AD LAB — ORIGINAL SCRIPTS

Fictional cases, with no real campaigns or claimed results.

Vertical 9:16 format. Label the piece as AI fiction.


Lume · narrated ad

Cylindrical magenta flashlight, black handle, and one front lens. Dark stone and blue surroundings. It does not depict real performance.

1. 0–2s — Product in low light
Image: Whole flashlight on dark stone; magenta outline visible, with room for text in the editor.
Motion: Static camera; light mist.
Check: You can recognize the product by its outline.

2. 2–4s — Beam appears
Image: Same flashlight lit; soft beam in blue mist.
Motion: Mist moves without changing the lens.
Check: One lens, one handle, shape preserved.

3. 4–6s — Full product view
Image: Same flashlight fully visible on the stone.
Motion: Small, slow push-in.
Check: No part disappears.

4. 6–8s — Body texture
Image: Close-up of the magenta body, same proportions and material.
Motion: Light changes softly; product stays still.
Check: Texture does not turn into letters or a brand.

5. 8–10s — Handle
Image: Close-up of the black handle attached to the body, with natural support.
Motion: Steady camera.
Check: The handle stays attached and does not multiply.

6. 10–12s — Scale in the setting
Image: Whole product on the same stone, with more blue space around it.
Motion: Mist passes in the background.
Check: Scale stays consistent throughout the clip.

7. 12–14s — Light and material
Image: Lens and body visible, with controlled shine.
Motion: Small change in light.
Check: Shine does not erase the product’s shape.

8. 14–16s — Return to the full view
Image: Same main composition, whole product visible.
Motion: Movement slows to a stop.
Check: Handle, lens, and body stay consistent.

9. 16–18s — Ending
Image: Whole product, with a clean background for a line added in the editor.
Motion: Static camera.
Check: The name and label fit without covering the object.

Sound: When everything looks the same, a pause changes how you see it. Lume: imagine another path. Record at a natural pace, listen, and spread the line across 18 seconds. Do not cut words to make it fit.

Ready: Nine scenes, 18 seconds, consistent product, and clear voice. Label it: fictional AI concept.


Vento · music sequence

Adult in coral clothing, green glass corridor, same light and direction. Four edited clips; not one real take.

1. 0–5s — Start
Image: Person shown in full at the start of the corridor, arms relaxed.
Motion: Starts walking slowly; camera follows.
Check: Body, clothes, and pillars stay steady.

2. 5–10s — Move forward
Image: Use the previous clean final frame; same person a few steps ahead.
Motion: Continues in the same direction.
Check: Position and scale fit the transition.

3. 10–15s — Strip of light
Image: Same person crosses a strip of sunlight on the floor.
Motion: Small walking movement, no turn.
Check: Light belongs to the space; it does not change the outfit.

4. 15–20s — Stop
Image: Same corridor; person finishes the step.
Motion: Movement slows to a stop.
Check: Feet do not slide, and the music ends deliberately.

Sound: Use the kit’s instrumental throughout the 20 seconds. Listen before choosing the transitions. These intervals are a study structure, not an analysis of the track’s beats.

Ready: Four segments, three checked transitions, continuous sound, and credits preserved. Label it: fictional AI study.


Senda · product in five shots

Translucent amber frames, two dark lenses, a bridge, and two temples. Cream paper. No protection or health claims.

1. 0–3s — Full view
Image: Whole glasses on cream paper, with room around the temples.
Motion: Slow push-in.
Check: The full product is recognizable.

2. 3–6s — Paper
Image: Same frames, fold in the paper in the foreground.
Motion: Fold moves gently.
Check: Paper does not transform the product.

3. 6–9s — Material
Image: Close-up of the amber frames and hinge.
Motion: Camera nearly still.
Check: Material and joints stay consistent.

4. 9–12s — Angle
Image: Same reference, slightly different angle.
Motion: Minimal movement, no full turn.
Check: Two temples do not pass through the lenses.

5. 12–15s — Ending
Image: Whole product against a clean background.
Motion: Static camera.
Check: The Senda line “notice the details” is readable.

Sound: Optional licensed music. The piece works without a voice. If you use the kit excerpt, trim it deliberately to 15 seconds and keep the credit.

Ready: Five scenes, 15 seconds, a coherent build, and a readable line. Label it: fictional AI concept.


Pulso · shortened sports ad

Coral sneaker, cream sole, coral laces; navy outfit and blue track. No performance or health claims.

1. 0–4s — Product
Image: Side view of the sneaker resting on the blue track.
Motion: Static camera or a slight push-in.
Check: The sneaker meets the track, and the sole looks consistent.

2. 4–8s — Getting ready
Image: Adult standing still, arms relaxed, wearing navy and the reference sneaker.
Motion: Small shift in weight.
Check: Two legs, two feet, and a balanced stance.

