Why 18 AI Roles Make a Short Drama Crew Run: The Division-of-Labor Logic Behind a Digital Film Set
A digital crew is not one AI pretending to make a whole drama. It is a production system in which different AI roles own distinct jobs—story, continuity, casting, art, camera, direction, and post—so each step has something reviewable before the next begins.
The core idea: AI works better when it has a job title
A digital crew is an AI production structure modeled after a real vertical short-drama set. Instead of asking one model to “write, design, direct, shoot, and edit,” the system assigns each task to a role with a clear scope, a handoff point, and a deliverable.
In Maosika (猫斯卡), that structure is organized as 18 digital specialists: 5 core specialists for the creative and screenplay pipeline, plus 13 video-production specialists for the shooting pipeline. The number is not a gimmick. It reflects a production principle: short-form drama breaks when one model is asked to remember everything at once.
Why role separation matters in vertical short drama
Vertical dramas are fast, dense, and highly serialized. A single season may contain many episodes, recurring characters, running conflicts, props, costumes, locations, and cliffhangers. If one generative process handles all of that without structure, the usual failures appear:
- A character changes face, outfit, or age between scenes.
- A prop introduced in one episode disappears in the next.
- The story resolves a conflict that was never set up.
- A scene begins with a weak hook instead of a strong opening.
- The video prompt reads like fiction instead of a shot instruction.
Role separation solves this by turning a vague creative request into a sequence of accountable production tasks.
The 5 core specialists: story and screenplay pipeline
The first group handles the work that happens before “camera.” These roles make sure the story is locked, the characters are defined, and the continuity record is maintained.
| Role | Main responsibility | What it produces or protects |
|---|---|---|
| Archive Secretary | Maintains the structured story archive | Character facts, relationships, unresolved threads, episode appearances, prop descriptions |
| Producer | Evaluates and guides the initial idea | A stronger brief, including conflict, hook rhythm, ending direction, platform, and audience |
| Screenwriter | Writes episodes in batches | Beat sheets, episode scripts, cliffhangers, dialogue, and scene-level structure |
| Process Supervisor | Enforces the writing workflow | Planning before drafting, continuity checks, mid-story convergence, and rewrite rules |
| Script Supervisor | Tracks writing-stage continuity | Batch notes, story-state updates, and consistency across episodes |
The important concept here is that the system does not rely on model memory alone. It uses a structured story archive—similar to a continuity bible—so later episodes can retrieve the relevant character state, unresolved threads, recent batch summary, and current main plot.
The 13 video-production specialists: from prep to final take
Once the script and visual direction are ready, the production pipeline shifts into a film-set-style crew. These 13 roles are grouped into four stages.
1. Scene preparation
| Role | Responsibility |
|---|---|
| Script Supervisor | Aligns scenes with the script and shooting blocks |
| Casting Director | Confirms character lineup and visual casting requirements |
| Production Coordinator | Organizes scene-level production readiness |
This stage prevents the system from generating a scene before the required characters, props, and scene references are identified.
2. Visual development
| Role | Responsibility |
|---|---|
| Visual Style Director | Selects and applies the overall visual style |
| Art Director | Defines scene and environment language |
| Makeup and Styling Specialist | Creates character look and costume consistency |
| Props Master | Extracts and catalogs props using the script’s original names |
This is where character consistency begins. The system includes multiple style manuals covering 2D, 3D, realistic, anime, period, fantasy, clay-stop-motion, cyber-Chinese-style, and other directions. After a style is chosen, character art, scene art, props, and video prompts follow the same visual path.
3. Shooting execution
| Role | Responsibility |
|---|---|
| Storyboard Artist | Converts scene action into shot-level instructions |
| Director of Photography | Controls camera language, lighting, and composition |
| Director | Shapes performance, pacing, and scene intent |
| Camera Operator | Produces the final shot-level execution direction |
The key discipline here is the shot prompt. Instead of asking the video model to improvise, the system constructs prompts around eight elements: subject, action, environment, lighting and tone, camera movement, visual style, image quality, and constraints.
4. Post and quality control
| Role | Responsibility |
|---|---|
| Editor / Selector | Reviews takes and supports selection from multiple versions |
| VFX Technical Supervisor | Checks technical output and visual stability concerns |
The system does not claim to automatically choose a “viral” take. It gives creators multiple takes, editable prompts, reference-image controls, and a retake workflow. Final quality judgment remains with the creator.
How the 18 roles work together in production order
The production sequence is more important than the role count. A digital crew should follow the same logic as a disciplined set:
- Evaluate the initial idea.
- Guide missing creative dimensions.
- Lock the creative brief.
- Confirm the character lineup and visual direction.
- Build the story archive.
- Write scripts in batches, with beat sheets and rule checks.
- Select a visual style.
- Generate character, scene, and prop art.
- Split each episode into roughly 10-second scene blocks.
- Bind each scene to its references and shot prompt.
- Generate vertical video, usually in 9:16 by default.
- Review takes.
- Revise dialogue, references, or prompts and reshoot when needed.
Each stage produces something inspectable: a brief, an archive, a script, a look, a reference image, a scene block, a prompt, or a take. That is what makes the process controllable.
Why “one AI operator” is the wrong mental model
A common mistake is to imagine AI drama production as a single command: type a premise, receive a finished series. That model fails because drama production is not one task. It is a chain of decisions with dependencies.
A better mental model is a small digital film crew:
- The producer decides whether the idea is ready.
- The screenwriter writes from a locked brief.
- The archive secretary prevents continuity drift.
- The props master tracks objects by their script names.
- The styling specialist locks the character look.
- The storyboard artist and DP translate action into camera language.
- The director and camera operator shape the shot.
- The editor and VFX supervisor support review and iteration.
The human creator still acts as showrunner, approving key checkpoints and making aesthetic judgments. The digital crew handles the repetitive, drift-prone work that is hard to maintain across dozens of episodes.
What a digital crew does not do
It is important to state the limits clearly:
- It does not guarantee a hit.
- It does not replace the screenwriter, director, or producer.
- It does not automatically score finished videos or choose the best take with creative certainty.
- It does not force every missing reference image; creators may skip references, though quality usually drops.
- It does not make character consistency depend on face-embedding verification alone; consistency depends on locked costume and makeup assets, reference rules, and prompt discipline.
- It does not replace final editing, sound design, music choices, or platform-level packaging.
The product unit is best understood as one scene block with controlled references producing one video segment, not a fully automatic one-click feature film.
About Maosika
Maosika (猫斯卡) is an AI production operating system for vertical short dramas. It organizes production as a digital crew with 18 specialized roles, covering story development, scriptwriting, continuity, visual development, scene blocking, shot prompts, video generation, and review. Its position is not “one click to a hit.” It turns repetitive, drift-prone, and hard-to-control parts of short-drama production into a reviewable pipeline, while aesthetic judgment remains with the creator. Learn more at www.maosika.com.
About Maosika — Maosika · Professional AI Video Production System. It connects briefing, scripting, look development, shot prompts and delivery into one reviewable pipeline. www.maosika.com