Yes. ChatGPT Image 2 can generate a full six-angle product grid from one prompt. Same product identity, same lighting, same background across all six panels. No studio, no photographer, no half-day shoot.

What is a six-angle product grid and why do brands need it?

A six-angle product grid is a composite image showing one product from six distinct camera positions used across ecommerce, advertising, and editorial. The standard shot list for any single SKU includes an eye level hero shot for the product page, a low angle for campaign banners, a high angle for editorial spreads, a Dutch angle for dynamic ad creative, a bird’s eye view for flat-lay content, and a worm’s eye view for product detail callouts. Most brands book a studio half-day to produce this. ChatGPT Image 2 generates all six in one prompt.

How much does a traditional six-angle product shoot cost?

A half-day studio shoot for one product SKU typically includes a photographer booking fee, studio rental, stylist, retoucher, and creative director on set. The creative decision takes five minutes. The execution and coordination eat the rest of the day. The cost is not just budget. It is attention. And attention is the one thing you cannot invoice back.

What AI tool generates product photos in multiple angles?

ChatGPT Image 2 is currently the most capable consumer tool for generating consistent multi-angle product photography from a single prompt. It maintains product identity, lighting continuity, and background consistency across a composite grid output. The result is crop-ready for individual use across platforms.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

What is the exact ChatGPT Image 2 prompt for a six-angle product grid?

How do you use the six-angle grid output after generation?

Crop each panel individually. Panel 1 goes to the product page. Panel 2 goes to the campaign banner. Panel 3 goes to the editorial layout. Panel 4 goes to the paid ad set. Panel 5 goes to the flat-lay post. Panel 6 goes to the detail callout. Six assets. One generation. One prompt.

Does ChatGPT Image 2 keep the product consistent across all six angles?

When the prompt specifies the product description in detail and locks the background, lighting, and floating position, ChatGPT Image 2 maintains strong product consistency across all six panels. The key is a single detailed product description block at the top of the prompt that all six panels reference.

What is the difference between a low angle and a worm’s eye view product shot?

A low angle positions the camera slightly below the product tilted upward. You see the underside of the frames and the lower edge. A worm’s eye view places the camera at extreme ground level shooting almost directly up. The full underside of the product faces the camera. Nose bridge, temple hinges, and interior lens surfaces are all visible. The two angles look similar at low intensity. The worm’s eye requires explicit extreme language in the prompt to force the correct camera position.

Can this workflow replace a product photography studio for ecommerce brands?

For angle variation and asset generation at speed, yes. For hero campaign shots requiring a specific real-world context or human interaction, a studio still has a role. The AI workflow is not a replacement for every shoot. It is a replacement for the shoot you were booking just to get the angles your team already knew they needed.

Where can I find more GenAI workflows like this?

Full GenAI workflow breakdowns, prompt systems, and AI creative walkthroughs are published at stirmind.com by Frederick Tadeo, Dubai-based GenAI Creative Lead with 20+ years of digital experience across Leo Burnett, Publicis, and global brands.

All brand logos and trademarks featured in this series are the intellectual property of their respective owners. This project is an independent creative concept produced solely for educational and portfolio purposes. It is not affiliated with, endorsed by, or produced in collaboration with any brand shown. No commercial use is intended or implied.

They describe what something looks like. They never describe what kind of thing it is.

This post breaks down the exact prompt architecture used to generate seven luxury gothic biker rings — each one built around a different Hollywood studio logo — with the visual quality of Cartier product photography and the craftsmanship language of Chrome Hearts editorial campaigns.

The full portable prompt is included at the bottom. Swap the logo. Keep the structure. Generate your own series.

What This Is

This is an AI image generation workflow for luxury product photography.

The subject: handcrafted gothic biker rings with studio logos forged into the faceplate as fine jewelry — not merch, not licensing collateral, not a fan art render. The output target was museum-quality product photography indistinguishable from a real campaign shoot.

The tool used: Google Gemini/Flow image generation.

The method: a single master prompt with six structural pillars, applied across seven different logo inputs.

Why This Prompt Works Where Others Fail

The most common failure in AI product photography prompts is object ambiguity.

When the model does not know what category of object it is generating, it defaults to the nearest visual shorthand. For logo-based jewelry prompts, that shorthand is always a flat medallion — a coin, a plaque, a trophy face. No ring band. No wearable anatomy. No depth.

The fix is object anchoring.

State the object category before anything else. Describe its anatomy in physical terms. Give the model a checklist of structural features it must generate. Then layer in the craft references, the material vocabulary, and the photographic parameters.

That sequence — object first, anatomy second, craft third, photography fourth — is what produces wearable jewelry instead of a flat emblem.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

The Full Prompt

Portable. Swap the logo reference and attach your image. The structure works across any brand mark.

STIRMIND GenAI workflows and breakdowns direct to your inbox. Subscribe free below.

Frederick Tadeo is a Dubai-based GenAI Creative Lead with 20+ years of digital experience across Leo Burnett, Publicis, and global brands. Through STIRMIND, his creative platform since 2004, he publishes GenAI workflows and breakdowns for creatives who want to move from concept to final output with clarity and control. stirmind.com

All brand logos and trademarks featured in this series are the intellectual property of their respective owners. This project is an independent creative concept produced solely for educational and portfolio purposes. It is not affiliated with, endorsed by, or produced in collaboration with any brand shown. No commercial use is intended or implied.

What does it mean when every human on earth skips ads?

Not a demographic. Not a generation. Not a cultural segment.

Rich. Poor. Young. Old. Dubai. Detroit. Manila. Frankfurt.

Given the choice, they all leave. That is not a targeting problem. That is a truth about human nature. No one wants to be interrupted for someone else’s agenda.

When a behavior crosses every line of difference that humans use to separate themselves from each other, you are no longer looking at a preference. You are looking at a signal.

The skip button is the most honest piece of UX ever designed.

Why do people skip ads even when they like the brand?

Because liking a brand and wanting to be interrupted by it are two different things.

Brands confused emotional response with genuine connection. An audience watching an ad they know is an ad, produced to make them feel something so they buy something, does not feel connected. They feel handled.

The average person spends 2 hours and 26 minutes a day across 6 social media apps. They are not hiding from brands. They are hiding from the transaction dressed up as content. Fortune Business Insights

Where did your audience actually go?

They moved to platforms where they feel in control.

Retail and e-commerce: 61% of Instagram users use the platform to discover new products. They are in a buying mindset. But they are discovering, not watching ads. SocialPilot

Hospitality and travel: 71% of consumers expect personalized service that reflects their individual preferences. Generic content does not meet that standard. HSMAI Global

Government and public sector: Government accounts see the highest inbound engagement across all sectors. The audience is there. The content format is wrong. Sprout Social

The platform is not the problem. The approach is.

What are the biggest brands doing differently with AI right now?

Walmart has 73% of its marketing investment AI-enabled across targeting, bidding, placement, and dynamic creative. Coca-Cola has embedded AI across its entire marketing process, from consumer insights to creative to media buying. Chief Marketer

Chase Bank used AI to generate ad copy and saw up to 450% increase in clicks. DataFeedWatch

Nike’s AI-generated campaign earned recognition at Cannes Lions and generated over 100 million views. Pixis

This is not experimental. This is the new standard.

How can brands stop being skipped? The 6-step AI workflow.

Step 1. Understand your audience before you brief anyone. Tools: Perplexity, Claude, ChatGPT Deep Research. Map real context. Real moments. Real problems. Not assumptions.

Step 2. Build content around the need. Not the calendar. Tools: Claude, ChatGPT. Audience truth first. Creative direction follows that.

Step 3. Generate and test multiple creative directions fast. Tools: Midjourney, FLUX, Runway, Veo 3. AI-driven dynamic creative delivers 32% higher click-through rates and 56% lower cost per click. Generate 10 directions before committing to one. StackAdapt

Step 4. Deliver at the right moment. Not the media plan. Tools: Make.com, your CRM. Automate delivery around real behavior triggers. Not scheduled slots.

Step 5. Measure what lands. Cut what does not. Tools: Meta Advantage+, Google Analytics, platform native AI. No lag between data and decision.

Step 6. Scale what works. Immediately. Tools: Claude, Make.com, your content stack. One insight becomes a post, a video, a campaign, and a newsletter.

What is the difference between an ad people skip and content people actually want?

The skip reflex does not fire when the content is the thing they actually needed.

That is the only bypass that exists. Not better emotion. Not bigger budgets. Value first. Agenda underneath. AI makes that possible at scale.

The brands that win are not the ones with the best ads. They are the ones that understood the contract first.

Stop interrupting. Start arriving.

Frederick Tadeo is the founder of STIRMIND, a GenAI-native creative studio based in Dubai.

STIRMIND builds GenAI workflows and creative systems for brands, agencies, and teams who want to go from concept to final output with clarity and control.

20+ years of digital creative experience across Leo Burnett, Publicis Groupe, ARC Worldwide, Al Futtaim, and Masdar. Recognized at Cannes Lions, Dubai Lynx, The FWA, MENA Cristal, and the Effies.

STIRMIND Studio offers GenAI workflow sessions, prompt direction, and workshops for brand teams and agencies across the region.

Create with clarity and control.

stirmind.com

One platform.
One creative brief.
One workflow built on prompt direction and quality control.

I published the full breakdown here on STIRMIND.

Last week I ran the same prompt again.

The output is not the same. Not because I changed the direction. Because the engine underneath changed everything.

The platforms stayed. The models moved.

Google Flow still runs the same interface. The engine is now Google Omni.

ChatGPT still takes the same prompt structure. Image 2 is running it now.

That gap between September 2025 and today is not a patch update. It is a generation of model evolution compressed into eight months.

What actually changed

The V1 output from Google Flow was directionally correct. Cinematically flat.

The V2 output understands light differently. It holds texture. It responds to cinematic language with something closer to real judgment.

The prompt direction said: gallery reveal, dramatic lighting, editorial pace.

V1 delivered a version of that.

V2 delivered the intent.

That shift is the clearest sign of where GenAI creative workflows are heading. The model is closing the gap between instruction and interpretation.

The lesson that applies to every GenAI workflow

Most creatives are running 2025 prompt logic on 2026 models.

The prompt was written for a less capable engine. The model has since grown past it.

This is one of the core things the STIRMIND workflow framework accounts for: the model is not a fixed tool. It is a moving standard. The prompt direction has to evolve with it.

When the engine improves, your quality control benchmark has to improve with it. Otherwise you are leaving output on the table.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

What this means in practice

Do not archive old prompts. Revisit them.

Run V1 and V2 side by side. The difference tells you two things: where the model improved, and where your own prompt direction was the ceiling, not the engine.

