Using AI for Pre-Production Mood Boards and Look Development – New Lesson in “Directing the Future” on MZed

MZed Education

by Nino LeitnerApril 3^rd, 2026

Subscribe to How To

The ever-evolving MZed course “Directing the Future: Ethical AI Video for Filmmakers” has received another update. In the latest lesson, cinematographer and photographer Drew Geraci demonstrates a practical, real-world use case for Higgsfield AI: generating pre-production mood boards and wardrobe looks for an actual photo shoot, then comparing the AI-generated references with the final on-set results.

One of the recurring themes throughout Drew Geraci’s course (original report here) is that AI should serve as a creative tool, not a replacement for real production work. This latest lesson puts that philosophy into practice more directly than any previous module. Rather than exploring a platform’s features in isolation, Drew walks viewers through a complete workflow where he used Higgsfield AI to plan looks and lighting setups for a portrait session with a model, then took those references on set and shot the real thing. The results sit side by side, and the takeaway is clear: AI-generated imagery can be a surprisingly effective pre-visualization tool for photographers and cinematographers alike.

From AI-generated looks to the real shoot

The lesson centers on a dual-model photo shoot that Drew planned using Higgsfield AI’s character creation and image generation tools (Drew added another lesson exploring the Higgsfield AI platform to the course recently). By uploading real photographs of his model, Michaela, into the platform, he was able to generate a wide variety of looks exploring wardrobe, lighting, and environment combinations. The brief was specific: professional, trendy, clean, and modern.

Drew started with the platform’s “Photo Dump” feature, which takes a character and automatically places them into a range of pre-designed scenes with different clothing, backgrounds, and lighting setups. The initial batch of generated images gave Michaela a broad spectrum of options to review before the shoot even began. When Drew showed her the AI-generated looks, she was taken aback by how realistic the results appeared. For just a few cents per generation (or essentially free with an unlimited subscription), the platform produced visuals convincing enough to serve as genuine wardrobe and mood references.

After the initial round, Drew refined his prompts to match the actual shoot location and concept more closely. He described a woman in modern clothing inside an old brick house with trendy furniture, white walls, soft white light, a pantsuit, and a tube top. Without further direction, Higgsfield generated images that closely matched the intended aesthetic. When Drew arrived on location for the real shoot, he already had a clear sense of how he wanted to position his subjects, how the lighting should fall, and what the overall tone of the session would be.

Why this matters for photographers and cinematographers

The practical value here extends well beyond novelty. Being able to generate realistic reference images of your actual subject in your intended environment, wearing approximate versions of their planned wardrobe, fundamentally changes how a photographer or director can communicate their vision. Drew points out that showing a model or actor these AI-generated references before the shoot gets them into the right mindset. It also provides a concrete visual brief for stylists, wardrobe departments, or anyone involved in the production’s look.

For the session itself, Drew photographed two models in the pantsuit collection and was able to recreate the moods he had pre-visualized through Higgsfield. The AI references guided not just wardrobe choices but also lighting placement and character interaction. One of the more compelling aspects of the lesson is seeing the AI-generated mood board images alongside the final photographs. The resemblance in tone and atmosphere is striking, even though the final images are, of course, authentic photographs shot on location.

Drew also explored generating images with multiple characters to see how different pairings would look in the same scene. While these were quick experiments rather than polished outputs, they demonstrate how a director could use the tool to test casting combinations or group compositions before committing to a setup on the day.

A deeper look at Higgsfield’s toolset

Beyond the mood board workflow, the lesson briefly touches on several other Higgsfield features that could prove useful in a production context. The platform offers face swapping, image upscaling, in-painting, and a relighting tool that Drew finds particularly interesting. If the lighting in a photograph did not turn out as planned, the relighting feature allows users to adjust it after the fact, with exports available at up to 4K resolution. The platform also provides a wide selection of style presets and supports multiple AI models, giving users flexibility in how their generated images look and feel.

Drew demonstrated the workflow live during the lesson, creating a character profile for Michaela and then generating new images using a custom prompt describing an old abandoned house with volumetric lighting, dramatic three-point lighting, and a dark, moody atmosphere. Generation times vary depending on server load, ranging from a few seconds during off-peak hours to several minutes during busy periods.

Keeping it authentic

The most important point Drew makes throughout the lesson is one that runs through the entire course: the AI-generated images are references, not deliverables. The technology is being used to plan, communicate, and pre-visualize, while the final creative output remains entirely human-made. For Drew, this approach represents a genuine shift in how productions can be planned. If you have a location already scouted (or even just found online), you can combine those real backgrounds with your character’s likeness to create a near one-to-one preview of the final scene before anyone steps on set.

As the tools continue to improve, with faster rendering, higher accuracy, and more detailed prompt interpretation, the pre-production applications for photographers, cinematographers, and directors will only become more refined.

If you would like to see Drew Geraci’s full demonstration of this workflow and his broader exploration of AI tools for filmmakers, watch the complete course “Directing the Future: Ethical AI Video for Filmmakers” on MZed. The new lesson is already available for MZed Pro subscribers. The course is continually updated with new modules covering the latest tools and techniques as they emerge. Previous lessons have covered Sora 2 for storyboarding, hybrid production workflows with Google Flow, and an earlier basic introduction to Higgsfield AI.

With MZed Pro, you can stream hundreds of hours of filmmaking education. We continually add new courses (several are currently in production). Starting at just $29/month (billed at $349 in the first year, $199 from the second year, or billed monthly at $49), here is what you will get:

• Highly produced courses from educators with decades of experience and awards, including a Pulitzer Prize and an Academy Award.
• Unlimited access to stream all content during the 12 months.
• Offline download and viewing with the MZed iOS app.
• Discounts to ARRI Academy online courses, exclusively on MZed.
• Course topics include cinematography, directing, lighting, cameras and lenses, producing, indie filmmaking, writing, editing, color grading, audio, time-lapse, pitch decks, and more.
• 7-day money-back guarantee if you decide it is not for you (on annual billing only).

Full disclosure: MZed is owned by CineD.

Join MZed Pro now and start watching today!

Have you ever used AI-generated images as mood boards or wardrobe references for a real shoot? Would this kind of pre-visualization change how you plan your productions? Don’t hesitate to let us know in the comments below!

Feature image source: Drew Geraci / MZed.