School Project

2026

Trendly: AI Powered Shopping Assistant

An AI-powered shopping assistant for a mobile e-commerce platform, simplifying complex cross-site comparisons and improving the decision-making workflow for fashion-conscious users.

Role

UI/UX Designer

Timeline

Feb 2026 - Mar 2026

Team

Esha Macha

Platform

Shopping Application

Role

UI/UX Designer

Timeline

Feb 2026 - Mar 2026

Team

Esha Macha

Platform

Shopping Application

Overview

This case study focuses on my design work for an academic project within the GXD program at NC State. The project challenged me to develop a comprehensive user experience and interface for AI powered personal shopping assistant application, guiding the product from initial concept to a final interactive prototype.

During this project, I executed an end-to-end design thinking process spanning empathy, definition, ideation, testing, and prototyping. This case study highlights my journey in creating an intuitive mobile shopping experience. Through user research, journey mapping, and iterative wireframing, I identified opportunities to streamline the browsing and checkout workflows, reduce navigational complexity, and support a more engaging e-commerce experience for users.

Problem

Summary

Through my user research, I discovered that the process of finding and comparing similar clothing styles across multiple e-commerce websites was disjointed. The challenge was to design an integrated AI feature to eliminate manual cross-checking, reduce friction, and help shoppers discover their desired items more efficiently.

My Role

My role was to conduct user interviews and consolidate a specific problem with online shopping that can be optimized through an AI shopping feature. My work spanned conducting user research, developing journey maps, creating user flows and low-fidelity wireframes, facilitating usability testing, and designing the final high-fidelity interactive prototype.

Research

User Interviews

To understand the current friction points in the e-commerce experience and evaluate the potential impact of an AI-driven discovery tool, I conducted interviews with two frequent online shoppers. I inquired about the specific hurdles they face during their shopping journey and how an integrated AI feature could alleviate those frustrations.

User Models

Based on our initial research and user interviews, I synthesized our findings into two primary user models. These archetypes map the varying motivations, behaviors, and priorities of our target audience when navigating online fashion e-commerce.

Define

User Story

Moving into the define phase, I narrowed the project scope to focus on our primary user archetype: The Fashion Designer. To ensure the ensuing design decisions and feature developments remained grounded in this user's core needs, I established the following guiding user story:

"As a Fashion Designer, I want AI suggestions based on my inputted preferences by cross-referencing products across multiple websites, to have a variety of tailored options to select from .”

Following the research process, I decided to select one user to continue with and developed a user story:

"As a Fashion Designer, I want AI suggestions based on my inputted preferences by cross-referencing products across multiple websites, to have a variety of tailored options to select from .”

Following the research process, I decided to select one user to continue with and developed a user story:

"As a Fashion Designer, I want AI suggestions based on my inputted preferences by cross-referencing products across multiple websites, to have a variety of tailored options to select from .”

User Journey Map

To visualize the emotional highs and lows of our primary persona, I mapped out their typical e-commerce journey. This exercise helped pinpoint exactly where friction occurs and where design interventions could have the highest impact.

User Flow

To define the architecture of the AI shopping assistant feature, I mapped out the primary user flow. This flowchart illustrates the path a user takes to find, refine, and purchase an item using the tool, eliminating the need for manual cross-referencing. By mapping this flow, I ensured that the AI feature acts as a natural extension of the browsing experience.

Design Process

Low Fidelity Wireframes

To quickly visualize the core user journey and test the placement of the new AI feature, I sketched a series of low-fidelity wireframes. My primary focus during this stage was establishing an intuitive navigational structure and ensuring the AI integration felt like a natural extension of a traditional e-commerce interface, rather than an intrusive add-on.

Mid Fidelity Wireframes

Building upon the initial low-fidelity sketches, I developed mid-fidelity wireframes to establish the visual hierarchy, component layout, and step-by-step logic of the AI discovery feature. This phase was critical for mapping out exactly how the conversational AI interacts with traditional e-commerce browsing.

Usability Testing

To validate the design and identify areas for improvement, I conducted usability testing sessions. The feedback was overwhelmingly positive, yielding an average usability rating of 4.5/5. Every participant indicated they would use the application again.

The main improvement I implemented following my usability testing was enhancing the loading screens to communicate data processing more effectively. I transitioned from utilizing static conversational text bubbles to incorporating a clear visual loading spinner paired with explicit status updates

Final Designs

The final designs were created using Figma. The visual design prioritizes a clean, modern aesthetic that allows the fashion pieces to remain the focal point, while ensuring the powerful AI tools feel intuitive and accessible rather than overwhelming.

Contextual AI Discovery: The "Discover alternatives with AI" feature is embedded directly onto the product detail pages. Rather than navigating away, users open a clean modal overlay to input their preferences

Conversational Iteration: The dedicated AI Assistant screen utilizes a familiar chat-style interface. This allows users to naturally refine their generated results and update their feed without losing their previous context or having to restart the search.

Unified Wishlist & Checkout: To solve the core issue of tab fatigue, this design centralizes saved items from various external retailers into a single wishlist. The checkout screen calculates the grand total across all websites and uses user auto-fill data to eliminate friction before final purchase.

Usability Testing

Final Designs

ey features of the redesign include a streamlined information architecture, user friendly navigation bar, a AI guided narrative-writing experience, an integrated review and submission process, and an AI review agent to aid the investigator. Together, these improvements reduce manual effort, improve discoverability, and help investigators create accurate, high-quality SARs with greater efficiency and confidence.

Mid Fidelity Wireframes

Lessons

Final Thoughts

My experience developing this AI-powered personal shopping assistant, teaching me the balance between advanced technical functionality and intuitive usability. I learned firsthand how even small design choices, like clear system feedback during loading states, can significantly influence user trust and behavior. Throughout the process, I was able to expand my UX design, research, and prototyping skills while gaining practical experience in rapid iteration and human-centered problem-solving.

It was a rewarding challenge to tackle a real-world problem and build a comprehensive solution from the ground up. Learning directly from usability testing participants was especially impactful; their constant feedback helped me refine the interface, grow professionally as a designer, and acquire new perspectives on seamlessly integrating AI into everyday digital routines.

My experience developing this AI powered personal shopping assistant, teaching me the balance between advanced technical functionality and intuitive usability. I learned firsthand how even small design choices, like clear system feedback during loading states, can significantly influence user trust and behavior. Throughout the process, I was able to expand my UX design, research, and prototyping skills while gaining practical experience in rapid iteration and human-centered problem-solving.