Transforming customer feedback into a product performance dashboard
Executive Summary
Root offers car insurance through a digital app, using AI models to calculate rates based on driving behavior.
Our research team struggled to assess customer data across scattered data sources. I redesigned UX research workflows and used AI-assisted coding to build a centralized dashboard prototype. Finally, I also synthesized every user research workflow into a package of AI skills that automated internal processes and provided actionable feedback for product managers.
- UX Research Workflows Automated
- +40
- Usability Issues Discovered
- +18
Problem
Previous user research contained valuable insights, but it was difficult for product teams to take action because:
1- Most data collection occurred outside the app experience
2- Usability testing sessions did not follow a standardized approach
3- Qualitative data on customer experience was often overlooked
Research
Working closely with the SVP of product, I conducted discovery interviews with Product Managers, Software Engineers, and other departments.
I mapped which research insight mattered most to stakeholders. Their organizational priorities shaped the information hierarchy and design choices after multiple rounds of rapid iterative prototyping.
Round 1
Because Root did not have a design system, my process started with building the variables, components, and layout constraints in Figma. I then used AI-assisted coding to turn the design into a minimum viable prototype for stakeholder feedback.
Round 2
Across several rounds of usability testing, I asked stakeholders to find information, explain its meaning, and describe how they would use it. I compared their responses with my design intent.
Round 5
For handoff, I delivered the dashboard prototype in code and created reusable AI skills for data preparation, analysis, and visualization, which allowed project managers to analyze product data autonomously.
Results
The final dashboard brought product metrics and customer feedback into one place, with a working interface and automated data analysis workflows.
- UX Research Workflows Automated
- +40
- Usability Issues Discovered
- +18
Reflection
The challenge was to decide (1) which research insights mattered, and (2) how to gather customer data from different sources into a single interface. For that, rapid iterative prototyping was the easiest way to align stakeholder expectations.
And thanks to my experience working with UX research, design, and coding, I was able to take ownership of every product development stage, shipping multiple high-fidelity prototype variants within a week.
My final lesson is that involving AI in my workflow has allowed me to collect more user feedback and move faster, drastically improving the reliability of my design choices.
Credits
- UX Research
- Anna Nguyen, Hala Daher, Julie Harrison, Layilah Campbell
- Executive Leadership
- Jill Kellet
- Product Design
- Klew Still, Brooke Webster, Allison Chen, Jason Jeong, Ryan Farham
- Creative Direction
- Wes Kull
- Product Management
- Brandon Hill