Story architecture
Define the central question, emotional arc and reason the audience should care.
Yousuf Khalid Albastaki
AI filmmaking
Story before technology.
Through AI filmmaking, visual storytelling and music, Yousuf explores how disciplined learning, imagination and service can connect sport, education and creative technology.
Writethe human idea
Builda coherent world
Directcamera and emotion
Finishsound, edit and meaning
AI expands the camera. Human judgement still shapes the story: what it means, what it protects and why it should exist.
Creative method
A pipeline built around intention.
World and identity
Lock characters, locations, materials, light and cultural details before generating scenes.
Shot direction
Design action, lens, movement, continuity and the exact emotional purpose of every shot.
Editorial rhythm
Shape the film through performance, pacing, sound, music and disciplined selection.
Human judgement
Use AI to execute creative decisions—not to replace taste, responsibility or authorship.
Selected visual work
Three stories. One human point of view.
A concise selection from Yousuf’s public AI-film work. Each card links to the original post; the complete archive remains on Instagram.
01
AI short film · Sci-fi world-building
SABR: The Still Point
Open original02
AI cinematic journey · Heritage, identity and the future
The Soul of Dubai
Open original03
AI visual story · Maritime heritage and future champions
No Wind. No Limits.
Open originalCreative note 36 · 13 October 2026
Light the story—not only the subject.
Google Cloud’s current video-generation prompt guide treats lighting, tone or mood, artistic style and ambience as distinct parts of a video’s visual language. Its lighting examples range from natural and artificial sources to high-key, low-key, backlit and side-lit treatments. For a filmmaker, the useful decision comes before the label: choose what the audience must notice or feel, then direct the source, direction, contrast and colour of the light to support that purpose. A generated result still requires review; lighting words guide the model, but they do not guarantee continuity or meaning.
- 01
Name the story purpose
Decide what the light should reveal, conceal or make the audience feel in this moment.
- 02
Choose the source
Identify the motivated natural or artificial source that belongs in the scene, such as window light, moonlight, firelight or practical fixtures.
- 03
Direct its character
Describe direction, softness, contrast and colour only where those choices change the story.
- 04
Protect continuity
Carry the important light direction, time of day and palette into adjacent shots so the sequence still belongs to one world.
- 05
Review the generated frame
Check faces, materials, shadows, exposure and cultural details; keep only lighting that serves the scene and remains coherent across the edit.
Lighting is not a decorative adjective. Give it a source, a direction and a story purpose—then review whether the generated shot belongs in the sequence.Google Cloud · Video generation prompt guide · Visual style and lighting
The archive
Keep exploring. Keep learning.
Creative note 33 · 10 October 2026Write the line the moment can carry.+
Creative note 33 · 10 October 2026
Write the line the moment can carry.
Google Cloud’s current Veo 3.1 prompting guide recommends putting specific dialogue in quotation marks and directing dialogue, sound effects and ambient noise as distinct parts of the soundstage. For a filmmaker, that is also a useful writing discipline: decide who speaks, what the moment truly needs them to say and how the surrounding sound supports the beat. A short, attributed line is easier to judge inside one shot than a paragraph that competes with the image. The generated result still requires review; a prompt is direction, not a guarantee.
- 01
Name the story purpose
Write what the audience should understand or feel because this line is spoken.
- 02
Attribute the speaker
Identify who speaks so the dialogue belongs to a visible character or a clearly defined narrator.
- 03
Quote only the essential words
Put the intended line in quotation marks and remove explanation the image or performance can carry.
- 04
Direct the performance
Add only the tone, pace or delivery detail that changes the meaning of the line.
- 05
Shape and review the soundstage
Specify the useful ambience or effect separately, then inspect speech, synchronisation, clarity and continuity before using the shot.
Dialogue is strongest when the speaker, exact words and purpose are clear. Write the line the moment can hold, direct the sound around it and review the generated result.Google Cloud · Ultimate prompting guide for Veo 3.1 · Directing the soundstage
Creative note 30 · 7 October 2026Build the prompt in layers—not adjectives.+
Creative note 30 · 7 October 2026
Build the prompt in layers—not adjectives.
Google Cloud’s official Veo prompting guide presents a structured prompt as a combination of cinematography, subject, action, context, and style and ambience. The value of that structure is creative clarity: each part makes a different decision visible, so the filmmaker can review what the camera should show, what must remain recognisable, what changes during the shot and how the world should feel. Structure does not guarantee a result; it makes the direction easier to test and refine.
- 01
Set the cinematography
Name the shot size, angle and only the camera movement the story beat needs.
- 02
Define the subject
Describe the stable identity, wardrobe, object or material details required for continuity.
- 03
Give one clear action
State what changes during the shot with a readable beginning and end.
- 04
Ground the context
Specify where and when the moment happens, including only the background conditions that affect the story.
- 05
Shape the ambience, then review
Choose the light, mood and visual style; generate the shot and compare the result with the intended beat and adjacent shots.
