🦎 How PJ Ace Created Viral AI Videos & Built a 7‐Figure AI Ad Agency

Hey there, Creator!

What if I told you there is a replicable system to create viral AI video ads that rack up tens to hundreds of millions of views — and that you could use that same system to build a six- or seven-figure business?

That’s exactly what PJ Ace discloses in his recent interview (I will link the YouTube video at the end.

He peeled back the curtain on the full process his agency uses to scale to six and seven figures in a matter of months.

And now I’m giving it to you, in sausage-level detail.

 In this edition of The Creator Report, you’ll get:

• The complete AI video creation “stack” PJ uses (and how you can too)

• His secret strategies for writing scripts that captivate and convert

• The “image-first” trick that separates amateurs from pros

• Advanced hacks — like reference-image injection — that make VO3 clips realistic

• How Sora might disrupt all of this (and how to stay ahead)

If your goal is to build faceless creative ad campaigns, client work, or viral content — this is your playbook.

Let's. Crawl. In. 🦎

1. The End-to-End AI Video Creation Process: From Script to Viral Video

PJ’s agency didn’t creep to success. They rocketed — using a repeatable, lean pipeline:

• Script / Big Idea

• ChatGPT → Shot List

• Figma Board / Visual Layout

• Image Generation (Rev / Nano Banana / Freepik)

• VO3 Animation (Frames to Video)

• Final Edit + Music / Text Overlays

What made it scalable? Starting with comedy. When viewers laugh, they engage — and they’re less likely to revolt at “AI doing the creative work.”

Using jokes, historical IP, absurd juxtapositions, and meme-friendly content neutralizes pushback and hooks viewers in.

Make people laugh first, then sell to them. That’s the rule.

2. The Script Is Everything — Use Recognizable IP + Juxtapositions + Internet-Native Hooks

Your job isn’t to impress people with fancy tools — it’s to make them care in 3 seconds or less. That means:

• Use familiar settings: Pompeii, Titanic, Marie Antoinette, Greek myths

• Twist the scene with modern absurdity: “Beanie Baby vaults,” “meme coins funding Roman aqueducts”

• End with a native punchline: NFTs, slang, or ironic meta-humor

Think like a meme historian. Your script is the bridge between intrigue and virality.

PJ often uses ChatGPT to generate 100+ weird script combinations — most are garbage, but one will be gold.

3. Don’t Go Straight Text-to-Video — Use the Image-First Approach

This is where most AI video creators typically fall short. Going straight from text → video feels janky. Instead, PJ’s agency uses an image-first flow:

1. Take script → ChatGPT → Shot-by-shot scene list

2. Convert shots into detailed visual prompts

3. Feed prompts into the Rev app (get 3 variations per prompt)

4. Iterate conversationally: “Add more smoke,” “Zoom in on the hands,” “Shift lighting to sunset”

5. Export the best version of each scene

That means every scene is strong, consistent, and on-brand before you animate a single frame.

Even if you’ve never touched a camera, this gives you pro-level results.

4. VO3 Animation: Making Static Art Come Alive

Once you’ve got polished scenes, you animate them using VO3 (Google Flow):

• Upload image(s) as the starting frame

• Add your dialogue + motion prompt

• VO3 handles facial expression, lip sync, ambiance, and movement

• Combine multiple frames for scene cuts, transitions, zooms, pans

This beats generic text-to-video. Why? Because you’ve already pre-directed the visuals.

VO3 is your cinematographer. You’re just giving it the storyboard and camera instructions.

5. Advanced Hacks That Make Your Video Look Real

Want your content to feel cinema-grade real? Use these PJ-approved pro tricks:

• Reference Images: Hide a Beanie Baby or meme inside your image. VO3 will understand what to reveal later.

• Picture-in-Picture: Prompt VO3 to “zoom into the drawer” or “pan from the window to the desk.”

• FAST Mode in Google Flow = 80% of quality, 20% of cost

• Skip VO3’s built-in music — use Epidemic Sound or Storyblocks for better cohesion

• Enhance with texture tools (like Remini, DaVinci passes) to reduce AI uncanny valley

These hacks are how PJ’s team makes ads that feel human-shot, even if they’re entirely AI-built.

6. Editing & Tools: Assemble It Like a Pro (Even as a Beginner)

Want your content to feel cinema-grade real? Use these PJ-approved pro tricks:

• Reference Images: Hide a Beanie Baby or meme inside your image. VO3 will understand what to reveal later.

• Picture-in-Picture: Prompt VO3 to “zoom into the drawer” or “pan from the window to the desk.”

• FAST Mode in Google Flow = 80% of quality, 20% of cost

• Skip VO3’s built-in music — use Epidemic Sound or Storyblocks for better cohesion

• Enhance with texture tools (like Remini, DaVinci passes) to reduce AI uncanny valley

These hacks are how PJ’s team makes ads that feel human-shot, even if they’re entirely AI-built.

7. The Sora Disruption: Why This Workflow Might Change Soon

PJ admits it: Sora might blow this whole process up.

Sora will soon be able to:

• Auto-generate scripts

• Create image + video + audio in one pass

• Extend clips beyond 10 seconds

• Add emotion, physics, and realistic camera movement

That said… ideas are still king.

