Nate Zuar - Projects

2025-Current

In the AI creative platform, Runway, I have developed projects ranging from in-store retail use case videos, high production style social content, off-roading event branding and mockups, and more.

Runway is a very powerful tool that allows for a creative like me to take an idea from a concept to an execution level asset. My experience in Runway spans across many AI services, including Claude and ChatGPT for prompt engineering, Nano Banana Pro and GPT Image 2 for image generation, Seedance 2.5 and Kling 3.0 Pro for video generation, and several other tools for quality of life improvements like TopazAI for upscaling and Aleph for video edits.

Image & Video
Generation,
Creative Curation,
Reference & Prompt Engineering

As a Senior Art Director, my workflow for an AI-produced video project typically includes the following:

Shot list + reference development

Working with Claude and ChatGPT to develop what shots are needed for the video, and creating or researching the references needed for both image and video generation stages

Creation of video stills

Using image platforms like Nano Banana Pro and GPT Image 2 to establish starting frames for videos, as well as ending frames if needed. Depending on the project and products being showcased, this step is crucial to preserving exact accuracy of product details.

Photoshop + curation cleanup

After creating my starting frames, the next step is to QC the stills so that products are shown correctly. Typically, this step in the process requires re-generating for specific details, then masking the updated areas via Photoshop to prevent AI artifacting, warped geometry & oversaturation of colors.

Video generation

Using final highly-detailed stills, I supply them to ChatGPT or Claude with any additional video references that will be used to craft the prompts needed for realistic, high quality videos. This step can range from the shortest or the longest in the AI project development process, solely based on the realism of the asset needed by the client. For example, a clip of a person using a new cutting-edge phone is easier to achieve compared to crafting a big-budget CGI style clip of something that’s more conceptual. This is because there’s more data for video models to pull from of day-to-day movements rather than big hyper-specific ideas that are more loosely based on reality.

Quality of life improvements + visual proofing

Arguably the second most important step in the process, making sure the end product doesn’t look like AI slop. This step is often overlooked in most AI productions, but it’s crucial for audience’s immersion when viewing the final product. In addition to visual proofing, this is also the step where I use additional tools on Runway like raising the resolution to 4K on all videos, tweaking frame rates or fine-tuning any generated audio. Lastly, I’ll typically edit my clips in Adobe After Effects to create a rough cut of the full video containing all of my clips. This is sometimes the final step, but often when working with a dedicated video editor, reference cuts are crucial as there are some moments in clips more beneficial to keep than others.

Below are a collection of projects I’ve developed from concept to final delivery.