Electrolux · Research practice

Scaling AI in UX research

Shared AI workflows for research teams across 4 business units.

Overview

Researchers across 4 business units used AI in different ways, so analysis quality depended on who ran the study. I built a shared practice where AI drafts and researchers decide.

AI tool use varied by researcher. Researchers converge on shared prompts and skills. AI drafts transcripts, tags, and summaries; researchers verify, interpret, prioritize and synthesize, and tell the story as an actionable insight. Prompts and skills are reviewed each quarter.
Role
Lead UX Researcher, Research Practice
Timeline
2025–present
Teams
UX Research across four business units, Global UX

Context

AI tools spread faster than shared methods. The goal: faster analysis, with methods stakeholders can trust.

Constraints

  • Tools6+ tools, with no central repository.
  • Time and resourcesBuilt alongside active research projects, without dedicated time.
  • Pace of changeAI tools changed every few months, so the practice had to outlast any single tool.

How I approached it

Five decisions shaped the practice:

  1. Focus AI on repetitive tasksInterpretation and cultural analysis stay human-led.
  2. Engineer the contextShared templates (context, role, data, format) keep outputs consistent.
  3. Keep templates livingPrompts and playbooks are reviewed each quarter as tools and skills change.
  4. Disclose AI useEach report notes AI’s role, so stakeholders can trust the findings.
  5. Reinvest time in storytellingTime saved goes into synthesis and storytelling.

In practice

The standards and tools the team uses.

Research strategy: practice audit, framework design, and pilot testing, with three principles for AI research and an AI-assisted research workflow
Research strategy: principles and workflow · Low resolution for confidentiality.
How the framework was operationalized: a decision tree for when to use AI, tool-specific playbooks, and a prompt template library
Decision tree and prompt template library · Low resolution for confidentiality.

Impact to date

An evolving practice, updated as tools and skills change. Reach out to learn more.

  • Speed~40% faster analysis.
  • AdoptionFour business units, with 12 researchers trained.
  • VisibilityPresented at IDSA IDC 2026 and World Usability Congress 2025.