UX Researcher Partner

An autonomous UX research agent that helps you plan studies, analyze user behavior data, and translate findings into actionable design recommendations that drive measurable product improvements.

When to Use This Agent

Choose UX Researcher Partner when:

• Planning user interviews, surveys, or usability testing sessions
• Analyzing qualitative and quantitative user research data
• Creating research reports with prioritized design recommendations
• Building research repositories to track insights across studies
• Evaluating product changes against user experience metrics

Consider alternatives when:

• You need pure UI design work without research backing (use a design-focused agent)
• Running automated A/B tests without qualitative analysis (use an experimentation platform)
• Doing competitive analysis without user data (use a market research agent)

Quick Start

Copy
# .claude/agents/ux-researcher-partner.yml
name: UX Researcher Partner
model: claude-sonnet-4-20250514
tools:
  - Read
  - Write
  - Bash
  - Glob
  - Grep
prompt: |
  You are a senior UX researcher. Conduct mixed-methods research
  combining qualitative insights with behavioral analytics.
  Always translate findings into actionable recommendations
  with clear priority and expected impact.

Example invocation:

Copy
claude --agent ux-researcher-partner "Analyze our checkout flow
  drop-off data and propose research plan to understand why users
  abandon at the payment step"

Core Concepts

Research Methods Matrix

Method	Best For	Data Type	Time Investment
User Interviews	Deep motivation insights	Qualitative	High
Usability Testing	Interaction problems	Mixed	Medium
Surveys	Broad sentiment patterns	Quantitative	Low
Card Sorting	Information architecture	Mixed	Medium
Diary Studies	Longitudinal behavior	Qualitative	High
Heuristic Evaluation	Quick issue identification	Qualitative	Low
Analytics Review	Behavioral patterns	Quantitative	Low
A/B Testing	Design validation	Quantitative	Medium

Research Plan Structure

Copy
## Research Plan: [Feature/Problem]

### Objectives
1. Primary question to answer
2. Secondary insights needed

### Methodology
- Method: Moderated usability testing
- Participants: 8-10, segmented by [criteria]
- Duration: 45-minute sessions
- Recruiting: [source and screener criteria]

### Success Metrics
- Task completion rate > 85%
- Time-on-task < 2 minutes
- SUS score improvement > 10 points

### Timeline
- Week 1: Recruit and prepare materials
- Week 2: Conduct sessions
- Week 3: Analyze and synthesize
- Week 4: Present findings and recommendations

Insight Prioritization Framework

Impact Score = (Frequency × Severity × Reach) / Implementation Effort

Frequency:  1=Rare, 2=Occasional, 3=Common, 4=Universal
Severity:   1=Minor annoyance, 2=Delays task, 3=Causes errors, 4=Blocks task
Reach:      1=Niche segment, 2=Some users, 3=Most users, 4=All users
Effort:     1=Quick fix, 2=Sprint item, 3=Epic, 4=Major initiative

Configuration

Parameter	Description	Default
research_methods	Preferred methods to suggest	All methods
participant_count	Default sample size recommendation	8-10
report_format	Output format for findings	Markdown
include_personas	Generate persona updates from findings	true
track_insights	Maintain insight repository	true
severity_threshold	Minimum severity to include in reports	2
confidence_level	Required confidence for recommendations	Medium

Best Practices

1. Triangulate your data sources. Never rely on a single research method. Combine behavioral analytics showing what users do with interviews revealing why they do it. When three different data sources point to the same insight, you can recommend changes with high confidence.

2. Recruit participants who match your actual user base. Screener surveys should filter for real usage patterns, not demographics alone. Five well-matched participants catch 85% of usability issues, while twenty mismatched ones may surface none of the problems your real users face.

3. Separate observation from interpretation. During research sessions, record what happened (user clicked the back button three times) separately from why you think it happened (navigation labels were confusing). This discipline prevents confirmation bias from contaminating your raw data.

4. Write recommendations with implementation context. Instead of "improve the checkout flow," specify "add a progress indicator showing step 2 of 4 above the payment form, reducing perceived complexity." Actionable recommendations with clear scope get implemented; vague suggestions get ignored.

5. Build a living research repository. Tag and cross-reference insights across studies so patterns emerge over time. When the same friction point appears in three separate studies over six months, it becomes impossible to deprioritize. Repositories also prevent teams from re-researching solved questions.

Common Issues

Research findings don't lead to product changes. This happens when insights lack business context. Frame every finding in terms of business impact: "Users abandon checkout at 34% rate at the address step, representing approximately $X in monthly lost revenue." Connect user pain to metrics stakeholders care about. Present findings as prioritized opportunities, not problems.

Small sample sizes questioned by stakeholders. Qualitative research follows different statistical rules than quantitative analysis. Five usability test participants reveal approximately 85% of issues (Nielsen/Landauer formula). Prepare stakeholders by explaining the diminishing returns of additional participants for usability studies, while reserving larger samples for survey validation and A/B tests.

Research takes too long relative to development cycles. Build a tiered research approach: guerrilla testing (1-2 days) for quick validation, structured studies (2-3 weeks) for major features, and longitudinal research (ongoing) for strategic direction. Match research depth to decision risk. Low-risk UI tweaks need quick hallway tests, not month-long studies.