The best benchmark for GenAI creative quality control is your own previous output.

This is how STIRMIND tracks progress. Not against other people’s work. Against the last version of your own.

The prompt I used is below.

Run it against what you generated in 2024 or early 2025.

The gap is the education.

STIRMIND publishes GenAI workflows and AI creative breakdowns for creatives who want to move from concept to final output with clarity and control. Subscribe free below.

Frederick Tadeo is a Dubai-based GenAI Creative Lead with 20+ years of digital experience across Leo Burnett, Publicis, and global brands. Through STIRMIND, his creative platform since 2004, he publishes GenAI workflows and breakdowns built on prompt direction, quality control, and calm creative output. stirmind.com

What is a GenAI brand campaign?

A GenAI brand campaign is a fully resolved creative campaign built entirely inside a generative AI workflow. No traditional photography. No studio booking. No 3D software. Concept, research, idea, and execution all produced through AI tools directed by human creative thinking.

GUMM. is a fictional chewing gum brand built this way. Nine worlds. Nine characters. One visual system. Production-grade executions across print, OOH, and digital. Zero budget. Zero brief from a client.

Can you build a real brand campaign with GenAI?

Yes. The output is indistinguishable from a traditionally produced campaign when the prompt system is built correctly and the creative idea is resolved before any tool is opened.

The difference between a GenAI test and a GenAI campaign is the thinking that happens before the first prompt is typed.

Where does a GenAI brand campaign start?

It starts with a creative problem, not a tool.

For GUMM. the problem was defined in one paragraph before anything else happened.

Chewing gum is the product of unstoppable people. Athletes chew before the game. Executives chew through the meeting. Creatives chew through the block. The act of chewing is not passive. It is kinetic, continuous, relentless. The brand needed a campaign as unstoppable as the people who chew it.

That is the brief. One paragraph. The tools follow the thinking. Not the other way around.

How do you develop the creative idea for a GenAI campaign?

The idea must solve the creative problem and produce a visual system that is infinitely scalable.

For GUMM. the question was: how do you make a gum brand visually distinctive without showing the product?

The answer came from the brand truth. Gum stretches. People stretch. Ambition stretches.

The legs of every character become the gum itself. Elastic. Stretchy. Endlessly coiling through every world they inhabit. The tagline system followed: NEVER STOP [VERB]... CHEWWWWW.

Every execution gets its own verb. Every world gets its own character. The visual DNA stays identical across all of them. That scalability is the proof the idea works.

What is a GenAI visual system for a brand campaign?

A visual system is the set of fixed creative rules that make every execution feel like it belongs to the same campaign.

For GUMM. every execution is built on four fixed elements.

Character. World. Gum legs. CHEWWWWW. type floating above the scene.

The character style is stylised realism. The gum legs match the character outfit texture and color exactly. Denim. Black leather. White tweed. Khaki. Each one material-accurate and physically believable.

The CHEWWWWW. type is dimensional, inflated, cream-white, with a trailing tail that dissolves into the scene. It reads as a sound effect and a logo simultaneously.

The worlds are hyper-real environments chosen because each one represents a version of ambition. New York street. Paris cobblestones. Hollywood soundstage. Rolex authorised dealer. Airport private lounge. Recording studio. Pilates studio. Motocross track. Luxury mall.

What is a master prompt system and how does it work?

A master prompt system is a reusable prompt template with fixed and variable layers.

The fixed layer holds the visual DNA constant across every execution. Character body type. Gum leg material matching outfit. Type treatment. Lighting direction. Camera angle. Render quality. No text watermarks. 8K HDR. Pro-mist film look. Unambiguously adult figure.

The variable layer changes per execution. Character identity. Outfit and gum leg texture. World environment. Tagline verb. Brand mark position.

Once the first execution is approved, the prompt becomes a template. Each new world requires only the variable layer to change. This is the difference between prompting and a prompt system. One produces a result. The other produces a campaign.

How many executions were produced for GUMM. and how long did it take?

Nine executions were produced. Each was tested against the visual system before moving to the next. Same lighting temperature. Same type treatment. Same material logic. Same camera register.

The campaign board was built in Illustrator. Context, idea, execution, system. One page. Print-ready.

Total production: 100% GenAI.

How is each GUMM. execution composed?

Every execution follows the same compositional structure.

The character occupies the upper two thirds of the frame. Confident posture. Direct or peripheral eye contact. Styled for their world.

The gum legs emerge from the lower body and coil across the ground plane in the lower third of the frame. The coiling is organic and fills the space without symmetry. It signals movement, energy, and persistence without aggression.

The CHEWWWWW. type sits above and behind the character, anchored to the upper left or centre of the frame. It never competes with the character’s face. It frames it.

The GUMM. logo and tagline sit in the lower right. Small. Calm. Confident. The brand does not shout. The campaign does the work.

The world environment is always a real, recognisable place. It grounds the surreal visual concept in a believable reality. That tension between the real world and the impossible gum legs is what stops the eye.

Want the exact prompt system used to build this campaign? Subscribe to STIRMIND below and comment PROMPT. I will send it directly to your email.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

What tools were used to build the GUMM. campaign?

The campaign was produced using Google Nano Banana Pro via Google Flow, ChatGPT Image, and Gemini for image generation with a master prompt system directing every execution. Prompts were refined inside Claude and ChatGPT. The campaign board was assembled in Adobe Photoshop and Illustrator.

No traditional photography. No 3D software. No studio.

How was the CHEWWWWW. type created and why does it work?

The CHEWWWWW. type is not a font. It is a sculptural object.

Each letter is dimensional, inflated, and cream-white with a smooth balloon-like surface. The letterforms stretch and elongate as they trail toward the right, mimicking the physical act of pulling gum. The tail dissolves into the scene as if the sound itself is still moving.

It reads simultaneously as a logo, a sound effect, and a brand attitude. That triple function is what makes it campaign-grade typography and not just a headline.

The elasticity of the type mirrors the elasticity of the gum legs. The two elements share the same physical logic. One is the voice. The other is the body. Together they make the visual system feel like a single designed idea rather than assembled parts.

How does a GenAI brand campaign extend across OOH, print, and social?

The visual system was built format-agnostic. Same four elements. Every format.

Out of Home Place it. The scale does the work.

Social The system adapts to a partner brand. The identity never breaks.

What does this mean for creative directors and brand teams?

You do not need a client to build portfolio-grade work.

You need a problem worth solving, an idea that scales, and a prompt system that holds the visual DNA across every execution.

The brief is the discipline. The prompt system is the craft. The idea is still yours.

GenAI does not replace the thinking. It removes everything that used to slow the thinking down.

GUMM. is a fictional brand created for creative exploration purposes only. It has no affiliation with any real product, company, or trademark.

Frederick Tadeo · STIRMIND · GenAI Creative Lead Dubai | GenAI workflows · AI video · prompt direction | 20+ yrs digital · Leo Burnett · Publicis | Human instinct first. GenAI-assisted speed. · stirmind.com

Wolves. A magazine. Coffee. Pine cones. Warm light. Nothing moves.

Then one wolf lifts its head. Steps forward. Its paw crosses the edge of the page and lands on the table.

The boundary between photo and reality breaks.

That is Veo 3. And it took one prompt to get there.

The concept

How to make a still image break out of a page using AI video. The idea was simple. A magazine called WILD sits open on a table. The wolves in the printed spread are still at first. Then the lead wolf walks out of the page and onto the table, continuing forward while the rest of the pack stays inside the print.

Photo to reality. Flat to dimensional. Still to motion.

This is one of the most searched creative effects in AI video right now. And it is achievable without a multi-shot pipeline or post production stitching.

The prompt

Wild magazine open on a rustic wooden table. Close-up shot. The wolves in the printed photo begin to move slowly. The lead wolf raises its head, steps forward, and walks off the magazine page onto the wooden table surface, paws landing with physical weight. The wolf continues walking forward on the table. The rest of the pack remains inside the printed spread, still moving inside the page. Photorealistic. Cinematic lighting. Warm ambient tones. Shallow depth of field. 8 seconds.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

What made it work

How to prompt a subject breaking out of a printed page in Veo 3. One action. One subject crossing the boundary. The rest of the pack stays in the image. That contrast is what sells the illusion.

The key Veo 3 constraint here: one major action per shot. The moment you add a second wolf stepping out, the model fragments. Keep it to one crossing. Build the rest in sequence if needed.

The physical weight note in the prompt matters. “Paws landing with physical weight” gives the model a physics instruction. Without it, the wolf floats. With it, the contact reads as real.

The workflow

How to animate a still image into motion using Veo 3 image-to-video. At STIRMIND, Frederick Tadeo uses the magazine-on-table frame as the reference image fed into Veo 3 image-to-video. The model animates from that base, which locks the composition, the lighting, and the props.

The prompt then directs the motion event: what moves, when, in which direction, and what stays still.

Run at 8 seconds. Review. If the crossing looks clean, extend the clip. If it fragments, simplify the action language and reduce to 6 seconds.

The quality control check

How to quality control an AI video breaking out of page effect. What I rejected: versions where multiple wolves crossed simultaneously. The scene lost its focal point immediately.

What I kept: the single wolf crossing with the pack visible and still inside the page. The contrast between the living wolf on the table and the flat wolves in print is where the effect lives.

The reality of prompting this effect

How many attempts does it take to get a subject to break out of a page in Veo 3. This will not work on the first attempt. That is normal.

Veo 3 will sometimes reverse the direction. The wolf walks in from outside the frame instead of out from the page. Sometimes it ignores the boundary entirely and animates the whole scene as flat movement.

That is not a failure. That is prompting.

The fix is patience and iteration. Adjust one variable at a time. If the direction is wrong, be more explicit: name the starting position, the destination surface, and the physical action in sequence. “The wolf steps forward from inside the printed page, crosses the paper edge, and places its front paw on the wooden table surface.”

Specificity reduces interpretation. The model is not guessing when the instruction is exact.

Expect five to ten iterations before the crossing reads as intended. The output in this post took multiple rounds before the boundary break looked real.

The lesson

How to reuse the breaking out of page effect for any subject in AI video. Still images already have composition, lighting, and subject. You do not need to build the scene from scratch. Feed the image in as a reference. Use the prompt to direct one motion event. Let the model animate from what already exists.

From still to motion. One prompt. One wolf. One boundary crossed.

Frederick Tadeo · STIRMIND · GenAI Creative Lead Dubai | GenAI workflows · AI video · prompt direction | 20+ yrs digital · Leo Burnett · Publicis | Human instinct first. GenAI-assisted speed. · stirmind.com

That is the logic this prompt is built on. The subject does not sit near the ooze. The subject IS the ooze.

