIO Energy Drink Ad Template on Zoviz Canvas is a study in how AI generated ads are actually produced when the goal is craft rather than novelty: 33 prompt nodes covering two styles of product photography, multi view character sheets, photorealistic storyboard grids, director level sequence prompts and structured shot lists. This article walks the pipeline stage by stage, with the real prompts.

Stage 1: one can, two kinds of truth

The product photography in this canvas runs in two opposite directions at once, and the pairing is the smartest thing in it.

Everything a photographer would fix — the off center framing, the hotspot, the grain — is deliberately requested. Imperfection is the aesthetic of authenticity, and this prompt treats it as a spec.

The second node goes the other way, transforming the can into commercial AI product photography:

Note the discipline in the first line: keep the branding exactly, change only the photography. That consistency starts even earlier when you create a brand name with ai and establish a clear identity before generating campaign assets. The canvas also carries four uploaded reference images feeding the pipeline, so generated shots stay anchored to one consistent can design throughout.

Stage 2: a cast with faces — and reference sheets to protect them

Where our luxury watch film case study hid every face, this commercial does the opposite: the hero's gold capped tooth is practically the logo of the film. The consistency strategy changes accordingly — each of the three characters gets a photorealistic multi view reference sheet:

The cast is a trio built for coverage: Relâmpago the hero (raio means lightning in Portuguese — the brand and the man share an identity), a wiry neon-haired dribbler, and a tall bald keeper with a long thin braid, mismatched gloves and a warpainted face. Distinctive silhouettes, distinctive props: every trait doubles as a recognition anchor when the AI regenerates them shot after shot. The same principle applies to a brand kit, where consistent colors, typography, logos and visual elements help keep the brand recognizable across every campaign asset.

Stage 3: photorealistic storyboards

Before any video renders, the film exists as storyboard grids — but unlike hand drawn boards, these are pitched as finished frames:

Eight of the 33 prompts are storyboard sheets. As in every serious canvas we've documented, the economics are the point: scenes get designed, judged and locked as single cheap images before the video nodes spend real credits.

Stage 4: the film in 15 second sequences

The film prompts read like a director's treatment. The opening:

And the climax closes the loop from sport back to product:

Each treatment is paired with a JSON shot list the video node consumes — indexed shots with their own framing, lens and duration ("Extreme close-up, locked-off tripod shot, cropped from mid-shin down. The hero's scuffed worn cleats and mismatched socks tap a worn leather ball…").

Between the big sequences sit insert packs — macro footwork studies, and a shot from directly behind the goal net with the netting sharp in the foreground — exactly the B roll a real commercial shoot would grab for the edit.

One production decision runs through every sequence: no music, no dialogue. The prompts ask only for diegetic sound — can crack, ball impact, crowd roar — and explicitly leave the sequences "open for post-production sound design." AI generates the picture and the atmosphere; the final mix stays a human craft decision.

What this template teaches about AI ads

1. Authenticity is a spec, not an accident. The fake iPhone photo works because its flaws are itemized in the prompt. If your ad strategy includes UGC style creative, write the imperfections in.

2. One branding, many registers. The same can appears as a kitchen counter snapshot, a studio hero shot and a prop in cinematic film — the range a real beverage brand needs across TikTok, retail and broadcast, all locked to the same reference design. An AI logo generator can help establish that visual identity from the start, giving the rest of the campaign a consistent brand foundation to build on.

3. Choose your consistency strategy per project. Hidden faces worked for the luxury watch film; multi view reference sheets with signature details (gold tooth, warpaint, neon hair) work here, where faces carry the emotion.

4. Fifteen seconds is the unit of work. Every sequence is scoped to what current video models sustain, and the campaign is assembled from sequences plus insert packs rather than one long generation.

5. Leave room for post. Foley listed, music absent, sound design deferred. The best AI generated ads treat the model as a cinematographer, not the whole crew.

Who this is for?

Beverage and CPG teams get the obvious blueprint: energy drink marketing lives on exactly this asset mix of product shots, athlete adjacent film and social ready verticals. Agencies and freelancers get something subtler — a way to pitch a full campaign for a brand, real or invented, before a single production dollar is spent. And indie founders get proof that a brand can be art directed into existence: RAIO's can, cast, world and film language were all decided in prompts, which means yours can be too.

FAQ

Can AI generate a complete ad campaign?

This canvas produces the creative assets — product photography in multiple styles, cast, storyboards and film sequences. Strategy, media buying, final edit and sound mix remain human work, and the template deliberately leaves the last two open.

How do you make AI product photography match a real product?

Two mechanisms here: reference images uploaded into the canvas anchor the can design, and every transform prompt orders the model to keep shape, colors, logo and text unchanged while only the photography changes.

Are fake UGC ads ethical?

Styled amateur photography is a long standing advertising technique, but disclosure rules for AI generated content vary by platform and country. Present AI creative as advertising, follow platform disclosure requirements, and never pass generated content off as real customer testimony.

Why are all the sequences 15 seconds?

That is the practical ceiling where current AI video stays coherent per generation, and it matches how vertical ads are consumed. The campaign cuts together from 15 second sequences and shorter insert packs.

Do I need a real product to use this template?

No — RAIO itself is fictional. Start from a described can as this canvas does, or drop your real product photos into the reference nodes and let the transform prompts do the rest.