Faraz Mirza
Ed.gov Project Thumbnail.png

FIGMA MAKE



Figma Make Conversational Sidebar
Figma Make Version Control Bar

FIGMA MAKE

IMPROVING AN AI-DRIVEN DESIGN INTERFACE

CONCEPTUAL CASE STUDY

PURPOSE

This redesign aimed to make Figma’s latest AI-powered design tool, “Figma Make”, more intuitive by streamlining accessibility to prompt-driven iterations, guided by heuristic and ethical AI considerations.

TIMELINE

2-week sprint from August 19 – September 2, 2025.

TEAM SIZE

I executed the redesign as both UX Researcher & UI Designer.


FEATURE FOUNDATION

When Figma Make exited beta testing, I experimented with prompt-driven design by creating "AI Spy", a functional dashboard that scrapes job boards to list the top 5 AI tools in demand.

I noticed that each prompt automatically labeled each design iteration with:
• An unchangeable title
• A chronological version number
e.g. "AI Tools Demand Dashboard, Version 5".


This causes a few observable usability issues:

• Past versions cannot be “viewed” natively, but only “restored” as the next chronological iteration (e.g. Version 1 can only be revisited as Version 6)

• Past versions with identical titles are indistinguishable


The result is a conversational sidebar cluttered with stacks of duplicate versions, complicating access to desired iterations.


RESEARCH & FINDINGS

To formalize observations into actionable insights, a heuristic evaluation of Figma Make revealed:

Version control is confusing: Unchangeable labels, misleading renumbering, no timestamps, and poor version differentiation force users to rely on memory, increasing errors and complicating decision-making.

Version navigation is complex: Blended past and new iterations, redundant versions, and difficulty of comparison reduce clarity and workflow efficiency.

The unique dimension of these issues is that AI plays a significant role them. According to the EU’s 2019 “Trustworthy AI” ethical guidelines:

  • AI-assisted version management limits human agency & oversight by automating version states and guiding workflow decisions.

  • AI-driven naming and numbering reduce transparency, making it difficult for users to track changes or understand version history.


INTERACTIVE PROTOTYPE

To clean up the AI interface, I designed a version control bar prototype providing consistent, scalable access to prompt-driven iterations across desktop, mobile, or code preview, on Figma Make’s interface.

Each iteration included updated UI components, wireframes, and interactive functionality.

The mechanism would change such that:

• Version numbers would be retained once assigned
• Versions could be viewed without producing duplicates
• Restoring a version would automatically scroll to the design-relevant prompt

Theoretically, user experience would significantly improve without fundamentally altering the existing interface.


LIMITATIONS & NEXT STEPS

Due to a lack of usability testing, the effectiveness of the updated version control workflow in real-world scenarios remains unvalidated:

• Users may experience friction when navigating restored iterations or interpreting retained version numbers.
• AI-driven prompt labeling could still create ambiguity in distinguishing iterations, particularly for complex projects.

Further testing is needed to measure efficiency, clarity, and error reduction.

• Given more time, iterative refinement based on user feedback could optimize accessibility, transparency, and overall workflow.
• Future steps may explore enhancing user control through iteration rename-ability, or a hierarchal restructuring of AI-assisted naming conventions.


 

TOOLS USED

Copyright © 2015-2025