Re-Imaginging Retail:
Building an AI Shopping Assistant That Bridges Customer Trust and Business Growth

Focus

  • Mixed-method research

  • Longitudinal/rolling research

  • Behavioral & sentiment measurement

  • Cross-team strategy

Client: Fortune 500 retailer

Role: Manager and strategist

Impact: Meaningful lifts in revenue per visit and conversion, unlocking a significant new revenue opportunity

AI experiences don't ship once. They evolve continuously, and so does the research behind them. This project taught me that research can't just validate an AI experience; it has to keep pace with one.

The Challenge

A national retailer set out to explore how AI could show up in the shopping journey. The initial direction was narrow, focused on a single use case, in a space where most of what worked was still unknown. My job was to widen that lens before the team built too far in one direction.

What I Did:

  • Audited existing customer feedback and engaged people already using AI tools in their daily lives to find where the real retail opportunity lived

  • Reframed the initial direction into a broader vision: an assistant that helps customers compare, decide, and trust what they're seeing

  • Established a rolling research cadence — recognizing that a probabilistic, constantly-shifting experience couldn't be validated with a single round of testing

  • Built out a framework for evaluating AI-specific UX patterns (onboarding, input, output interpretation, feedback, and system learning) rather than testing features in isolation

  • Partnered with engineering to define customer intents and train internal models to classify conversations against them

  • Designed and piloted new methods for capturing sentiment directly within the experience, since traditional transcript analysis wasn't robust enough on its own

What Happened Next:

  • Insights from this work shaped product capabilities beyond what was originally scoped, strengthening both the retailer's experience and the broader platform

  • The rolling research model became the foundation for how the team continues to evaluate and evolve the assistant today

  • The work delivered a measurable lift in conversion and revenue per visit, translating into a significant new revenue opportunity for the business

Want to go deeper?

This case study includes proprietary business details and is protected for privacy. Reach out to request access.