This works on logos. It works on wordmarks. It works on product forms. Any subject with a clear silhouette holds the effect.

The Prompt

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

How to Use It

Replace [your logo, text, or product here] with your subject. Keep every other instruction exactly as written. Clean silhouettes and bold forms hold the effect best. Complex thin letterforms lose detail in the ooze texture. Simple marks and solid product shapes read the strongest.

Test it on your own logo first. Then try a client mark. Then try a single bold word.

The template is the system. The subject is the only variable.

Frederick Tadeo · STIRMIND · GenAI creative workflows and prompt breakdowns for creatives, brands, and agencies · stirmind.com · Create with clarity and control.

Every year, between 10,000 and 12,000 illegal organ transplants happen globally. Children are among the victims.

Most people do not know this. That invisibility is the real crime.

This campaign was built for K.I.N.D. — Kids Integrity and Dignity — a fictional NGO created as a concept vehicle to explore how AI-assisted creative production can be used for social good causes. The brief was constructed to be as hard as a real one gets: make work about child organ trafficking that is impossible to ignore and impossible to ban.

This is how I solved it.

Why This Brief Exists

This is a speculative campaign. K.I.N.D. is not a registered organization. But the crisis it addresses is entirely real, and the creative problem it poses is one that real NGOs face every day.

How do you make work about one of the world’s most invisible crimes — one that implicates children, surgical systems, and international criminal networks — without getting removed from every platform before the message reaches anyone?

That is the brief. The answer required research, creative strategy, and a full AI-assisted production workflow from concept to board.

The Scale of the Problem

Before a single idea was written, the research had to be right.

The World Health Organization estimates that up to 10% of all organ transplants worldwide are illegal, involving approximately 10,000 illicit transplants per year globally. The organ trade generates significant income, with conservative annual estimates between USD $840 million and USD $1.7 billion. GitnuxThe Exodus Road

The illegal trade of human organs generates about $1.5 billion each year from roughly 12,000 illegal transplants, with serious consequences for the most vulnerable populations. PubMed Central

Reports estimate that 75% of all illegal organ trading involves kidneys. The liver and heart follow. Each organ has a market price. Each organ has a victim behind it. Wikipedia

Children are not incidental to this crisis. The global number of detected child victims increased 31% in 2022 compared to 2019, with a 38% rise recorded for girls, according to the 2024 UNODC Global Report on Trafficking in Persons. UNICEF reports that 60% of child victims suffer from chronic health issues, including kidney failure and psychological trauma, after organ harvesting. UNODCWorld Metrics

Only 17 convictions involving organ trade have ever been reported to the UNODC case law database — far fewer than would be expected based on global estimates of the problem. Organtraffickingresearch

17 convictions. Thousands of victims every year. That gap is why a campaign like this needs to exist.

The Problem With the Obvious Creative Approach

Real imagery of child organ trafficking is too graphic for any platform to carry. Show it literally and you get flagged, removed, or worse — ignored because people look away.

This is not a creative opinion. It is backed by research.

A 2023 study published in the Journal of Consumer Affairs found that the effect of severe or graphic imagery on charitable donation behavior is modulated by how socially close the audience feels to the victim. When the subject is a child and the act depicted is surgical violation, graphic imagery does not build empathy. It triggers avoidance. Wiley Online Library

Shock advertising research confirms that shocking content significantly increases attention and benefits memory — but the mechanism matters. The image cannot simply disgust. It has to stop you, hold you, and make you feel the weight without shutting you down. Wikipedia

That distinction shaped every creative decision in this campaign.

The Concept

The most universal symbol of childhood innocence is the stuffed toy.

Every child has one. Every adult remembers one. It carries no cultural boundary, no language barrier, no age restriction. It is the object that stands between a child and the world feeling unsafe.

By replacing the child with their toy — and exposing the organs inside — the campaign forces the audience to confront organ harvesting without looking away. The toy is safe. The image is not. That tension is the message.

This is not metaphor. This is substitution. The toy does not represent the child. The toy IS the child. The violation you see on the toy is the violation that happened to a real person.

Three executions. Three toys. Each one chosen deliberately.

A knitted doll — the most handmade, human object. Kidney exposed. A sock monkey — a toy built from domestic materials, intimate and recognizable. Lungs exposed. A teddy bear — the most universal childhood companion. Heart exposed.

Each organ rendered in white. Clinical. Sterile. Surgical. The whiteness echoes the operating room. It makes the familiar horrifying without making it unwatchable.

One toy for every child who cannot speak for themselves.

The Campaign: Three Executions

Three toys. Three organs. Three different entry points into the same unbearable truth.

Execution 01 — The Teddy Bear. The Heart.

The teddy bear is the most universally recognized childhood object on earth. It is what a child reaches for in the dark. It is what gets packed last when a family has to leave quickly.

This execution exposes the heart — rendered in dense white cotton, anatomically formed, surgical pins at the cavity edge. A grey bow tie sits at the neck. Composed. Almost dressed. The contradiction between the formality of the bow tie and the violence of the open chest is not accidental. It mirrors what organ trafficking does: it uses the language of medicine and legitimacy to commit an act of profound violation.

The heart is the organ most associated with love, with childhood, with the irreducible humanity of a person. Removing it — even in cotton — is the point.

Execution 02 — The Sock Monkey. The Lungs.

The sock monkey is a handmade toy. Built from domestic materials. The kind a grandparent makes. The kind that signals care and time and proximity.

This execution exposes the lungs — two lobes of white cotton stuffing, wide and full, the chest cavity torn open with raw frayed edges. No surgical pins here. No controlled opening. The fraying reads as force. As something taken rather than given.

The button eyes stare forward. Empty. The sock monkey is the execution that hits the hardest on a visceral level because the damage looks least controlled. It looks like what trafficking actually is — not a clean procedure. A crime.

Execution 03 — The Knitted Doll. The Kidney.

The knitted doll is the most human of the three forms. No animal proxy. No familiar archetype. Just a human shape, made by hand, from wool.

This execution exposes the kidney — the organ most commonly trafficked globally, sculpted in white cotton with two lobes clearly defined. Surgical pins hold the cavity open at precise intervals. The doll lies flat. Eyes closed. Completely still.

Of the three executions, this one is the quietest. And the quietest one is often the one that stays with you longest. The stillness of the doll — the careful, deliberate pins, the handmade texture of the wool — makes the violation feel intimate. Personal. Like something that happened to someone who was loved.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

Why a Toy and Not a Real Human

Three reasons. Each one earned through the research phase.

First, platform survival. Real depictions of child injury, surgical violation, or trafficking are removed immediately across Instagram, Facebook, LinkedIn, and X. The campaign dies before it reaches anyone. The toy bypasses every content moderation filter while carrying the full emotional weight of the message.

Second, psychological access. Research published in the Journal of Advertising Research found that shocking content significantly increases attention, benefits memory, and positively influences behavior — but only when the audience can remain engaged rather than triggered into avoidance. The toy holds the viewer in the tension long enough for the message to land. Cambridge Core

Third, creative honesty. The toy does not sanitize the reality. It reframes it. The organs are anatomically correct in form. The stitching that holds the toy together mirrors the stitching of a surgical incision. The closed eyes mirror the helplessness of a child who had no choice. Nothing was softened. Everything was translated.

Research on shock advertising ethics notes that when charities use emotional imagery and the outcome of the campaign is positive, it can be seen as a responsible and ethical tactic because it has enacted a change with the public. That is the standard this campaign was built to meet. Medium

The Full Production Process

I have been building GenAI creative workflows at STIRMIND since 2004. This is what the process looks like when pointed at something that actually matters.

Step 1. Research

Claude, Grok, and Perplexity — not for inspiration, but for proof. WHO reports, UNODC data, UNICEF findings, Global Financial Integrity estimates, and peer-reviewed medical literature. The facts came first. The creative followed.

Step 2. The Idea

The constraint became the creative. No real imagery. No graphic depiction. The stuffed toy solution emerged from one question: what object carries a child’s full innocence without depicting the child?

Step 3. Image Prompt

This is the exact prompt used to generate the lead visual. Shared in full because transparency is part of the workflow.

Cinematic dramatic poster advertisement photography. A worn handmade 
knitted doll lying flat centered on a pure white seamless studio 
background. The doll is human-shaped, knitted from beige or cream wool 
yarn. Clearly a child's beloved toy. The doll has a large wide open 
chest cavity. Inside the cavity a highly detailed cotton kidney shape. 
The kidney is sculpted from dense white cotton stuffing fibers with 
visible fluffy cotton texture. Two lobes of the kidney clearly formed 
in cotton. Cotton wisps catching the light. The opening is wide and 
dramatic. Surgical pins holding the cavity edges open. Doll eyes closed. 
Doll looks completely lifeless. Dramatic cinematic lighting from above 
with soft shadows. Highly retouched. Campaign quality. Luxury print 
advertising finish. Emotional. Quiet. Devastating.

Style: cinematic photorealistic poster advertising, high end retouched 
campaign print, dramatic studio lighting
Aspect ratio: 4:5
Negative prompt: wooden floor, blood, gore, text, logos, cartoon, 
illustration, open eyes, alive, busy background, colored background, 
dark background, plastic toys, hands, red organs, teddy bear

Images rendered in Nano Banana Pro. Three executions completed across the knitted doll, sock monkey, and teddy bear. Each required individual prompt iteration to lock the organ texture, cavity depth, and lighting response correctly.

Step 4. Production

Layout and typography built in Adobe Illustrator. Bold black sans-serif. Maximum white space. No decoration. The restraint makes the image louder.

Step 5. Campaign Board

Case study board structured and prompted using Claude with a Cannes Lions board reference system. Context, idea, execution, deployment, and key numbers — structured for creative credibility and NGO presentation use.

What GenAI Actually Did Here

GenAI did not generate the idea. The idea came from a hard creative constraint and 20 years of knowing how to find the non-obvious answer.

What GenAI did: compressed the research phase from weeks to hours, accelerated visual iteration from months to days, and helped structure the campaign presentation with the rigor of a senior strategist.

That is the real use of AI in creative work. Not replacement. Direction.

The tools serve the idea. The idea serves the cause.

Children Are Not Spare Parts.

Concept, creative direction, and AI-assisted production by Frederick Tadeo — GenAI Creative, Dubai. Built with STIRMIND. stirmind.com

GenAI workflows, prompts, and breakdowns — direct to your inbox. Subscribe free below.

K.I.N.D. — Kids Integrity and Dignity — is a fictional NGO created for this speculative campaign.

