Producing complex digital experiences requires alignment. UX and Product Designers have a big bag of goto artifacts and workshops at their disposal to achieve alignment. One of the tried-and-true artifacts is the storyboard. In this experiment I take a look at using AI to generate storyboards for a few key scenarios supporting the Vehicle Passport concept.
Explore emerging tools and technologies to identify a storyboarding workflow that might augment and/or replace classical methods.
Direct: exploit dedicated storyboarding AI platforms for one-shot storyboard comic strip creation
Indirect: generate each comic panel independently and then assemble in Figma to create the comic strip
Manage character and environment consistency across multiple comic strips in order to achieve a unified look and feel
Use the tools and AI agents I’m already paying for to produce satisfactory storyboard results (keep costs down)
Storyboarder.ai
Storyboarder.ai is one of several dedicated platforms that help creatives and cinematic storyboard artists generate storyboards. I toyed with the platform’s Free plan but was not compelled to move forward with a paid version. The platform felt quite capable. However, I consider myself a “casual user” and as such the platform was a bit overkill for my needs.
Insights
Format: traditional page and form flow model
Robust features and functions to build and edit storyboards
Upload character assets; high quality renderings (pencil sketch effect was great) and character and scene consistency was only mildly difficult to achieve
Multiple export options
Geared more towards the dedicated storyteller working on movies, books or other
Storyboarder.ai traditional page/form interface as seen on desktop display
Gemini Pro + Nano Banana
The experience was cumbersome. I must admit… I was was pretty lazy with the prompting and the results reflected generalizations that felt sloppy.
Insights
Format: Chat
Difficult to achieve character and scene consistency. Renderings got increasingly “sloppy”
Multiple export options
Gemini Pro + Nano Banana chat interface as seen on desktop display
Weavy.ai + Gemini + Nano Banana Pro
I was most skeptical about Weavy… and then I tried it. A few short Youtube videos and I was generating comic panels that maintained scene and character consistency around 80-90% of the time. Weavy’s Prompt Concatenator node will generate a well structured prompt using an agent which helps speed things along.
Insights
I found Weavy’s node user interface paradigm to be intuitive
The rendered images were consistent and of good quality
The platform provides a library of functional nodes as well as a variety of models to work with
Weavy.ai node interface as seen on desktop display
