One challenge of integrating AI into a design process is accounting for the meta aspects of what design is all about.
Prototype a design using AI? Sure!
Use AI to build multiple prototypes sooner in the design process? Sounds possible! Why not bring higher fidelity to the thinking sooner.
Turns out, this is more difficult than you might initially think. Articulating that there will be a system within a system up front is the missing step to starting in a coherent direction.
We developed a PRD that lays the groundwork for a modular prototype environment. This gives us clear ways to articulate different variations, as well as ways to switch between aesthetics and views.
The bottom-bar here shows the value. There are multiple variations that we can switch between that show drastic differences in layout approaches based on patterns popularized by other platforms. We can also swap in different themes and switch views using the bottom bar, avoiding the disorientation of navigating between different, potentially confusing, layouts.
This process proves out a few key things:
- PRD Value - Thinking through the process up front, even with prompting, helps to better articulate a system that can be built out coherently.
- Prototype Value - What decision needs to be made that the prototype is facilitating? Is it important to have the form, the functionality, or both?
- Design Process - One drawback of the design process has been articulating interaction and feel. This type of prototyping process allows for fluid experimentation and iteration within similar workspaces without getting overly caught-up in whether we’re looking at the final product.
Lots more to learn from here, but was excited to share some initial approaches.
You can explore the prototype here!