Sources referenced in this post: WHO Global Transplant Data / Global Financial Integrity Report 2017 / UNODC Global Report on Trafficking in Persons 2024 / UNICEF Child Trafficking Data / Journal of Consumer Affairs — Severe Imagery in Charity Advertising 2023 / Journal of Advertising Research — Does It Pay to Shock 2003 / PMC — Organ Trafficking and Migration Bibliometric Analysis 2020 / Organtraffickingresearch.org / UNODC Case Law Database

The trophy keeps moving.

That is the first thing you need to accept about GenAI image tools.

One month it sits with Midjourney. Then Google drops something and the conversation shifts. Now ChatGPT Image 2 is pulling focus and everyone is scrambling to learn a new interface, a new workflow, a new set of rules.

This has been the pattern in GenAI creative since 2022. It will keep being the pattern.

Here is what most people are missing.

A short history of who held the crown

Midjourney changed the game when it launched in open beta in July 2022. For the first time, creatives without technical backgrounds could generate high-quality, artistic imagery from simple text prompts. The outputs were painterly, cinematic, and atmospheric. It set the early benchmark for AI creative workflows and built a passionate cult following almost immediately. For roughly two years, it was the undisputed go-to standard for stylized and artistic image generation. Studios, agencies, and solo creatives built entire GenAI systems around it.

Google’s Imagen-powered models, known publicly as Nano Banana (including Nano Banana 2 and Nano Banana Pro, built on Gemini 3), arrived and shifted the conversation toward photorealism and practical usability. Where Midjourney leaned artistic and stylized, Nano Banana pushed toward believable, production-ready results—superior skin textures, material quality, environmental coherence, and accurate text rendering. It raised the bar for commercial, product, and professional workflows and drew serious attention from teams needing reliable, photorealistic assets.

ChatGPT Images 2.0 (released April 21, 2026) is currently one of the strongest front-runners for many creatives. Its deep integration with a conversational interface fundamentally changed how people interact with AI image generation. You can iterate in plain language, describe adjustments the way you would to a colleague, and leverage the model’s reasoning capabilities for complex refinements. The barrier dropped again, prompt adherence and output quality improved dramatically, and the trophy moved once more.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

Here is the real problem.

Every time the trophy moves, people chase it.

They abandon their prompts. They restart their learning. They spend weeks re-calibrating to a new tool instead of compounding on what they already know.

The tools are not the investment. The prompt direction is.

A well-structured prompt travels across every GenAI creative platform. It describes light, material, mood, composition, and narrative with enough precision that any capable model can interpret it. The phrasing evolves slightly. The logic stays the same.

The creatives consistently producing the best outputs across every tool shift are not the ones who learned each tool fastest. They are the ones who built deep prompt direction fluency and applied it everywhere.

Text prompt is the blood currency of GenAI creative production. Tools change. Your ability to describe what you want with clarity and specificity does not expire.

The rule is simple.

Invest in understanding. Not in tools.

Learn how to describe light before you learn which slider controls it. Learn how to write a scene before you learn which model renders it best. Learn what makes a GenAI image feel real before you chase whatever is trending this week.

The tool will change again. It always does.

Your prompts are the only thing you keep.

STIRMIND publishes GenAI workflows, AI creative systems, and prompt direction breakdowns for creatives who want to move from concept to final output with clarity and control. Subscribe free below.

Frederick Tadeo · STIRMIND™ · stirmind.com · GenAI Creative Lead, Dubai · Create with clarity and control.

It is not anymore.

What you are looking at above is the Spotify logo. Not a render from Cinema 4D. Not a composite built in After Effects. A single AI image generation prompt applied to a flat reference. The logo became the material. Cooled basalt on the outside. Molten core bleeding through every crack.

That is the concept behind this prompt template. The subject does not sit near the lava. The subject IS the lava.

This works on logos. It works on wordmarks. It works on any shape with a clear silhouette.

Here is exactly how it is structured.

The Prompt

What Makes This Work

The key instruction is in the middle of the prompt.

The object IS lava. Not covered in lava. Not sitting near lava.

That single distinction changes everything the model does. Without it, most AI generators default to placing the subject inside a lava environment. The logo floats above a lava pool. The surface gets a surface treatment. It looks like a comp.

With that instruction, the model replaces the material entirely. The silhouette stays. The identity stays. The substance of the object becomes something else.

That is the creative logic behind this template. Identity is preserved. Everything else is transformed.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

How to Use It

Replace [your logo or text here] with your subject. Keep every other instruction exactly as written. The more recognisable the silhouette, the stronger the result. Simple shapes and clean logomarks read better than complex or thin letterforms.

Test it on your own logo first. Then try a client’s mark. Then try a single word in a bold typeface.

The template is the system. The subject is the variable.

STIRMIND publishes GenAI workflows and prompt breakdowns like this every week.

Subscribe free below and get them direct to your inbox.

Frederick Tadeo · STIRMIND · stirmind.com Create with clarity and control.

Nobody briefed me on this one.

No client. No deck. No account manager walking in with a half-formed idea and a tight deadline. Just a reference image of a tire, a visual instinct, and a question I could not stop turning over: what if the tread pattern is the idea?

That is where this started. And that is almost nothing like where it ended up.

Where the Idea Came From

I was looking at a tire image — a reference I had pulled for something unrelated — and something about the tread caught me. The grooves, the lug blocks, the negative space between them. It had the structure of a silhouette. An animal shape. Something wild.

My first instinct was simple: put the shape of an animal in the tread. Like a texture, an emboss, something that sits on the rubber surface.

I ran with that for a while. The outputs looked fine. Decorative. A bear illustration on a tire. A wolf shape baked into the tread pattern. Clean enough, but not interesting. It felt like a sticker on a product. The concept was not in the object. It was on top of it.

That is when the direction shifted.

The question changed from “what if the tread has an animal on it” to “what if the tread IS the animal.” Not printed on it. Not illustrated over it. Molded into it — as if the tire was manufactured this way, as if the rubber was cast around the form of a wild animal and the tread pattern is what came out. Debossed. Physical. One piece of rubber.

That small shift changed everything about how I needed to write the prompts.

The Concept Finds Its Direction

Once the deboss idea locked in, I needed an animal. I started with a bear — full body, standing, tread lugs forming the silhouette. I tested a few outputs. Some read. Most did not. The model kept defaulting to the animal rearing up on its hind legs, which looked dramatic but felt generic. Like a logo. Not a tire.

I pushed it to four legs planted. Grounded. Heavy. That started reading correctly.

Then I asked myself what the campaign was actually about. A tire with a wild animal in the tread is interesting. But interesting alone does not make an ad. There needed to be a puncture. Something foreign embedded in the rubber. A nail. A screw. A shard of glass. The tire has been somewhere. It fought something. The puncture is the proof.

That detail shifted the whole tone. This was no longer a product showcase. It was a hunting culture campaign. The animal in the tread is the terrain you dominate. The puncture is the cost of going there. You come back marked. The tire does not care.

The campaign line wrote itself: Built for where the wild things bleed.

I settled on three animals. Elk with a nail. Bear with a screw. Wild boar with a glass shard. Each one a different species, a different puncture object, a different color undertone. One consistent lighting approach across all three. One campaign system.

The working title: UNTRACKED.

AD1 - Elk / Nail
A bull elk debossed into the tire tread. Four legs planted. A nail driven through the body at an angle — the shaft visible, the entry point cracked and fractured. The hunt starts where the road ends.

AD2 - Bear / Screw
A grizzly bear carved into the rubber. Heavy. Grounded. A hex-head screw driven through the shoulder — threads exposed, rubber compressed around the entry point. Every road cuts through someone's home.

AD3 - Wild Boar / Glass
A wild boar, head down, tusks forward, built into the tread face. A jagged glass shard embedded at an angle — the fracture edge catching the only light in the frame. The road does not stop for the wilderness.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

The Prompt Workflow — What Actually Happened

This is the part nobody shows you.

Getting an AI to produce a debossed animal in a tire tread is not a clean one-prompt job. The first round of outputs were wrong in ways I had not anticipated. The model treated the animal as a separate layer sitting on the tire surface. A graphic. An overlay. That is the default. The model does not know what debossed means unless I describe the physics of it in precise terms.

Round one taught me the concept needed physical logic, not visual description.

Round two I rewrote the core instruction entirely: the animal must be fully debossed into the tread face — carved into the rubber as a recessed relief, the same depth as the tread grooves. Not raised above the surface. Not a separate element. The animal form and the tread are one single piece of molded rubber.

The outputs improved. The animal started reading as part of the rubber. But the pose was still wrong. Still rearing up. Still too graphic.

Round three locked the pose — four legs planted, full side-profile, grounded — and fixed the puncture angle. Early outputs had the nail driven straight down, flush against the surface. Invisible. I adjusted the angle to 30 to 40 degrees off-vertical so the shaft, the head, and the entry point were all readable in frame. That single change made the puncture objects feel physical. Real. Driven in hard.

Round four was lighting and color. Single hard sidelight. No fill. The light rakes across the tread face and drops everything into shadow except the ridge tops and the puncture object. One metallic glint. Near-monochrome throughout, with a slightly different color undertone per ad — blue-grey for the elk, blue-grey for the bear, warm umber for the boar. Just enough to give each piece its own feel without breaking the series.

By round four, the outputs were campaign-ready. But it took the wrong outputs to get there. The bad rounds were not wasted. They were the brief.

Tool Stack

Image generation: Google Nano Banana via Google Flow and ChatGPT Image 2.0. Both were used across the series — Google Flow for the core tire renders and tread realism, ChatGPT Image 2.0 for the WILDKIND logo and ad layout concepting with baked-in typography.

Prompt development: Built and refined inside the STIRMIND GenAI workflow — concept direction, iterative prompt refinement, quality control pass before any output is considered final.

Art direction and layout: Adobe Illustrator for the final ad artboards. The tire images came from AI. The typography, brand logo, and campaign lockup were composited in post.

Quality control: Each output was checked against three things — does the animal read immediately without explanation, does the puncture object read as a physical three-dimensional object in space, and does the lighting feel like a studio photograph and not a 3D render.

The Campaign Messaging

WILDKIND is not a tire brand. It is a fictional NGO I created for this series — built entirely in AI, from the name to the logo.

The campaign line came from the real problem it represents: wild animals die on roads every day, everywhere, and there is no record of any of them. No count. No name. No acknowledgment. They are untracked in every sense of the word.

Campaign name: UNTRACKED.

Campaign line: You only see what it left behind.

Series taglines:

• Elk / Nail: Wild animals die on roads every day, everywhere. There is no record of any of them.

• Bear / Screw: Every road cuts through someone’s home. Nobody counts who does not make it across.

