This spring, I gave a talk at Figma’s Design Systems Meetup to share the story of Stick’s first Communication Design System prototype. Our goal, from the beginning, was to establish a Design System that balanced consistency and control.
Here’s a bit more behind the journey:
The Design Challenge at hand: Can you turn a manual service into an automated system?
The Vision behind it: enable anyone to communicate visually, not just those with “design talent”
The video that inspired the design challenge: going from process to program
Step 1. Content: Reverse engineer the parameters based on an extensive analysis of existing visual communication assets. In this case, we started with slides and decks, publicly available on SlideShare, and took an inventory of all visual structures used, frequency, and use cases.
Step 2: Style. Understand the “levers” of branded visual communication
The challenge: ultimate simplification of branded communication design. Could we simplify the factors driving a company’s look and feel into a few simple levers to turn the art into a science?
The goal: increasing complexity and personalization. Sidenote: in the early days we experimented with neural networks style transfer as a way to generate branded assets. We were a bit early :)
Experiment: Could we take structured text and systematically “translate” it into visuals? Ran the first experiment with class content from Stanford MBA curriculum, provided by “third-party” authors (classmates). Left: Before, text. Right: After, styled visual.
Using NLP APIs to extract key ideas from the text content.
Step 2: Using Typeform form to force author to make styling decisions by choosing objectives, not outcomes.
The result: Bringing it all together.
A closer look: The Prototype’s style parameters: (1) Casual vs. Formal determined typography, (2) Hot vs. Cold determined color, (3) “Rubik’s Cube” vs. “Rubber Band Ball” determined visual structure, (4) Left Brain vs. Right Brain determined icon illustration style, and (5) Coke vs. Pepsi determined Background vs. Foreground balance.
The result at scale: Dozens of unique visuals, all from Arial text input, and quantitative style survey data
About the Author: Alli McKee is the CEO and Founder of Stick.ai, a communication visualization platform that illustrates your ideas, in real-time, using natural language processing and machine learning. Built for B2B Sales and Success teams, Stick helps you communicate more more clearly and quickly so your ideas actually stick. For more, visit www.stick.ai.
