On July 3rd, SELISE Total Experience Lab organized a TXL Hackathon workshop to explore how Figma’s 2026 AI updates can streamline our daily design operations. Instead of a standard lecture, we turned it into a hands-on problem-solving sprint. 
We divided the UX participants into three teams with a clear challenge: identify a real, everyday workflow bottleneck and build a solution using Figma Agent.

The strict rule for the exercise was that there could be zero manual workflows; everything had to be executed using AI agents. After brainstorming dozens of everyday frustrations, the teams narrowed their focus and built out three core solutions: 

1. AI-powered white labeling control 

The problem: White-labeling UI designs are tedious and time-consuming because every brand requires unique colors, logos, and styling updates across countless screens. 

 The solution: The team conceptualized an AI system to quickly customize existing UI designs and manage all modifications from a single control point. This makes the white labeling process significantly faster, easier, and more visually consistent. 

2. “Draft to design” for pitch projects 

The problem: We frequently end up with a stack of approved screens, like a successful client pitch, that were built fast and never assigned to a formal design system.

The solution: A plugin that runs on unassigned screens to analyze and extract the raw styling (colors, typography, border radius, gaps, padding, and shadows). It instantly turns them into actual Figma Styles, generates a separate style guide page, and exports a design markdown (.md) file to integrate seamlessly with other AI tools. 

3. Taming mid-process design drift

The problem: Deep into versioning and creative exploration, designs naturally drift away from the core system, leaving detached, raw values across the canvas. 

The solution: Expanding the extraction plugin, this team created a one-click fix that converts extracted styles into Figma Variables and applies them directly back onto the original frames, ensuring no element is left detached. They also added a “refine and reconcile” layer to clean up near-duplicate values and flag inconsistencies, enforcing strict design consistency.

The takeaway

The hackathon proved that AI’s biggest immediate value isn’t just generating new concepts; it’s wrangling our chaotic workflows. By automating tedious white labeling and instantly pulling structured design systems out of raw drafts, we can spend less time doing cleanup and more time solving real problems.