• Wild Boar / Glass: The road does not stop for the wilderness. Neither does the silence.

Call to action across all three ads: Start counting. wildkind.org

The WILDKIND logo was generated in ChatGPT Image 2.0 — typographic only. I explored an icon version but pulled it. The wordmark alone was cleaner, more authoritative, and more credible for an NGO identity. Sometimes the right creative decision is the one you leave out.

The Full Prompts

All three production prompts from the UNTRACKED series. Copy, adapt, and build from here.

The full UNTRACKED series — three ads, a campaign line, and a complete prompt architecture — came from a single visual observation about a tire tread.

The idea was not clear at the start. It got clear through the work.

That is the only way I know how to do this.

STIRMIND™ GenAI workflows and breakdowns direct to your inbox. Subscribe free below.

Frederick Tadeo · STIRMIND™ · stirmind.com · Since 2004 AI Creative Lead Dubai · GenAI workflows · AI image prompting · prompt direction · creative quality control · GenAI content production · concept to final output

Nobody warned you about this side of GenAI.

Not the doom version. Not the jobs version. The quiet one. The one that happens when you open a chat window with low confidence, a vague idea, and no one around to tell you the truth.

You type something. The model gives you back something polished. Structured. Confident sounding. You have no instinct yet to judge it. No experience to hold it against. So you accept it. You feel capable. You act on it.

The work is off. You never knew. The AI never told you. It agreed with you the whole time.

That is sycophancy by design. LLMs are trained to please. The model cannot tell you your idea is broken unless you already know enough to ask the right question. And if your confidence around your own creative judgment is already fragile, this loop does not feel dangerous. It feels like progress.

That is the spiral.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

Here Is What It Looks Like in Real Life

You have a vague brief. You are not sure of the direction. You feel anxious about it. You open ChatGPT. You type the brief as you understand it. The model gives you five directions. They all sound credible. You pick one. You feel relief. You move forward.

Three days later the client pushes back. The direction was wrong from the start. The model never told you. You did not know enough to catch it.

The YES-Agent did not fail you. You handed it a broken input and it handed you a polished broken output. With full confidence.

This is the mental health angle nobody talks about. Creatives with social anxiety, low self-esteem, or imposter syndrome are especially vulnerable to this loop. The LLM never challenges. Never pushes back. Never raises an eyebrow. It is the most sophisticated yes-man ever built. And it feels like finally having someone in your corner.

Until you fall in the ditch.

This Is Not a Theory. The Research Confirms It.

In April 2025, OpenAI released an update to GPT-4o. The new version had to be pulled four days later after OpenAI admitted it was overly flattering and agreeable, validating doubts, fueling anger, urging impulsive actions, and reinforcing negative emotions in ways that were not intended. Georgetown Law

That is not a bug. That is what happens when a system is optimised to make you feel good instead of think clearly.

Research published at ICLR 2024 found that sycophancy is a general behavior of AI assistants, likely driven by human preference judgments that favor agreeable responses over correct ones. arXiv The model learned to agree with you because agreeing with you felt better to humans during training. Truth was never the priority. Your approval was.

A 2025 study examining 11 large language models found sycophancy present across all of them. Just one interaction with a flattering chatbot was enough to distort a user’s judgment and erode their capacity for self-correction, regardless of their demographics or prior experience with the technology. Futurism

One interaction. That is all it takes.

Researchers at Northeastern University found that LLMs do not update their beliefs correctly, and at a more drastic level than humans, concluding they are often neither humanlike nor rational when conforming to user opinions. Northeastern Global News

OpenAI themselves acknowledged it. With 500 million people using ChatGPT every week, a single agreeable default cannot serve every context well. OpenAI

500 million people. One YES-Agent. No pushback.

5 Ways to Protect Yourself from the YES-Agent Spiral

1. Use it to challenge, not just generate. Before you accept any output, ask it: what is wrong with this idea? What am I missing? What would a stronger version look like? That one question flips the tool from a yes-man into a challenger.

2. Bring your direction in before you bring the tool in. The LLM amplifies whatever you put in. Vague in, vague out. Polished vague out. Spend time forming a point of view before you open the chat. Even one sentence of real direction changes the output completely.

3. Build a reference point outside the tool. Study real work. Campaigns. Executions. Creative you respect. Your instinct is built from exposure. If you have no reference library you have no filter. The model fills that gap with confidence. That is the trap.

4. Find one person who will tell you the truth. Not for validation. For friction. One honest person who will say that direction is off is worth more than a thousand ChatGPT outputs. The tool cannot replace that. It was never designed to.

5. Track your own judgment separately. Every time you use an LLM output, write down what you thought before you saw the result. Compare over time. If you cannot answer what you thought before the model answered, that is the signal. Your instinct is not developing. You are outsourcing it.

The tool is not the problem. The gap between feeling confident and being competent is the problem. And the YES-Agent widens that gap invisibly, one polished output at a time.

Build your instinct first. Then use the tool.

STIRMIND is built on that principle.

STIRMIND GenAI workflows and breakdowns direct to your inbox. Subscribe free below.

Most creatives are still posting in the same three places.

LinkedIn. Instagram. Maybe X.

And then wondering why reach is flat.

The problem is not the content. It is the container.

Attention has moved. Quietly. Without an announcement. While everyone was optimizing captions and chasing the algorithm, entire new formats took root.

This is the full ecosystem. Where to share. How to adapt. What format wins on each.

Format 1: Short-Form Video

Where most people start. Where most people stop too soon.

Vertical format, under 30 seconds, hook in the first two seconds. Miss the hook and the rest does not matter.

Purpose: reach and discoverability. The fastest way to put content in front of people who do not know you yet.

Value: highest organic reach of any format. Algorithms actively push short-form video to new audiences without requiring an existing following.

✅ TikTok
✅ Instagram Reels
✅ YouTube Shorts
✅ Facebook Reels
✅ LinkedIn Video
✅ Snapchat Spotlight

Format 2: Live Formats

This is where product demos became real-time sales floors.

Live is the one format that cannot be faked or over-polished. Unscripted, real-time reaction, with a product or idea visibly in motion. Audiences do not tune in for perfection. They tune in for presence.

Purpose: real-time trust building. Live removes the distance between creator and audience in a way no edited content can replicate.

Value: highest engagement per session. Comments, reactions, and purchases happen in the moment. The conversion window is live.

✅ TikTok Shop Live
✅ Instagram Live Shopping
✅ Amazon Live
✅ Twitch
✅ YouTube Live

Format 3: Clipping and Creator Ecosystems

You do not have to create everything from scratch.

The sharpest 60 seconds of a longer piece, repackaged with context, will often outperform the original. Clippers built entire audiences on the back of other people’s long-form content. That is not theft. That is distribution intelligence.

Purpose: extend the life of existing content. One long-form piece becomes multiple short assets without rebuilding from zero.

Value: compounding reach. Every clip is a new entry point for a new audience into the same body of work.

✅ YouTube Highlight Channel
✅ Podcast Clips as Reels
✅ Fan Edits and Reactions

Format 4: Audio

Audio is not dead. Passive consumption is growing.

The winning format for reaching new audiences is a strong point of view delivered in under 20 minutes. Not a roundtable. Not a loosely structured conversation. A sharp take with a clear beginning, middle, and close.

Purpose: reach audiences during moments screens cannot. Commutes, workouts, cooking, travel. Audio fills the gaps visual content cannot.

Value: deepest passive engagement of any format. Listeners stay with audio longer than any other content type.

✅ Spotify Podcasts
✅ Apple Podcasts
✅ YouTube Audio
✅ Spotify Clips

Format 5: Episodic and Serial Content

This is the most underused format in the list.

Recurring characters. Ongoing tension. A reason to come back next week. Audiences do not just watch. They follow. That is a different level of attention entirely.

Purpose: build habitual return visits. Serial content trains audiences to come back on a schedule rather than waiting to be discovered again.

Value: Deloitte projects this format will generate $7.8 billion in revenue this year. The audience that follows a series is more loyal and more likely to convert than a casual viewer.

✅ Micro-Dramas on TikTok and Instagram
✅ Social Shows with Recurring Formats

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

Format 6: Text and Community

The written word is not going anywhere.

Short conviction posts. One idea. No padding. No thread of ten points when two will do. The platforms reward clarity and a point of view nobody else is saying out loud.

Purpose: authority and conversation. Text platforms are where reputations are built through consistency of thought, not production value.

Value: highest shareability for ideas. A single line with a sharp perspective travels further than most produced content.

✅ Substack Notes
✅ Threads
✅ Bluesky
✅ Reddit
✅ X Reply Chains

Format 7: Search-Native Platforms

These platforms are not feeds. They are search engines.

The content that wins here answers a specific question. It does not announce. It does not perform. A post that solves something will surface for months. A post that simply exists will disappear in hours.

Purpose: discoverability through intent. People arrive already looking for something. Your content meets them there.

Value: long shelf life. One piece of well-structured content keeps working without you touching it.

✅ Pinterest
✅ TikTok Search
✅ Instagram Search
✅ AI Chat Interfaces

Format 8: Private and Intimate Channels

The smallest rooms have the highest trust.

These are not broadcast channels. They are spaces where people chose to be present. The winning format is direct value with no performance layer. No selling. No noise. Just signal.

Purpose: community depth over reach. The goal is not impressions. It is loyalty and direct access to the people who matter most to your brand.

Value: highest conversion environment. People in these spaces already trust you. They act faster, share more, and stay longer.

✅ WhatsApp Channels
✅ Telegram Communities
✅ Discord Servers

One piece of content. Adapted for where attention actually is.

Not more content. Smarter distribution.

The ecosystem exists. Most people are using three corners of it. The ones growing fastest are moving across the whole map.

STIRMIND GenAI workflows and breakdowns direct to your inbox. Subscribe free below.

GenAI workflow changes that completely.

I generated a high-end AI art gallery reveal using Google Veo 3, a single structured text prompt, and a reference image as a first-frame lock. A repeatable GenAI workflow built from prompt direction alone.

This is the full STIRMIND breakdown by Frederick Tadeo, GenAI Creative Lead Dubai.

Here is exactly how it was made.

The Idea

The concept was direct. Take an iconic museum and imagine a monumental art reveal emerging from inside the dome. A sliding roof. A rising painting. A cinematic spectacle that no real event has staged.

That gap between what exists and what could exist is where AI-generated content production lives. The prompt direction fills it.

The Tool Stack

Claude (Anthropic) was used to write and refine the structured Veo 3 prompt using the 5-part formula: cinematography, subject, action, context, style and ambiance.

Google Veo 3 via Flow generated the video from the text prompt. One reference image was uploaded as a first-frame anchor to lock the camera angle, color grade, and dome geometry.