A prompt becomes directable when every layer has a job. Separate camera, subject, action, context and ambience—then review the result instead of treating the words as a guarantee.Google Cloud · Ultimate prompting guide for Veo 3.1
Creative note 27 · 30 September 2026Move the camera for a reason.+
Creative note 27 · 30 September 2026
Move the camera for a reason.
Google Cloud’s official video prompt guide distinguishes a static shot from movements such as pan, tilt, dolly, truck and zoom. Its examples connect movement to a change in story information: a tilt can reveal a letter, while a dolly out can emphasise isolation. In AI filmmaking, camera movement should therefore answer a question: what does the audience discover or feel because the frame moves?
- 01
Start with the story beat
Write what should change in the audience’s understanding or feeling during this shot.
- 02
Test the still frame
If the meaning works without movement, a static shot may be the clearer choice.
- 03
Choose one purposeful move
Select a movement that follows action, reveals information or changes emotional distance—not one that merely decorates the frame.
- 04
Describe direction and pace
Name where the camera begins, how it moves, its speed and the final composition.
- 05
Review the transition
Compare the result with the shots before and after it; reject movement that distracts or breaks continuity.
A moving camera should reveal, follow or reframe meaning. If it only decorates the shot, the story may be stronger without it.Google Cloud · Video generation prompt guide · Camera movements
Creative note 24 · 27 September 2026Choose what the audience knows.+
Creative note 24 · 27 September 2026
Choose what the audience knows.
Google Cloud explains that a camera angle defines the viewpoint of a generated shot and directly influences how the audience perceives its subject. Its examples also distinguish shot sizes by narrative function: a wide or establishing shot gives location and context, a medium shot balances a subject with the surrounding scene, and a close-up emphasises emotion or detail. In AI filmmaking, shot size is therefore a story decision—not a decorative label.
- 01
Ask the story question
Decide what the audience needs to understand or feel in this moment.
- 02
Establish when context matters
Use a wide shot when place, distance or the subject’s relationship to the environment carries the meaning.
- 03
Move closer for the right reason
Use a medium shot for action with context, or a close-up when emotion or one detail should lead.
- 04
Compare one variable
Test the same subject and action at different shot sizes, then keep the version that reveals the right information at the right time.
Shot size decides what the audience sees first and how close the moment feels. Choose it for the story, then review the result.Google Cloud · Video generation prompt guide · Camera angles
Creative note 21 · 23 September 2026Name the distraction—without telling the model ‘don’t.’+
Creative note 21 · 23 September 2026
Name the distraction—without telling the model ‘don’t.’
Google Cloud describes negative prompts as a way to specify elements that should not appear in a generated video. Its guidance recommends listing the unwanted elements directly instead of writing instructions with words such as ‘no’ or ‘don’t.’ The creative purpose is simple: keep the main prompt focused on the shot you want, and use the negative prompt to protect the frame from specific distractions.
- 01
Direct the shot
Use the main prompt for the subject, action, setting, camera and sound the story needs.
- 02
Spot the distraction
Identify only the visual elements that would weaken clarity, continuity or mood.
- 03
List, do not command
In the negative prompt, describe the unwanted elements directly rather than writing ‘no’ or ‘don’t.’
- 04
Review the result
Treat the prompt as guidance, not a guarantee; inspect every output before it enters the sequence.
Tell the main prompt what belongs in the shot. Tell the negative prompt which specific distractions do not. Then review the result.Google Cloud · Video generation prompt guide · Negative prompts
Creative note 18 · 19 September 2026Give the model a reference—not a memory test.+
Creative note 18 · 19 September 2026
Give the model a reference—not a memory test.
Google Cloud documents reference-to-video generation in supported Veo models: a filmmaker can provide up to three images of one person, character or product, and the model uses them to preserve the subject’s appearance. The reference does not replace direction. It anchors identity while the prompt defines the action, camera and story purpose.
- 01
Choose the subject
Use clear, approved images that show the same person, character or product without conflicting details.
- 02
Protect the constants
Write down what must remain stable: face, wardrobe, key materials, proportions and identifying details.
- 03
Direct the change
Prompt the action, camera movement and environment instead of repeatedly describing the subject.
- 04
Review continuity
Compare the output with the references and adjacent shots; reject identity drift or changes that weaken the story.
Use references to anchor identity. Use direction to create change. Keep only the result that belongs in the sequence.Google Cloud · Generate videos from references
Creative note 15 · 15 September 2026Define the first frame—and the last.+
Creative note 15 · 15 September 2026
Define the first frame—and the last.
Google Cloud documents video generation from specified first and last frames in supported models. For a filmmaker, those two anchors are useful before generation: the opening establishes the composition; the ending states what must visibly change. The movement between them should preserve identity, setting and screen direction while earning the transition.
- 01
Set the opening
Choose the composition, subject position and visual state the audience meets first.
