Back office ops · Production

Uber builds agentic system uSpec to automate design specs in minutes instead of weeks

The problem

Uber's design systems team faced an enterprise-scale documentation bottleneck: manually producing component specs for hundreds of components across 7 implementation stacks was slow, inconsistent across authors, and drifted out of date whenever a component changed, turning documentation into a dedicated workstream that still could not keep pace.

Workflow diagram · grounded in source
1
Designer shares Figma link and context
trigger
“open Cursor and reference an agent skill with your Figma component link. Add context about states, variants, or platform-specific behavior that the agent can't infer from the design alone”
2
Agent connects and extracts Figma data
integration
“the AI agent connects to your Figma file through the Figma Console MCP, extracts component structure, design tokens, variables, and styles, then maps the parent-child relationships between layers and sub-components”
3
AI interprets and classifies component
ai_action
“AI judgment where interpretation matters—classifying accessibility semantics, selecting the right token mappings, deciding how to structure a spec”
4
Spec rendered directly in Figma
output
“it imports the matching documentation template from your library, detaches it, and populates it—filling text fields, cloning sections, building tables, and placing markers—all through the MCP”
Reported outcome

The agentic system generates complete component specs in minutes instead of weeks; a full screen reader spec covering 3 platforms now generates in under 2 minutes, and a system of hundreds of components that previously required months of spec-writing can generate complete specs in days.

Reported metrics
Spec generation timeminutes instead of weeks
Screen reader spec generation time across 3 platformsunder 2 minutes
Full system spec-writing timemonths to days
Reported stack
uSpecCursorFigma Console MCPFigma
Source
https://www.uber.com/us/en/blog/automate-design-specs/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

The agentic system generates complete component specs in minutes instead of weeks; a full screen reader spec covering 3 platforms now generates in under 2 minutes, and a system of hundreds of components that previousl…

What tools did this team use?

uSpec, Cursor, Figma Console MCP, Figma.

What results were reported?

Spec generation time: minutes instead of weeks; Screen reader spec generation time across 3 platforms: under 2 minutes; Full system spec-writing time: months to days (source-reported, not independently verified).

How is this back office ops AI workflow structured?

Designer shares Figma link and context → Agent connects and extracts Figma data → AI interprets and classifies component → Spec rendered directly in Figma.

Related back office ops cases