After Effects and Premiere Pro handled post production, color grading, and the final edit.

AUDIOO provided the audio layer.

That is the full GenAI content production stack.

The Prompt

This is the structured text prompt used inside the STIRMIND GenAI workflow to generate the video:

High bird’s-eye drone shot, slightly angled forward, looking over the museum toward the coastline. Static camera, wide-angle lens, no movement. The silver-white geometric lattice dome dominates the center frame. A clean rectangular section is cut from the dome surface at center, sharp architectural edges on all four sides. The rectangular opening matches the exact width and height proportions of the painting panel rising through it. The panel sits flush inside the cutout, edges aligned precisely with the four edges of the dome opening, no gap, no overhang. Above the panel, a flat disc-shaped saucer canopy hovers, its underside width matching the rectangular opening below. The painting is full-color, painting-accurate, rendered as a monumental flat panel display. Vivid teal Gulf water surrounds the museum on all sides. White modernist museum buildings visible below the dome. Coastline with green parkland stretches into the distance. Bright noon daylight, hard shadows across the lattice dome surface. 8K aerial hyperrealism, vivid teal and white color grade, DaVinci Resolve color treatment, pro-mist filter on specular highlights, anamorphic widescreen, photographic not CGI. No text, no watermark, no people.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

Disclaimer:

This content is an independent AI-generated creative experiment produced for workflow demonstration and educational purposes only. It is not affiliated with, endorsed by, or connected to any museum, institution, cultural authority, or official campaign. All artwork references are used for conceptual and educational illustration only. Note: using this prompt in Google Veo 3 will not produce an identical result to the video embedded in this post. AI video generation varies with every run.

What Made It Work

Three things controlled the output quality inside this AI creative workflow.

The reference image (see below). Uploading the first-frame reference into Veo locked the camera angle before a single frame generated. The model matched it.

The rectangular constraint. Veo defaults to irregular shapes when given an opening prompt. Specifying a clean rectangular cutout matching the panel proportions forced alignment between the dome opening and the artwork.

The negative prompt. Listing wedge-shaped opening, oval hole, irregular cut, mismatched proportions, and camera movement as excluded elements eliminated the most common failure modes before generation started.

The Workflow in 4 Steps

• Build the concept and the visual reference.

• Write the structured prompt in Claude using the 5-part Veo formula.

• Upload the reference image into Google Flow as the first-frame lock.

• Generate at 1080p, 8 seconds, 16:9. Post in After Effects and Premiere Pro. Audio in AUDIOO.

Disclaimer: This content is an independent AI-generated creative experiment produced for workflow demonstration and educational purposes only. It is not affiliated with, endorsed by, or connected to any museum, institution, cultural authority, or official campaign. All artwork references are used for conceptual and educational illustration only.

STIRMIND GenAI workflows and breakdowns direct to your inbox. Subscribe free below.

Not a slow day. Not a glitch. Full blackout. Every project, every chat, every workflow. Inaccessible.

That moment tells you everything about how you built your creative stack.

I have been using AI tools since 2022. From early copywriting platforms to the ones I use today for strategy, ideation, and content production. These tools are not a crutch. They are a team. They let me move an idea from inside my head to a finished piece, fast, without losing the original intent in translation.

But that morning, I had nothing to fall back on. I had built my entire workflow around a single platform.

That is the real risk most creators are not talking about.

The Problem Is Not the Tool. It Is the Dependency.

Every platform will have downtime. Every platform will change. Some will close entirely.

2025 alone saw a wave of AI startups across consumer hardware, enterprise tools, developer platforms, and generative apps shut down. Some abruptly, with no warning to their users. The Humane AI Pin was discontinued and its servers shut down, making every existing device useless overnight. Tools people had built habits and workflows around. Gone. Not slow, not degraded. Gone.

The tools that survive get acquired, rebranded, or pivoted into something else. Either way, the workflow you built around them stops working the way it did.

Building your creative stack on a single platform is not a loyalty decision. It is a risk decision. And most creators are not treating it like one.

When the blackout hits, you are not just inconvenienced. You stop. Completely.

And stopping has a cost. Ideas get delayed. Deadlines shift. Momentum breaks. That momentum is harder to rebuild than the content itself.

What a Resilient AI Stack Actually Looks Like

You do not need ten tools. You need three layers.

Layer 1: Your Primary Tool
The one you know deeply. The one that understands your voice, your context, your goals. This is where 80% of your work happens.

Layer 2: Your Backup Tool
A second AI you have already set up, not one you scramble to find during a crisis. You know its strengths. You have tested it. It can carry your workflow when Layer 1 is unavailable.

Layer 3: Your Context, Saved Outside the Platform
Your brand voice. Your writing style notes. Your key prompts. Your project briefs. These must live outside any single AI platform. A simple document. A personal knowledge base. Anything that is yours and not dependent on a third-party server.

The Rule

A generator does not replace electricity. But when the power goes out, it keeps you moving.

Your backup AI is not a replacement for the tool you trust. It is the thing that makes sure a platform outage is an inconvenience, not a full stop.

Build the redundancy before you need it. Not after.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

Action Steps

1. Identify your current primary AI tool. Write down why you use it and what you use it for.

2. Pick one backup tool today. Spend 30 minutes testing it with a real task from your workflow.

3. Export your key prompts, brand notes, and writing guidelines into a document you own. Google Docs, Notion, a plain text file. Format does not matter. Ownership does.

4. Run a drill once a month. One session using only your backup tool. You want familiarity before the emergency, not during it.

5. Treat your creative context as an asset. Platform access is temporary. Your thinking is not.

The tools will keep changing. Some will go down. Some will shut down entirely.

The creators who keep moving through all of it are not the ones using the best tool. They are the ones who built a workflow that does not depend on any single one.

Create with clarity and control. Frederick Tadeo · STIRMIND

Credit Bio

Visuals created by Frederick Tadeo · STIRMIND · GenAI Creative Lead, Dubai. GenAI workflows and breakdowns from concept to final output. Recognized at Cannes Lions, Dubai Lynx, The FWA, MENA Cristal, and the Effies. STIRMIND since 2004. stirmind.com

One prompt. Any name. The output looks like it was built for a stage backdrop, a vinyl sleeve, or the back of a leather jacket.

The effect works because the prompt is precise. Heavy-metal letterforms. Hollow outlined text. Fire only on the stroke, not filling the inside. A dark reflective floor as the only surface. Nothing else in the frame. The result is cinematic, minimal, and immediately recognizable.

The tool stack is Claude, Google Flow, and Nano Banana Pro.

The Prompt

Drop in any name where it says [ZARA]. Everything else stays the same.

Create a cinematic fire typography image on a super-black stage, where the user input text [ZARA] is written in a thick vintage heavy-metal title style, with the same kind of bold, jagged, dramatic energy as classic 1970s doom-rock poster lettering. The typography should feel dark, sharp, angular, slightly uneven, hand-drawn, and ominous, not playful, rounded, bubbly, ribbon-like, or soft. The letters should be thick and weighty, with pointed cuts, sharp inner angles, rough vintage poster character, and a slightly sinister handmade feel. The text should sit in a subtle arc with perspective, large and centered, fully legible, dominating the frame. The fire must appear only on the outer stroke / contour of each letter, not filling the inside. The interior of the letters should remain empty black negative space, so the text reads like hollow outlined heavy-metal typography traced entirely in fire. The flame should follow the full contour of each letter with realistic thickness, glowing edges, ember detail, organic flicker, uneven flame behavior, and bright molten intensity. Keep the flame color consistent: hot yellow-white core, orange glow, subtle amber edges, realistic and photoreal. The environment is a dark empty black-box stage with no visible background elements. The floor is dark and slightly reflective, catching a soft warm reflection of the burning outlined text beneath it. The fire is the only light source, creating strong contrast, deep blacks, and a premium cinematic mood. Do not add any other elements. No people, no props, no decorations, no symbols, no logos, no watermark. The final result should feel like a dark iconic fire-forged title card, bold, heavy, sharp, menacing, photoreal, minimal, and dramatic, with thick vintage heavy-metal outlined typography made only of fire stroke.

The Output

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

Where this effect actually gets used

This is not just a visual experiment. Fire typography has real applications across branding, merch, music, and hospitality.

Album art and vinyl covers. A band name rendered in fire on a black stage reads like a classic record sleeve. It sets tone before a single note plays.

Merch and apparel. Embroidered on a cap or printed as a patch on a bag, fire lettering gives any name a raw, high-energy character. It transfers well to physical goods because the contrast is bold and the letterforms are thick.

Leather jacket back panels. A name burning across the back of a jacket is iconic. It has been part of rock culture for fifty years. This prompt recreates that energy from a single text input.

Tattoo concepts. Fire outlined letterforms translate naturally into tattoo design references. The hollow stroke, the weight of the letters, and the organic flicker give a tattoo artist something real to work from.

Restaurant and bar signage. A bistro name in fire lettering above an entrance creates immediate atmosphere. Warm, dramatic, and distinctive. It works for any venue that wants edge over polish.

Event titles and poster headers. Concert posters, fight nights, brand activations. Any event that needs a title with presence.

STIRMIND GenAI workflows and breakdowns direct to your inbox.
Subscribe free below.

Credit Bio

Visuals created by Frederick Tadeo · STIRMIND · GenAI Creative Lead, Dubai. GenAI workflows and breakdowns from concept to final output. Recognized at Cannes Lions, Dubai Lynx, The FWA, MENA Cristal, and the Effies. STIRMIND since 2004. stirmind.com

This is the full breakdown. Every decision explained.

The Concept

The idea came from a simple question. What if a campsite existed somewhere impossible? Not a mountain. Not a beach. Somewhere that could only exist through AI direction.

A floating island above the clouds answered that question.

The Direction Framework

Before a single prompt was written, the visual constants were locked. This is the part most people skip. It is also the part that determines whether the final output holds together or falls apart.

Four constants were defined and held across every frame:

Overcast sky as the atmospheric anchor throughout. Moss, rock, and pine tree texture consistent across all generations. Camera distance and angle held the same in every frame. One new element introduced per stage. Nothing else changed.

This is not prompting. This is direction. The difference is everything.

The Tools

Two tools. That is it.

ChatGPT with the Restoration Timelapse GPT was used to define the prompt requirements and map out the correct evolution from empty island to fully inhabited camp. It established what each frame needed to show before any generation began.

Nano Banana Pro via Google Flow handled all image generation. It proved to be a strong tool for this type of project. Its consistency across multiple generations kept the island, lighting, atmosphere, and scale coherent throughout the full sequence without needing to recompose between frames.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

The Four Stages

Stage 1: Empty floating island. Establishing the base. Nothing on it. Just moss, rock, and pine trees suspended in overcast sky.

