Design Assist AI

AI UX Assistant for Product designers

Project Overview

Design Assist is an AI-powered tool that intuitively understands your needs, automates routine tasks and guides you toward design system compliance—all without interrupting your creative flow.
The design, development, and implementation of Design Assist improved how product designers at Meta interact with our “dim sum” design system. Integrating directly within Figma, Design Assist provides contextual guidance, automates compliance checking, and offers real-time assistance that significantly reduces the cognitive load of remembering complex design system rules. Through seamless integration with Meta’s Blueprint design system and partnership with the MetaGen AI team, our work has improved design consistency while reducing time spent on manual compliance tasks, with over 600 designers now using the tool daily to create more consistent, higher-quality designs.

My Role

As the Design Manager at Meta, I led an AI-powered design tool’s vision, UX strategy, and implementation, collaborating closely with cross-functional teams, including MetaGen AI engineers, product managers, and the Blueprint design system team. I directed research studies to identify designers’ pain points related to design system adoption and created and tested multiple iterations of the conversational UI and contextual features. Additionally, I oversaw the tool’s integration with Figma and Blueprint design system components, ensuring a seamless user experience.

Confidentiality: This case study's insights and design process reflect my perspective. Specific details have been modified to protect proprietary information while accurately showcasing my design approach and the project's impact.

Project Duration: April 2024- January 2025

Key Partners: MetaGen, Blueprint Design System team, Figma

Team: Fas Lebbie, EISA Engineering

Problem Context

Only 46% of designers consistently implement Blueprint design system components at Meta. Research identified that even experienced designers spent up to 25% of their time searching for correct components, struggling with compliance issues, or rebuilding existing components from scratch. This inefficiency stemmed from a fundamental gap between designers’ creative workflows and the growing complexity of the Blueprint design system, which contained over 3,000 components across multiple platforms. The traditional approaches to design system education—such as documentation websites and design reviews—proved inadequate for real-time implementation needs, creating a high cognitive load that disrupted creative processes. As Meta’s Enterprise Infrastructure Services and Analytics org product ecosystem continued to expand, this disconnect resulted in inconsistent user experiences, increased design debt, and longer production timelines. The situation highlighted a need for a more integrated approach to design system implementation—one that could provide contextual assistance without forcing designers to context-switch between creative work and system compliance.

Design Interventions

Our design intervention focused on product designers at Meta who regularly work with the Blueprint design system in Figma, based on research showing that 78% struggled with system compliance while maintaining creative flow. These designers spend approximately 12 hours per week searching for components, checking compliance, or recreating existing patterns. To address this challenge, we developed Design Assist, an AI-powered tool that integrates directly within Figma, providing contextual guidance through a conversational interface, automated compliance checking, and blueprint component suggestions.

Scroll to next section

My Approach

Design Process

Design Research & Strategy

Our research journey began with a foundational hypothesis: AI could serve as a “raw material” for design—not replacing human creativity but augmenting it at critical decision points. To test this, we needed a comprehensive understanding of designers’ actual workflow challenges and system implementation barriers.
We conducted multi-layered research combining 17 in-depth interviews with Meta product designers, 20 contextual inquiries observing real-time workflow patterns, and a quantitative analysis of 300+ Figma files. We also mapped the design system implementation journey and identified critical moments where designers abandoned system compliance, with compliance errors averaging 12 per file and system-related tasks consuming design time.

Summary of Findings

This mix of methodologies revealed Blueprint and dim sum the current design system contained over 6,000 components across multiple platforms, designers spent an average of 16 hours weekly on system-related tasks, and despite 92% acknowledging the importance of design systems, only 46% consistently used them. The research identified three key opportunity areas for intervention. First, designers needed real-time verification of design system compliance without disrupting their flow—87% reported abandoning system components when verification required multiple steps. Second, discovering relevant components within the context of their current work was crucial, as designers were 3.5x more likely to use system components when they could preview them directly in their working files. Third, designers needed just-in-time learning about system rules and patterns, with 92% preferring contextual snippets of information over comprehensive documentation. These insights revealed that successful design system implementation isn’t just about better components or documentation—it’s about seamlessly integrating assistance within the creative process. We determined that an AI-powered design assistant can be a raw material for designers who can adapt to individual designers’ workflows and project contexts.

Prototyping & Implementation Strategy

Our prototyping approach evolved through collaborative design workshops and iterative testing to balance technological feasibility with user experience goals. We conducted three structured design workshops (February- March 2024) where cross-functional teams collaborated to explore and refine Design Assist concepts. These workshops brought together product designers, AI specialists, and Blueprint system experts to co-create solutions using the XDS Design System components and GenAI tooling.

The workshop structure followed a progressive development path:

Concept exploration and ideation based on research insights
Feature refinement and prototype development with critique sessions
Demo preparation and implementation planning with engineering teams

This collaborative approach allowed designers to prototype the designs directly. The phased rollout strategy begins with essential Blueprint component detection, followed by automatic compliance checking (increasing adoption by 42%), and finally, introducing contextual recommendations and iteration features.

Each implementation phase incorporated user feedback, particularly around maintaining creative autonomy while providing system guidance. Technical implementation required close collaboration with the MetaGen team for AI capabilities and the Blueprint team for design system integration, ensuring accurate component detection and recommendations while adhering to Meta’s privacy and security.

Reflections & Impact

Within the first six months of launch, Design Assist achieved remarkable adoption, with over 600+ d signers using it daily. Blueprint component usage increased by 72% across product design teams, while designers reported a 46% reduction in time spent searching for and implementing system components. The quality scores for design system compliance improved by 13 points on average, a fundamental shift in how designers perceived system compliance—from a constraining requirement to an integrated aspect of their creative process. These changes became visible in more consistent user interfaces across Enterprise Engineering org Meta products. By embedding system guidance directly into creative workflows, we’ve begun influencing how other design tools are being developed with Enterprise Infrastructure org at Meta, putting to practice leveraging AI as a “raw material” for design.

All Projects

Next Project