3. 8–12s — One step
Image: Same person and direction, with room to move one foot forward.
Motion: One short step, then a foot planted on the track.
Check: The foot does not pass through the ground; the body does not multiply.

Sound: Three light taps you record can mark the edit. They set a rhythm; they are not a heartbeat recording. Or make the first cut in silence.

Ready: Three scenes, 12 seconds, with footing and anatomy checked. Label it: fictional AI concept.


Arquivo do Vento · first person

Four references: terracotta room, brass astronomical instrument, gray gloves, and circular ceiling light. It does not depict a real place.

1. 0–4s — Entering
Image: View from the entrance toward the table; light from the circular opening.
Motion: Slow forward movement.
Check: The table and ceiling match the map.

2. 4–8s — Moving closer
Image: Same room, table closer; the object stays at the center of the action.
Motion: A slight move forward without crossing the table.
Check: The approach matches the scale.

3. 8–12s — Looking closely
Image: Brass instrument nearby; gloves at the bottom of the frame.
Motion: Steady hands; slight shift in gaze.
Check: Hands and object stay undistorted.

4. 12–16s — Ending
Image: Same viewpoint; instrument and room remain recognizable.
Motion: Movement ends calmly.
Check: Doors, ceiling, and table stay in place.

Sound: Use subtle ambient sound you recorded or are allowed to use. You can start in silence to check the space. Music and a scare are optional.

Ready: Four scenes, 16 seconds, three spatial transitions checked. Label it: fictional AI world.


Dobra · seven-scene demonstration

Jo: brown skin, short black hair, cream top. Rectangular olive canvas bag with two cream handles. The character and product are fictional.

1. 0–3s — Introduction
Image: Jo shown from the waist up, holding the whole bag.
Motion: Brief speech, minimal gestures.
Check: Check the shape. Match the mouth to the voice, or label it as narration.

2. 3–6s — Full view
Image: Whole bag, same reference.
Motion: Static camera.
Check: Show the whole bag. Keep its rectangular shape.

3. 6–9s — Handles
Image: Close-up of the cream handles attached to the bag.
Motion: Hand holds a handle without twisting it.
Check: Inspect the handles. Hand and fabric stay separate.

4. 9–12s — Fabric
Image: Close-up of the olive canvas.
Motion: Slight push-in.
Check: Inspect the fabric. Its texture does not turn into another material.

5. 12–15s — Opening
Image: Open bag, with no capacity claim or objects appearing from nowhere.
Motion: Minimal movement.
Check: Inspect the opening. Edges and handles stay consistent.

6. 15–18s — Full scene
Image: Jo and the whole bag in a waist-up frame.
Motion: Simple pose.
Check: Compare the full scene. Face, outfit, and product stay consistent.

7. 18–21s — Ending
Image: Product visible, with room for its label.
Motion: Static camera.
Check: Introduce the Dobra concept. Keep the full line readable.

Sound: Use the seven short suggested lines, with pauses. Your own or a synthetic voice is fine. Narrating over details reduces the need to sync speech. Do not invent a shopping experience.

Ready: Seven scenes, 21 seconds, with speech, character, and bag consistent. Visible label: Fictional AI demonstration.

CASE:
SCENE:
FILE AND ATTEMPT:
REFERENCE USED:
PROMPT:
STATUS: approved / revise / pending
TIME SPENT ON ISSUE:
WHAT I OBSERVED:
CHANGE TO TEST:
ACTIVE WORK: measured / estimated / not measured
PROCESSING WAIT:
KNOWN ATTEMPTS AND USAGE:
LIMIT AND ALTERNATIVE:

CHOSEN CASE:
AUDIENCE:
MESSAGE:
FORMAT:
DURATION AND SCENES:
EXPECTED FILES:
CRITERIA: sequence / identity / sound / readability / labeling
PREVIOUS VERSION:
REVISED VERSION:
CHANGE AND OBSERVED EFFECT:
DECISION:
INDEX OF THE SIX STUDIES:
Lume — 18s — file and status:
Vento — 20s — file and status:
Senda — 15s — file and status:
Pulso — 12s — file and status:
Arquivo do Vento — 16s — file and status:
Dobra — 21s — file and status:
FINAL FINISH — file and status:
SOURCES AND CREDITS:
REAL PENDING ITEMS:


CREDIT
Carefree — Kevin MacLeod (incompetech.com). CC BY 4.0 license: https://creativecommons.org/licenses/by/4.0/. First 20-second excerpt, converted to MP3. Source: https://incompetech.com/music/royalty-free/index.html?gt=&isrc=USUAN1400037. Obtained by INEMAVOX on 25/09/2026. If you change the volume or duration, or add fades, include that in the credit.
