Trust and Explainable AI (XAI) in E-Commerce
Indian fashion shoppers abandon carts at high rates even when they like the product. The question wasn't "how do we show recommendations", it was "why don't users trust what the algorithm is showing them, and what would actually change that?"
Core Problem
Indian fashion shoppers were distrusting AI recommendations even when the recommendations were accurate, leading to cart abandonment that had nothing to do with product fit
Role & Scope
UX Researcher & Interaction Designer — research design, user interviews, behavioral observation, prototype testing, and synthesis across 3 major e-commerce platforms
The Challenge
Trust is invisible until it breaks. Isolating which specific design patterns were eroding it, versus price, brand, or context factors required careful research design and ruthless synthesis
Other projects
DRAS Assessment Portal for MSMEs
How do you digitize a government evaluation process when your users have never filled a form online and the entire system runs on WhatsApp forwards and printed sheets?
Logsy: Tracking Growth Through Play
Parents want to track their child's development. Children want to play. How do you design one product that genuinely serves both, without turning learning into a chore or play into a test?
A Predictive AI & Automation Ecosystem Design
Hyderabad drivers waste 20+ minutes per trip searching for parking. When you build an AI that predicts availability, the real design problem isn't the algorithm, it's whether users will trust a prediction they can't verify.
Reimagining a School’s Digital Identity
A school's website is its first handshake with an anxious parent. When navigation is built for internal logic instead of parent intent, the result is a beautiful site that nobody converts on.