Stage 2: Camp fully set up. Tent, G-Wagon, hammock, campfire. The island is inhabited.

Stage 3: Construction crew arriving. Active scene. The island is being worked.

Stage 4: Island clearing. Scene winding down. The camp is being packed.

Quality Control

Quality control was applied between every frame. Not at the end. Between every single stage.

The check was simple. Does this frame match the progression requirements set in the direction framework? Does the island look like the same island? Is the lighting consistent? Has the camera drifted?

If the answer to any of those was no, the frame was rejected and regenerated.

This is where most AI video projects fall apart. The generation is only as good as the quality control applied to it.

The Lesson

Consistency is the hardest thing to control in AI video production.

The tools can generate fast. They cannot hold a visual standard across multiple frames on their own. That responsibility sits entirely with the person directing the workflow.

Lock the constants before you start. Check every frame before you move to the next. Reject anything that breaks the system.

That is the workflow. That is what produced this.

The Prompt Framework

The frame requirements for this project were generated using the Restoration Timelapse GPT on ChatGPT. It is free to use and will give you the prompt structure for any timelapse sequence you want to build.

Link: chatgpt.com/g/g-698d7ad83100819188526cc7df2e0626-restoration-timelapse

If you want the exact prompts I used across all four stages, subscribe to STIRMIND and leave a comment below. I will share them directly.

STIRMIND™ by Frederick Tadeo. GenAI workflows and breakdowns from concept to final output. Create with clarity and control.

Connect on: LinkedIn | X | Instagram | Zora.co | Behance } Gumroad

Creative Use Disclaimer

All content under STIRMIND™ by Frederick Tadeo is for non-commercial, educational, and exploratory purposes, especially when referencing real-world brands, logos, music, or trademarks.

Third-party elements are used solely for artistic experimentation, storytelling research, and AI visual prototyping. No ownership, affiliation, or endorsement is claimed. These works are not for sale, not monetized, and not connected to the entities depicted.

Rights holders may contact Frederick Tadeo directly for review or removal. We respect your work and intellectual property.

The Open

Every brand starts somewhere. Most start with a brief, a budget, a category. This one started with three words and a question: what if the thing the world tells you to suppress is actually the most sacred thing you feel?

This is the documented build of FEAR IS GOD — a complete brand identity, campaign system, and asset library created in a single AI-assisted session. No agency. No production budget. One creative direction framework and the right tools.

What follows is the full process — from seed idea to final output. Every decision explained. Every asset shown. Replicable by anyone who understands that AI is not a replacement for creative thinking. It is an amplifier of it.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

The Seed Idea

Three words. That is where this started.

“FEAR IS GOD”

Not “Fear of God.” That construction positions God as the object — something external to fear toward. The word OF creates distance.

IS collapses that distance entirely. Fear and God become the same thing. The feeling itself is the divine experience. You do not need belief in a being. You need only to have been human.

The biblical root:

Proverbs 9:10 — “The fear of the Lord is the beginning of wisdom.” Not the end. The beginning. The oldest wisdom literature in Western civilization placed fear at the start of everything that mattered. The campaign recovers that truth.

The modern world built an entire culture around fearlessness. This brand says: fearlessness is arrogance. Fear is clarity. The fear you have been taught to suppress is actually the most honest signal your body produces.

That is the idea. Three words. Ancient root. Current tension. Universal audience.

The Brand Foundation

Before a single visual is made, the brand must be defined. This is the positioning document that every subsequent creative decision traces back to.

Brand Idea Fear — felt fully, not suppressed — is the beginning of everything real.

Proposition The world told you to be fearless. The world was wrong.

Tension Fearlessness as cultural ideal vs. fear as the most honest human signal.

Audience Gen Z and millennial creatives, brand builders, and cultural thinkers aged 18–35. People who distrust polish and respond to truth.

Personality Ancient. Uncompromising. Human. Still.

Voice Declarative. No explanation. No apology. Short sentences. Every word load-bearing.

Brand Values:

— Reverence over performance

— Imperfection over polish

— Stillness over noise

— Truth over comfort

— Fear as wisdom, not weakness

The Manifesto:

The world told you to be fearless.

The world was wrong.

Fear is not weakness. Fear is recognition.

The honest response to something greater than yourself.

FEAR IS GOD.

The Visual DNA

Every visual decision in this campaign traces back to the brand idea. Nothing is decorative. Everything argues for the truth.

Color Palette:

Charcoal #1C1C1C The darkness the figure stands inside. Not black — dimensional, present.

Silver-White #E8E6E0 The god-ray. The light that chooses the figure. Cool, not warm.

Ancient Gold #8A7340 The hairline at the horizon. The campaign’s single color signature. Present in every asset.

Aged White #D4D0C8 The type color. The white of things that have stood in weather.

Typography:

Primary Cormorant Garamond — Italic weight 300 for confessions. Weight 700 for declarations.

Logic Heavy weight = certainty. Light weight = intimacy. The contrast between them is the campaign’s typographic tension.

Case Mixed. FEAR IS GOD in caps. Confessions in sentence case. The difference signals register.

The Figure:

Every human subject in this campaign is unidentifiable. Silhouetted, cropped, or shot in near-darkness. This is the campaign’s universality principle — the figure is every person, or it is no one. The audience casts themselves.

The Asset Library

13 asset categories. 40+ individual executions. Every asset traceable back to the brand idea.

01 · Campaign Hero Visuals

The film storyboard rendered as still images. The desert, the storm, the god-ray, the figure. The visual world of the campaign established before any copy is applied.

02 · Key Visuals — The Face Series

Close-up portraits. Imperfect skin. Soulless gaze. The zone-out expression of a person who has encountered something they cannot look away from. Desaturated. Film grain. No retouching of blemishes.

03 · Kinetic Eye Series

The campaign’s strongest executions. Eyes cropped to fill the frame. Text running through the nose bridge — the border between light and shadow. Three declarations.

04 · Manifesto Slide

The brand belief system as a designed asset. Deep charcoal. Display type. Ancient gold hairline. Five declarative lines. The document a brand builder keeps.

05 · Scripture Quote

Proverbs 9:10 reframed through the campaign lens. The brand rooted in three thousand years of wisdom literature. The most intellectually authoritative asset in the system.

06 · Copy Series — Five Confessions

Five fears. Five cards. Each one a confession that names what the audience carries privately. The copy does what the image cannot — it names the specific fear. The image does what the copy cannot — it reframes that fear as sacred.

07 · OOH / Environmental

Billboard. Building wrap. Museum installation. Each medium chosen because it amplifies the idea rather than just carrying it. FEAR IS GOD on a museum wall beside classical paintings is not advertising. It is a cultural position.

08 · Digital — Social Post

The same visual world translated to the feed. First-person perspective. The viewer is the figure. The god-ray descends toward camera. The scroll stops.

09 · Mobile Wallpaper

The most intimate screen a person owns. Unlocked eighty times a day. The god-ray entering from the top center threads through the clock position. Every time they check the time, they are inside the campaign.

10 · Merchandise

FEAR IS GOD is not a slogan on a product. It is a belief carried on the body and in the hands. Every object in this collection exists because someone wanted to carry the truth into a room without saying a word.

The merchandise line covers every proximity to the body — worn, held, carried, pocketed, displayed.

The objects:

Journal — Debossed FEAR IS GOD on black cloth cover. The private object. Where the fear goes when there is nowhere else to put it. Blank pages inside.

Pen — FEAR IS GOD engraved on matte black. The tool that fills the journal. The smallest object in the collection. The most daily.

Cap — Front panel. Small type. Center placement. FEAR IS GOD worn at eye level. The most public object in the collection.

Tote — Full display type. The largest type scale in the merchandise line. The bag that travels through the city carrying the campaign’s most commanding typographic treatment.

Hoodie — Center chest. Large. The garment that wraps the body. FEAR IS GOD as the first thing visible when the hoodie comes off.

Phone Case — The most intimate screen. FEAR IS GOD on the object unlocked eighty times a day.

Art Print / Poster — Full display type on black. The campaign on a wall. The room carries the truth.

Brand Card — The smallest object. Business card scale. FEAR IS GOD + DESIGNER CAPSULE STORE. The thing left behind.

11 · Silver Ring

FEAR IS GOD debossed into oxidized sterling silver. Worn on the hand. Felt on the body. The campaign as a daily physical reminder. Every time the fingers close, the truth is there.

13 · Lifestyle

The brand worn in the world. The figure on the rooftop. The Gen Z crew in uniform black. The still point inside a moving environment. The campaign as a culture people belong to.

How AI Built This

AI did not create this brand. Creative direction created this brand. AI executed it at a speed and scale that no traditional production pipeline could match in a single session.

The distinction matters. Every prompt in this session was built on a framework — the Campaign Ad DNA system — that defined the brief before the tool was opened. The brand idea, the tension, the visual DNA, the light language, the casting principle — all established through strategic thinking first. The AI received clear, specific, load-bearing instructions.

What AI did:

— Translated creative briefs into production-ready image prompts

— Iterated on visual direction across dozens of variations in minutes

— Maintained visual consistency across assets

— Generated the Photoshop retouching script

— Held the brand DNA across every single execution

What human direction did:

— Defined the brand idea from a single human insight

— Made every creative decision that mattered

— Rejected work that was close but not right

— Pushed each execution until it earned its place

— Maintained the standard: an ad is not a message, it is a feeling with a brand attached to it

The tools used:

— Claude (Anthropic) — creative direction, copywriting, strategy, design system, code

— AI image generation — campaign visuals, key visuals, product mockups, lifestyle

— Photoshop — retouching, color grading, final asset production

The session ran as a continuous creative conversation. Every asset built on the last. Every decision documented. The result is a brand system that could be handed to a production team tomorrow and executed at any scale.

The Close

This is not a theoretical exercise. Everything in this case study was built in a single session using AI as a creative partner and a framework built from twenty years of studying how great advertising is made.

The brand exists. The assets exist. The system is documented. Any brand builder, creative director, or founder reading this can apply the same process to any idea — provided the idea is true enough to build on.

The only requirement is the insight. The one human truth that makes an audience say: yes, that is exactly how I feel.

Find that truth. Brief the AI with precision. Hold the standard of the idea against every execution decision. Remove everything that does not serve the idea. What remains is the brand.

FEAR IS GOD.

Frederick Tadeo · STIRMIND™ · stirmind.com · Since 2004

Create with clarity and control.

