An AI director develops the idea, decides how it should look and move, and judges which images belong in the finished film. Generative tools can be part of that process. They do not turn a treatment, a promising frame or a rough animatic into an approved production by themselves.
For a music video, the starting point is the relationship between the song, the artist and the images. For a commercial, it is the brief and the product. For a narrative film, it is the scene and the characters. A useful production plan makes those choices explicit before selecting a tool.
Start with the scene, not the software
Define the intended audience, running time, release formats and the scene's purpose. Decide what needs an actual performance, a physical product or a real location, and what could be explored through generated reference material. Keep reference images separate from approved final assets.
A treatment should describe composition, lighting, movement, pacing and sound. A shot list should explain what each shot contributes. These are creative decisions that can be reviewed independently of whichever camera or model might eventually make the image.
Test one sequence before expanding
A practical starting point is a small sequence with a clear beginning and end. Review the frames together, not just as attractive stills. Check continuity, performance, spatial relationships, motion, product details and how the cut reads with sound.
Record which parts were filmed, generated, composited or edited. Keep the input references and revision history with the test. When a result cannot preserve the detail a scene needs, change the production method rather than promise that another generation will solve it.
Keep approval and finishing in the plan
Agree on who approves the treatment, test and final cut. Before production, establish permission to use supplied footage, music, products and likenesses, and the disclosure expectations for generated material. Include editing, sound, colour, versioning and final quality review in the schedule.
A previsualization may be useful for planning without being suitable for release. The delivery decision belongs to the complete film and its agreed requirements, not to the speed of the first generated clip.
Directing work with a public record
Amos Le Blanc's existing films provide a verifiable starting point for discussing his directing practice. Promo News credits Le Blanc as director of Young Empires' The Gates and describes filming across American locations and Toronto. Its coverage names Christopher Lew as cinematographer.
In The One Club's 2019 interview, Le Blanc discusses Run This, made with co-director Ohji Inoue, and his approach to finding the right form for a project. Campaign Brief's coverage of Tiger Beer's Sit Closer Together credits Le Blanc as director, with The Sweet Shop and Marcel Sydney.
These are records of directing work, not evidence that those productions used generative AI. They are not being relabelled as AI campaigns or endorsements of a current tool.
Measure what actually happened
A credible AI-production case study should show the brief, approved inputs, a dated test, the finished result, credited roles and the limits of the comparison. Record hands-on work, render time, review rounds and finishing separately. Compare equivalent deliverables before describing any saving.
The companion previsualization comparison guide sets out what to document. For current project inquiries, see AI direction; for completed films, see selected work.
Editorial note, 9 September 2026: Earlier timing comparisons and project-development assertions on this page were not supported by records available for review. They have been replaced with planning guidance. This article does not report a measured production experiment.