Moving From Still Images to Video

Once I had a visual direction, I began experimenting with AI-generated motion.

This was one of the more interesting parts of the project because a still image can look convincing while animation exposes inconsistencies very quickly.

The prompts had to describe not only what the scene looked like, but how it behaved over time.

For example, I experimented with instructions for subtle movements such as breathing, eye movement, small head turns and tail movement. I also worked on having the chameleon’s colors gradually transition through individual scales instead of instantly switching from one color palette to another.

Camera movement was another consideration. Slow zooms and controlled movement helped create more of a polished commercial feel while leaving negative space for typography that could be added later.

Editing in Adobe Premiere Pro

The AI-generated clips were only the raw material.

I brought everything into Adobe Premiere Pro to begin shaping the actual piece. This included:

• Selecting the strongest generated clips
• Trimming and sequencing shots
• Adjusting timing
• Testing transitions
• Working on fades
• Adding music and sound
• Correcting color differences between clips
• Improving the overall pacing and flow

One challenge was getting separate AI-generated clips to feel like they belonged in the same video.

AI can produce a great individual shot, but the lighting, movement, framing or color can change slightly in the next generation. A lot of the traditional editing process is still necessary to bring those pieces together.

Using Photoshop and Traditional Design Skills

I also used Adobe Photoshop for image cleanup, compositing and preparing supporting visual elements.

This is one of the biggest things I’ve learned while experimenting with generative AI: traditional design skills still matter.

Composition, typography, hierarchy, color, masking, retouching and an understanding of branding don’t disappear because AI is involved. If anything, those skills become even more important because someone still needs to decide what looks right and what doesn’t.

AI can generate options quickly. The designer still has to art direct them.

The Tools

The project combined several tools rather than relying on a single AI platform:

Adobe Firefly — AI image generation, visual exploration and experimentation with motion.

Adobe Premiere Pro — Video editing, sequencing, transitions, timing, music, sound and color correction.

Adobe Photoshop — Image editing, cleanup, compositing and preparation of visual assets.

Generative AI prompting — Developing and refining detailed prompts to control composition, camera movement, lighting, animation and color transformations.

Most importantly, I approached AI as another tool within an existing creative workflow rather than as a replacement for that workflow.

What I Learned

The biggest lesson from the Chameleon Project has been that AI generation is only one part of AI-assisted creative production.

Getting a usable result often requires multiple generations, revised prompts and plenty of trial and error. Then the generated material still needs editing, design decisions, timing, sound and polish.

There are also imperfections. Transitions can be smoother. Color consistency can be improved. Motion can occasionally feel unnatural. Those are exactly the reasons I wanted to work on a project like this.

The goal wasn’t perfection on the first attempt.

The goal was to learn.