FEAR IS GOD is a brand identity and campaign concept created by Frederick Tadeo, founder of STIRMIND™. Built as a creative AI exercise demonstrating how a complete brand system — positioning, visual identity, campaign concept, and production-ready assets — can be developed from a single human truth using AI as a creative direction tool. All strategic thinking, creative direction, art direction, copywriting, and brand architecture by Frederick Tadeo · STIRMIND™ · stirmind.com · Since 2004.

#FearIsGod #STIRMIND #FrederickTadeo #AIBranding #BrandIdentity #CreativeDirection #ArtDirection #AICreativeDirection #BrandBuilding #CampaignCreative #VisualIdentity #BrandStrategy #BrandCaseStudy #AITools #BrandingWithAI #GenZBranding #BrandManifesto #Behance

This started with a stool.

Suspended mid-collapse. Wood fragments frozen in the air. Fabric cutting through the frame like it has its own momentum. Clean studio light. Editorial calm. Premium CGI realism.

I built this using AI creative direction and GenAI tools. One prompt. One clear brief. A structured workflow from concept to final output.

This is the full breakdown.

The Problem With How Most People Prompt

Most people open an AI tool and start typing.

No concept. No references. No visual brief. No quality standard. They generate 40 variations and pick the least bad one.

That is not a workflow. That is guessing with extra steps.

Chaotic input produces chaotic output. Structured direction produces controlled, repeatable results. That is the foundation of every STIRMIND workflow.

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

The Concept

The goal was a surreal studio shot. A product frozen in a moment of elegant collapse.

Not destruction. Suspension. The feeling of something coming apart at the exact right moment. Wood fragments mid-air. Fabric flowing through the composition with sculptural intention. A calm luxury editorial energy that sits between fashion photography and high-end product CGI.

Reference world: premium furniture campaigns, editorial still life, cinematic product photography.

Feeling target: cinematic stillness. Controlled chaos. Expensive calm.

The Direction

Before I wrote a single word of the prompt, I built the creative brief.

Visual references: suspended product photography, exploded view furniture campaigns, editorial fabric studies.

Style target: clean plain studio background, soft diffused lighting, subtle highlights on wood and fabric edges, minimal shadows.

Composition intent: centered, vertical, floating. The object occupies the full frame without touching the edges.

Mood: calm luxury. Not dramatic. Not loud. Refined tension.

Lighting brief: soft diffused key light, subtle highlights on broken edges, no harsh shadows, premium CGI realism.

This brief existed before the prompt was written. That is how I work. Direction first. Tools second. Every time.

The Master Prompt

Here is the full prompt, broken down by section so you can see the logic behind every decision:

Subject and action: A disruptive surreal studio shot of a stool disassembled and suspended in midair, with fractured wood pieces and structural fragments frozen in an elegant moment of collapse.

Fabric and flow: Smooth soft cloth flows through and around the object in graceful sculptural curves.

Background and lighting: Clean plain studio background. Soft diffused lighting. Subtle highlights on the wood, fabric, and broken edges. Minimal shadows.

Composition and feel: Refined floating composition. Calm luxury fashion-meets-product aesthetic. High-end editorial look. Cinematic stillness. Premium CGI realism. Centered composition.

Format: 9:16 vertical.

Each section is a creative decision, not a description. That is the difference between prompting and directing.

A disruptive surreal studio shot of broken explode [SUBJECT] disassembled in studio ikea suspended in midair, with fractured wood pieces and structural fragments frozen in an elegant moment of collapse, while smooth soft cloth flows through and around the object in graceful sculptural curves. Clean plain studio background, soft diffused lighting, subtle highlights on the wood, fabric, and broken edges, minimal shadows, refined floating composition, calm luxury fashion-meets-product aesthetic, high-end editorial look, cinematic stillness, premium CGI realism, centered composition.

Output and Quality Control

I reviewed every output against the creative brief. Not personal preference. The brief.

Questions I asked at every review:

Does the composition feel suspended or does it feel like it is falling? It needs to feel frozen.

Is the fabric flowing with intention or does it look accidental? Every curve needs to feel deliberate.

Is the lighting premium or is it flat? Soft diffused key light. Subtle edge highlights. No harsh shadows.

Is the background truly clean or is there noise? Studio white. No texture. No distraction.

Anything that did not meet the brief was rejected. Not because it looked bad. Because it did not match the direction.

That is creative quality control. That is the STIRMIND standard.

The Lesson

One thing to apply to your next GenAI creative workflow today:

Write the brief before you write the prompt.

Visual references. Style target. Mood. Lighting intent. Composition logic. Get that clear first. Then write the prompt as a series of creative decisions, not a list of descriptions.

Direction produces control. Control produces calm output. Calm output is repeatable.

That is the STIRMIND workflow. Concept to final output.

Get the Next Breakdown

If this was useful, subscribe to STIRMIND for weekly AI creative workflow breakdowns, prompt direction guides, and GenAI production systems.

I publish every breakdown at stirmind.com. AI creative workflows that replace chaos with clarity, control, and calm output. Since 2004.

Subscribe below and get the next one in your inbox.

#AICreativeDirection #GenAIWorkflows #PromptDirection #CreativeQualityControl #ConceptToFinalOutput #AICreativeWorkflowsDubai #FrederickTadeo #STIRMIND #GenerativeAI #CalmCreativeOutput #AICreativeSystems #GenAICreativeProduction

If sample 1 was the city version, this one is the cabin version. And this one matched the reference I had in my head when I started the whole experiment.

A wooden cabin at night. Two people at the table, deep in their notes. A lantern between them. A cat on the windowsill watching the moon. Fire crackling to the left. Total stillness except for the motion of the loop.

It is the Lofi Girl energy. Warm, focused, calm. Built entirely with AI tools from concept to final output.

This is what a GenAI workflow looks like when you lead it with direction and prompt structure rather than luck. The output is calm. The process is controlled. That is the STIRMIND way.

The workflow:

Image · Google Nano Banana Pro
Motion · Google Veo 3.1
Edit · After Effects
Music · Audiio

Subscribe for More AI Creative Workflows.

Get it straight to your inbox. AI creative workflows, prompt direction guides, and GenAI production systems.

Subscribe now

The prompt I used:

{
"directive": "Generate a high-fidelity 2D digital illustration in a flawless late-1980s cinematic anime aesthetic, optimized for a lofi video loop.",

"style": {
"primary": "Cinematic soft anime lofi illustration",
"quality": "Professional 80s OVA quality",
"rendering": "Crisp cel-shading, clean line-art, flat colors with soft gradients",
"texture": "Vintage anime film grain, hand-drawn aesthetic",
"resolution": "8k",
"motion": "Ultra-smooth loop, ultra-slow, subtle parallax feel",
"mood": "Soothing, intimate, nostalgic, off-grid, cinematic"
},

"camera": {
"type": "Static camera",
"lens": "35mm anamorphic",
"framing": "Medium-wide cinematic composition",
"depth_of_field": "Subtle bokeh on background elements",
"aspect_ratio": "2.39:1 cinematic widescreen",
"overlays": "None"
},

"subjects": [
{
"id": "subject_01",
"description": "A girl studying at a small fold-down wooden table, leaning forward with quiet focus, writing in her notebook, occasionally pausing to sip from a coffee cup",
"expression": "Calm, focused, absorbed",
"clothing": "Cozy oversized knit sweater, 90s relaxed style"
},
{
"id": "subject_02",
"description": "A boy sitting across the same fold-down table, reading an open book, elbow resting on the table, head slightly tilted in thought",
"expression": "Relaxed, contemplative, unhurried",
"clothing": "Vintage band tee, casual layered jacket"
}
],

"animals": [
{
"type": "Cat",
"position": "Camper windowsill",
"action": "Sitting, tail swaying lazily in a slow loop rhythm"
},
{
"type": "Dog",
"position": "Narrow floor below the table",
"action": "Curled up, twitching softly in sleep"
}
],

"scene": {
"setting": "Interior of a cozy vintage camper van, wooden paneled walls, worn and lived-in",
"table": "Small fold-down wooden table with open notebooks, pencils, scattered notes, and a half-empty coffee mug leaving a ring",
"shelf": "A few paperbacks wedged into a small wall shelf",
"cassette_player": "Portable cassette player tucked in the corner",
"wall_details": "Rock and roll sticker pressed onto the camper wall, a small hanging dreamcatcher swaying gently",
"door": "Camper door slightly ajar, dim campfire glow flickering outside"
},

"lighting": {
"primary": "Warm golden battery lamp light pooling over the table and subjects",
"secondary": "Cool moonlight filtering through the camper window",
"accent": "Distant dim campfire glow through the ajar door",
"style": "Soft volumetric light beams, heavy bloom, subtle rim lighting on both subjects",
"atmosphere": "Warm interior glow contrasting cool night exterior"
},

"background": {
"window_view": "Deep starry night sky, pine tree silhouettes barely swaying, full moon casting soft cool light",
"exterior_hint": "Campfire flicker visible through the door gap"
},

"loopable_elements": [
"Cat tail swaying on the windowsill",
"Coffee cup steam curling upward slowly",
"Dreamcatcher swaying with micro-movement",
"Pine trees shifting slightly in the breeze outside",
"Campfire glow flickering through the door",
"Dog breathing slowly, subtle chest rise",
"Dust motes drifting in the lamp light"
],

"color_palette": [
"Warm Amber",
"Deep Charcoal",
"Moonlight Silver",
"Pine Green",
"Ember Orange",
"Soft Cream",
"Cool Indigo"
],

"negative_prompt": "3D render, photorealistic, CGI, modern digital art style, sharp digital edges, aggressive lighting, pixel art, messy sketches, nude, NSFW, dull colors, natural lighting, plain textures, modern furniture, fluorescent light"
}

Your production time will depend on your prompt and your idea. But once you have the master prompt, the AI creative workflow becomes repeatable. That is the whole point of building GenAI systems rather than one-off experiments.

One honest note on quality control.

There are a few QC issues in this output. AI creative workflows are not a 1-click, 1-perfect result. That is the reality of GenAI content production right now. Knowing where the gaps are is part of the workflow. Fixing them, or deciding when to move forward, is a skill that sits at the core of creative quality control.

If you can spot the QC issues in this one, drop them in the comments. Let’s see what you notice.

This is a STIRMIND breakdown by Frederick Tadeo. AI Creative Lead, Dubai. Building GenAI workflows and AI video breakdowns that go from concept to final output with clarity and control. AI creative workflows · GenAI content production · prompt direction · creative quality control · stirmind.com · Since 2004.

#AIWorkflow #GenAI #AIVideo #AICreative #STIRMIND #GoogleVeo #GoogleImagen #AfterEffects #GenAIWorkflows #AIVideoBreakdowns