- 02
Define the ending
Show one clear change that advances the story rather than only changing the picture.
- 03
Protect the constants
List what must remain stable: identity, wardrobe, setting, scale and screen direction.
- 04
Describe the bridge
Name one action or camera movement that can connect the two states.
- 05
Review the result
Reject continuity breaks and keep only motion that supports the intended story beat.
Design the change before asking AI to fill the frames between. Review the result; keep only motion that serves the story.Google Cloud · Generate videos using first and last frames
Creative note 12 · 11 September 2026Generate after the sequence makes sense.+
Creative note 12 · 11 September 2026
Generate after the sequence makes sense.
Google Workspace presents AI-assisted video work as beginning with a draft storyboard containing suggested scenes and a script. Used thoughtfully, that draft becomes a sequence test before time is spent generating polished clips: what changes from one beat to the next, why does each shot exist and how does it lead to what follows?
- 01
Define the purpose
Write one sentence describing what the audience should understand or feel.
- 02
Divide the beats
Turn the idea into three to five moments in which the story genuinely advances.
- 03
Assign the shot
Give each beat one essential frame or action.
- 04
Plan the transition
Name what connects the next shot to the last: movement, eyeline, sound or shape.
- 05
Review the sequence
Read the storyboard without effects; remove any shot that repeats or distracts from the idea.
Before generating another clip, make sure the sequence already works on paper. AI can accelerate production; story structure makes the film understandable.Google Workspace · AI-generated video with Google Vids
Creative note 09 · 7 September 2026A beautiful shot can break the story.+
Creative note 09 · 7 September 2026
A beautiful shot can break the story.
Adobe’s filmmaking guidance describes the 180-degree rule: keep camera positions on one side of an imaginary line between two characters to preserve their spatial relationship across cuts. Crossing that line can make them appear to swap sides and disorient the viewer. It is a storytelling convention, not an absolute law; breaking it should serve a deliberate purpose.
- 01
Map the scene
Before generating shots, sketch where each character stands and draw the line between them.
- 02
Choose a side
Plan the cameras on one side of that line so the conversation remains easy to follow.
- 03
Check the edit
Review adjacent shots for consistent screen positions and opposing eyelines—not just attractive individual frames.
- 04
Make exceptions intentional
If you cross the line, decide what the change should make the audience feel.
For AI filmmaking, apply the same continuity check to generated shots: a compelling frame must also belong in the sequence. Keep viewers oriented unless disorientation is your intention.Adobe · The 180-degree rule in filmmaking
Creative note 06 · 3 September 2026When the first frame is locked, prompt what moves.+
Creative note 06 · 3 September 2026
When the first frame is locked, prompt what moves.
Google’s official video-generation guidance treats a source image as the first frame of the film. Because the subject, scene and style are already visible, an image-to-video prompt works best when it directs motion instead of describing the picture again.
- 01
Choose the frame
Start with a sharp, well-composed image whose subject, light and visual language are already clear.
- 02
Name the action
Give the subject one deliberate movement with a readable beginning and end.
- 03
Direct the camera
Choose a purposeful camera move—such as a slow push or a controlled orbit—that supports the emotion.
- 04
Animate the world
Add only the environmental change the moment needs: wind, water, fabric, light or atmosphere.
- 05
Protect continuity
Review identity, screen direction, scale and setting before the clip enters the edit.
A strong image-to-video prompt does not repaint the frame. It directs the change inside it.Google Cloud · Best practices for generating videos
Creative note 03 · 31 August 2026Direct the soundscape—not only the picture.+
Creative note 03 · 31 August 2026
Direct the soundscape—not only the picture.
Sound is part of visual storytelling, not an afterthought. Google’s official Veo guidance recommends describing audio clearly and separately so dialogue, ambience and distinct effects can shape action, pacing and mood.
- 01
Voice
Define who speaks, the delivery and only the words the moment truly needs.
- 02
Ambience
Name the background sound that makes the location believable.
- 03
Action sound
Choose the one or two synchronised sounds that give movement weight.
- 04
Silence
Leave space when restraint carries more emotion than noise.
- 05
Continuity
Keep recurring voices, environments and sonic details consistent across the sequence.
A believable world is heard before it is explained.Google Cloud · Video generation prompt guide
Creative note 01 · 29 August 2026Prompt a shot—not merely an image.+
Creative note 01 · 29 August 2026
Prompt a shot—not merely an image.
The difference between an attractive frame and a useful film shot is intention plus continuity.
- 01
Subject
Who or what carries the moment?
- 02
Action
What changes during the shot?
- 03
Camera
Where are we, and how does the lens move?
- 04
Light
What time, direction and emotional temperature?
- 05
Continuity
What must remain identical to the shots before and after?
A strong prompt describes a decision. A strong sequence protects that decision from shot to shot.
Continue the conversation
Build a story people remember.
Open to creative commissions, cultural projects, AI-film conversations, music visuals and purposeful collaborations.