The New UX Workflow: From Figma-First to AI-Augmented Design
How I transformed the UX design process at Thomson Reuters by integrating AI tools—cutting project timelines from 6-8 weeks to 2-3 weeks while improving quality and team collaboration.
Old Workflow
Manual design in Figma, lengthy handoffs, weeks of iteration
New Workflow
AI-augmented design, continuous collaboration, rapid iteration
As Lead UX Designer at Thomson Reuters, I recognized that our traditional Figma-first workflow couldn't keep pace with the demands of modern product development. With weekly research cycles, faster engineering timelines, and the rise of accessible AI tools, I led the transformation of our design process—integrating AI throughout the workflow while keeping designers at the center of strategic decisions.
This case study documents how we evolved from spending weeks on manual design tasks to leveraging AI for rapid exploration, freeing our team to focus on strategy, creativity, and user empathy.
Impact at a Glance
Reduced project timelines from 6-8 weeks to 2-3 weeks by integrating AI throughout the design process
Cut research synthesis time from 3-4 hours to 15 minutes using Claude for transcript analysis
Led team of 4-5 designers to successfully adopt AI-augmented workflow across all projects
The Old Reality: Where We Spent Our Time
Before AI integration, our design process was heavily weighted toward manual, time-consuming tasks that left little room for strategic thinking and innovation.
Manual design work
Days spent creating wireframes, prototypes, and design iterations from scratch in Figma
Research synthesis
3-4 hours manually reviewing transcripts and clustering themes after each research session
Prototype creation
Hours building clickable prototypes for user testing, often requiring rework after feedback
Content writing
Lorem ipsum placeholders everywhere, leaving content decisions until later stages
The Catalyst: What Changed
Weekly Rolling Research
We started testing current features AND exploring future concepts simultaneously. This meant we needed to move faster—traditional design timelines couldn't keep up with continuous discovery.
Engineering Moved Faster
After acquiring Materia, our engineering team adopted faster development practices. Builds often started before designs were "done." We were no longer weeks ahead—we needed a new way to stay aligned.
Product Communication Evolved
Lengthy PRDs became rare. Teams started "vibe coding" with tools like Lovable and V0—show, don't tell. We needed to adapt our collaboration model to match this new reality.
AI Tools Became Accessible
V0, Lovable, Claude, and other AI tools emerged and became accessible to everyone. UX design as we knew it was changing—and we had a choice: resist or adapt.
Our old workflow couldn't keep up. We needed a new way of working.
The New Workflow: AI-Augmented Design
We transformed every phase of our design process by strategically integrating AI tools—not to replace designers, but to amplify our capabilities and free us to focus on what humans do best: strategy, empathy, and creativity.
Discovery & Requirements
Old way
- • Wait for complete PRD
- • Start from scratch
- • 1-2 weeks preparation
New way
- ✓ Use GPT/Claude to create detailed personas
- ✓ Generate user journeys and scenarios
- ✓ Prepare technical UX questions for architects
- ✓ Arrive at workshops ready to facilitate
The shift: "Gathering requirements" → "Facilitating informed collaboration"
Time saved: 1-2 weeks → 2 hours + meeting time
Ideation & Exploration
Old way
- • Blank Figma canvas
- • Sketch concepts manually
- • Days to explore variations
New way
- ✓ Describe requirements to V0
- ✓ Generate flow diagrams with Lucid AI
- ✓ Iterate on different approaches rapidly
- ✓ Refine and polish in Figma
Flow Diagrams (Lucid AI)
Created comprehensive workflow diagrams showing AI processing stages, decision points, and user actions in minutes instead of hours.
Wireframes (V0)
Generated functional, interactive wireframes instantly—3-panel layouts, progressive disclosure patterns, and complex UI structures ready for testing.
What we still do: Refine, ensure brand consistency, add micro-interactions and edge cases
Time saved: Days → Hours (2-3 days for ideation vs weeks before)
Research & Validation
Old way
- • Manual transcript review
- • 3-4 hours per interview
- • Days to synthesize patterns
New way
- ✓ Upload notes to Claude
- ✓ Structured themes in 15 minutes
- ✓ Validate findings and add nuance
What we still do: Validate findings, add human nuance, present with empathy
Time saved: Prototype setup: Hours → Instant | Synthesis: 3-4 hours → 15 minutes
Content & Accessibility
Old way
- • Wait for content writers
- • Manually write alt text
- • 4-5 days for content
New way
- ✓ GPT/Claude trained on brand voice
- ✓ Generate tooltips, errors, empty states
- ✓ A11y-compliant patterns from day one
- ✓ Content partners review AI drafts early
Benefit: Real content and accessibility-compliant patterns from day one—their expertise shapes the output
Working with Product & Engineering
The Reality
Engineers were building fast—sometimes before designs were "done." Instead of fighting this, we adapted our approach to continuous collaboration.
Our Approach
Early sharing
Shared interactive video and V0 prototypes showing panel interactions before build started—engineers had high-level interaction sense from day one
Staying ahead
By detailed build time, final designs were ready—matched the fast pace without being left behind
Close contact
Frequent working sessions and continuous conversation—not just handoff meetings
The Results
What was built was almost identical to what we designed. Tighter alignment between design and engineering. Surprising in a good way.
Key takeaway: Handoff became collaboration. Artifacts became continuous conversation.
AI Tools in Our Workflow
We integrated multiple AI tools throughout the design process, each serving a specific purpose while keeping designers at the center of strategic decisions.
V0
Generating functional prototypes and wireframes
Claude
Research synthesis and content generation
Lucid AI
Creating flow diagrams and system architecture
GPT
Persona development and content writing
Impact & Time Savings
Old Workflow Total
- • Discovery & Requirements: 1-2 weeks
- • Design & Exploration: 2-3 weeks
- • Research Synthesis: 1 week
- • Content & A11y: 4-5 days
- • Handoff & Iterations: 1-2 days
New Workflow Total
- ✓ Discovery (AI prep): 2 hours + meeting time
- ✓ Ideation (V0, Lucid AI): 2-3 days
- ✓ Research: Instant prototypes + 2-4 hours synthesis
- ✓ Content & A11y: Integrated throughout
- ✓ Collaboration: Continuous
Net savings: 4-5 weeks → More time for strategic work
What's Working
- Faster exploration and iteration
- Coming to meetings more prepared
- Content and a11y integrated from the start
- More time for strategic thinking
- Tighter engineering collaboration
What's Still Messy
- Learning which prompts work best
- Knowing when to use AI vs. do it manually (requires HTML, Tailwind, React knowledge)
- Team adaptability to new processes
- Balancing speed with thoroughness—sometimes we miss edge cases
- Managing expectations—"AI = instant" but quality takes time
Key Lessons
AI works best as a collaborator, not a replacement
We're still the experts—AI amplifies our capabilities
Document your prompts
What works once will work again—build a prompt library
Content and accessibility from day one
Not bolted on at the end—integrated throughout
Embrace the messiness
We're still figuring this out, and that's okay
The Bottom Line
AI tools handle the grunt work, so we focus on strategy, creativity, and user empathy