If you can come up with one scroll-stopping idea per week, you’ll always beat lazy creators, no matter how fast Sora gets.

BONUS: The Ultimate VO3 / Veo Prompt Template

Use this prompt to create the best videos:

A [MOOD] [CAMERA SHOT TYPE] opens on [PRIMARY CHARACTER: age, look, outfit, emotion], standing in [SETTING: be detailed – describe textures, objects, mood].

The camera [CAMERA MOTION – dolly in, slow pan left, static, tracking shot] as the character [ACTION – be specific: glances nervously, throws something, scrolls phone]. Their [FACIAL EXPRESSION] reflects [INNER STATE – worry, excitement, confidence].

Lighting is [LIGHTING STYLE – golden hour glow, harsh office fluorescents, candlelit shadows], casting [SHADOW DETAIL – soft cheek shadows, long floor silhouettes]. The environment has [ATMOSPHERIC CONDITIONS – light fog, dust motes, steam, rain].

Color palette features [3–5 COLORS – e.g., cold grays, glowing amber, desaturated teal], reinforcing a [PSYCHOLOGICAL EFFECT – tense, calm, energetic, nostalgic].

Background elements include [RELEVANT OBJECTS – sales dashboard, neon billboard, messy desk, angry crowd], creating a rich, immersive context.

The camera [CAMERA ADJUSTMENT – zooms in, tilts up, holds on close-up] as the character delivers the line:  
"[MAIN DIALOGUE – max 10 words, emotionally punchy]"

Text overlay appears on screen in [FONT STYLE – bold sans-serif, glitchy typewriter, hand-drawn serif]:  
"[HEADLINE – value prop or surprising twist]"

Sound design includes [IMPLIED AUDIO CUES – crowd murmurs, digital pings, thunder outside, calm ambient hum] to build immersion and pacing.

Final shot reveals [PRODUCT/CTA MOMENT – Stripe payment notification, AI building a website, rocket taking off, customer smiling]. Text overlay (centered):  
"[CALL TO ACTION – Tap to start, Build with AI, Claim yours now]"

This isn’t just a prompt — it’s a blueprint for realism.

Use it for every animated scene, and your VO3 results will appear as if they were recorded on a real film set.

Need some ideas for styling your next video? Here are some examples:

style: "epic, dramatic, gritty, dreamlike, [choose your mood/style]"

shot_type: "wide shot, medium shot, close-up, extreme long shot"

subject: "a [character or object] doing [specific, vivid action]"

setting: "a [location] filled with [textures, key objects, ambiance]"

lighting: "golden hour, harsh fluorescent, candlelit, moonlit"

color_palette: ["dominant color 1", "accent color 2", "highlight color 3"]

atmosphere: "fog, smoke, rain, haze, dust motes" camera_motion: "static, slow pan left, dolly in, zoom out, handheld shake"

composition: "rule of thirds, leading lines, silhouette, shallow depth"

textures: "rain-slick brick, rusted steel, velvet drapes, cracked stone"

implied_sound: "distant thunder, echoing footsteps, eerie silence"

notes: "maintain photorealism, cinematic continuity, dramatic tone"

How to Use the Cinematic Prompt for VO3 or Veo

Step 1: Set the Scene with Vivid Details
Begin by clearly defining the emotional tone, character, and setting. Think cinematically. Ask yourself:

• What mood should the scene evoke? (e.g., gritty, dramatic, calm)

• Who is the primary character? (age, outfit, expression, role)

• Where is the action taking place? (environment details, lighting, textures)

This sets up the opening visual for maximum viewer retention.

Step 2: Direct the Camera and Character Behavior
Now guide the “camera” and actor behavior:

• Define the camera motion (e.g., dolly in, slow pan, static)

• Describe exactly what the character is doing

• Include facial expressions and internal emotions

These cues will help VO3/Veo generate realistic movements and reactions.

Step 3: Control Lighting, Colors, and Atmosphere
Here’s where you add richness and tone:

• Choose a lighting style (e.g., golden hour, moody neon)

• Add atmospheric details (fog, dust, steam)

• Pick a color palette to evoke a specific psychological effect

• Describe background elements that enhance story depth (e.g., neon signs, crowded streets, glowing dashboards)

This makes the scene feel immersive and emotionally grounded.

Step 4: Layer in Dialogue, Text, and Sound
Wrap it all together with these cinematic elements:

• Include one powerful line of dialogue (10 words or fewer)

• Specify the on-screen text overlay and font style

• Describe the implied sound design (e.g., street noise, ambient hum)

• End with a clear visual call-to-action (e.g., “Start now,” “Join the rebellion,” “Claim yours today”)

This final layer turns a static prompt into a fully dynamic, branded story sequence.

What You Just Learned

• PJ Ace’s entire 7-figure video ad pipeline

• How to write viral scripts using humor + history + internet-native hooks

• Why image-first flows beat janky text-to-video

• Pro tricks to make AI animations look handcrafted

• Why Sora’s coming storm is a gift to creatives

• The cinematic prompt that levels up every shot