Summary
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
