Descriptions:
Dan Kieft breaks down his workflow for creating cinematic AI videos using Higgsfield Cinema Studio, covering five core skills he considers essential for serious AI filmmaking. The video opens with a short film composed entirely of AI-generated shots before diving into the techniques that produced them.
The first skill is structured prompting: Kieft uses a custom GPT trained specifically for Cinema Studio to transform vague ideas into detailed, camera-aware prompts. The GPT suggests lens choices—Sony Venus and Hogy Light at 14mm for wide establishing shots and 35mm for closer work—and generates multiple scene variants from a single concept. The second skill is shot consistency, achieved through Higgsfield’s multishot feature, which generates nine distinct camera angles from a single reference image automatically, replacing what previously required manual re-prompting for each angle. The third focus is character realism: Kieft argues that AI image generation has become polished enough that filmmakers must now actively prompt for imperfections—skin pores, stubble, under-eye texture—to avoid an artificially smooth look. He demonstrates this using Nano Banana Pro inside Cinema Studio with reference photos of himself, showing the measurable difference between a generic portrait and a photorealistic one by prompting explicitly for natural skin detail.
Viewers come away with a concrete, repeatable system for producing high-quality AI video content, grounded in specific tool settings and prompt strategies rather than general advice.
📺 Source: Dan Kieft · Published February 03, 2026
🏷️ Format: Tutorial Demo
