Context
Five years as a web developer, then a pivot to UX design in September 2024. I was the first designer in a B2B start-up, and I watched a junior developer use AI to outpace my own workflow. That observation, combined with the difficulty of getting design taken seriously in a fast-moving product team, shaped the research question.
Research Scope
UX, UI and Product designers and frontend/fullstack developers in French teams, across organisations from start-ups to multinationals. Primary research restricted to the French market to stay close to its culture and practices.
Year
June to September 2025
Research Question
How does AI integration transform collaboration between designers and developers, from scoping to handoff?
A Mixed-Methods, Action-Research Approach
Because the field was moving fast and existing data on the French market was thin, I chose action research: alternating between reading, observing, and testing inside my own company. Secondary research (newsletters, LinkedIn, international studies) set the baseline. Three field tests gave it a concrete anchor.
The tests covered each phase of the product cycle: briefing and scoping, prototyping, and handoff (Test 1: AI synthetic dev vs. real developers, Test 2: Figma traditional vs. Figma Make, Test 3: Figma traditional vs. Cursor with Figma MCP).
Research Methods
Semi-structured interviews (5 designers, 2 developers)
Qualified responses to the online questionnaire
Field tests covering scoping, prototyping, and handoff
Expert interview with a Lead Design Strategist
Key Findings
Individual vs. Collaborative
AI as Accelerator, Not Connector
Across interviews and the questionnaire, AI emerged consistently as a tool for individual productivity, not as a shared layer that improves design-dev communication. Adoption remains siloed by role and by habit.
Uneven Adoption
Developers Ahead, Designers Cautious
Developers adopted AI tools earlier and more systematically. Designers showed more reticence — part structural (tool immaturity for UX workflows), part cultural. This asymmetry didn't create the execution-speed gap initially hypothesised, but it did produce a perception gap.
Field Tests
Promising Bridges, Not Structural Shifts
The three field tests confirmed benefits at the task level (faster iteration, documentation support), but revealed structural limits: AI tools lack the contextual awareness to replace the human dialogue between designers and developers.
What Changes Things
Framework, Not Just Tools
Effective integration requires more than tool adoption: upskilling, a clear organisational framework, and adapted collaborative rituals. Without these, AI amplifies existing individual habits without bridging the designer-developer gap.
Full Analysis
Results and Recommendations
Published on UX Collective
The detailed findings, hypothesis validation, and 11 macro-recommendations for product teams are published on UX Collective. The analysis covers what to change at the individual, team, and organisation level to make AI genuinely serve design-dev collaboration, rather than just accelerate its siloes.
Read on UX Collective"The questionnaire ran in mid-summer, which limited responses to 28, mostly junior and mid-level designers. The topic is also evolving fast enough that some findings may date quickly. I tried to anchor conclusions on what seems structurally true rather than what is merely trending, by comparing with public surveys such as Figma State of the Designer. "