Identifying Generative AI's Impact on Brands
Role
Lead UX Designer, UX Researcher
Team
Megan Chen, Lucky Sharma, Evan Zhao
Duration
August 2025- Present
Project Overview
In a 4-month school project at Georgia Tech, we were sponsored by Slalom a human business and consulting company that helps people and businesses unlock the full potential of AI.
PROBLEM OVERVIEW
AI is transforming how consumers shop, creating new challenges for brands to stay visible and relevant in AI-driven recommendations.
Agentic AI is transforming online shopping by autonomously guiding consumers through complex purchasing journeys with greater personalization and efficiency. As shoppers increasingly rely on AI to discover products, brands must adapt their marketing and product strategies to remain visible within AI-driven recommendation systems or risk becoming overlooked. Our partnership with Slalom Consulting explores how consumers interact with AI while shopping and how retailers can strategically position their products for this emerging landscape.
MARKET SITUATION
Shoppers are consulting Gen AI for help with their shopping process instead of browsing retailer websites.
WHY THIS MATTERS TO BRANDS
Brands lose direct visibility and control over how they’re represented, since AI systems summarize or filter their information. If they are not represented in new forms of AI shopping tools, they would lose business.
OPPORTUNITY
To build a product that will help brands remain discoverable, trustworthy, and fairly represented in AI-driven shopping environments.
PROBLEM STATEMENT
PROCESS
WHO ARE THE USERS?
Shoppers
JTBD: Discover a broader range of relevant brands that fit their personal needs, increase relevance of AI recommendations on shopping.
Brands
JTBD: Ensures they are represented by AI, focus on creating products that genuinely satisfies customers and lead to good reviews, reach more shoppers
RESEARCH METHODS
Investigating our research questions & problem space
Through conversations with our sponsor, Slalom Consulting, and exploration of the problem space, we refined our focus into three key research questions centered on how shoppers are using AI in their purchasing journeys. Over the course of a month, we conducted a mixed-methods study incorporating both qualitative and quantitative research to investigate these questions, ultimately aiming to build a comprehensive understanding of consumer behaviors and expectations in an AI-driven retail environment.
Surveys
Surveyed 90+ shoppers in their 20's to understand to what extent they utilize AI in their shopping process
We distributed a survey to collect quantitative data from our target user group and identify patterns in AI usage across different shopper segments. The survey explored users’ willingness to adopt AI shopping agents, their comfort with personalized recommendations, and their motivations for using or avoiding generative AI during purchase decisions. This method allowed us to quickly capture broad trends in AI-enabled shopping behaviors, while the standardized questions ensured consistent data for clear comparisons across demographics and usage patterns.
Semi-Structured Interviews
Interviewed 8 shoppers who used AI before to understand how AI shapes shoppers' perceptions of brands
We conducted semi-structured interviews with target users who have previously used AI for shopping, aiming to understand the underlying reasons behind the survey trends. These conversations allowed us to dive deeper into individual experiences, uncovering how AI tools influence brand perceptions, loyalty, trust, and frustrations. The interviews provided rich qualitative insights that helped us clarify survey responses and better understand why users trust certain recommendations while disregarding others.
Contextual Inquiries
Conducted 9 sessions to observe how shoppers interact with AI
We conducted contextual inquiries to gather in-depth qualitative data on shoppers’ real-time processes, reactions, and perspectives when using AI to support purchase decisions—whether they were experienced or first-time users. This approach allowed us to observe authentic interactions with AI tools while probing deeper into participants’ thoughts, emotions, and preferences. By engaging a diverse set of users, we gained insight into the different roles AI plays throughout the shopping journey and how these experiences shape decision-making.
Competitive & Comparative Analysis
Compared most frequently used AI tools to identify differences in the way they present brand information and product recommendations
We conducted a competitive and comparative analysis to examine how existing generative AI shopping platforms present brand information and product recommendations. Our goal was to identify specific design features that make these tools effective for shopping purposes. We focused on the most frequently used AI platforms—ChatGPT, Gemini—along with Google Search and Google News, selecting features based on our research questions. Survey data indicated that users rely on different platforms for online shopping, so we analyzed the features each offers to understand these preferences. This analysis helped us explore the problem space in depth and informed which features to prioritize in our app to enhance the user experience while minimizing unnecessary elements.
RESEARCH FINDINGS
Shoppers use AI primarily to save time and streamline product research but adoption depends on trust, accuracy, transparency, and clear information presentation.
USER NEEDS & DESIGN IMPLICATIONS
Synethesizing our findings to guide design phase
By synthesizing and prioritizing our research, we identified 11 key user needs and mapped them to 13 design implications to directly address user pain points. These findings provide clear guidance on which features to prioritize and which to avoid as we move into the design phase.
PERSONA
Meet Fred
Creating this persona helped us better understand our users’ goals, pain points, and needs, enabling us to empathize with his experience and design solutions that align with his motivations, behaviors, and expectations throughout the shopping journey.
JOURNEY MAP
Visualizing the shopping process from discovery to purchase
To understand Fred's experience, we developed a journey map that divides his shopping process into stages, from product discovery to purchase, highlighting moments where he may encounter confusion, decision fatigue, or a lack of trust in the tools he uses.
IDEATION
Initial Brainstorming and Sketches
From our research findings, we developed three design concepts that addressed our design implications from our research. These concepts ranged from a mobile app to a browser extension to a full web application.
Design Idea #1: Mobile App
An AI-powered mobile app that assists users in the browsing and exploring stage of shopping online.
Design Implications
1. Design should enable users to give our website information that makes recommendations more confident (eg. images, links, profiles of what they like).
2. Design should suggest products from a variety of brands, making the recommendations relevant based on information users provide.
3. Design should enable users to see images of products recommended.
Design Idea #2: Broswer Extension
An AI chatbot extension that can assist users while they’re shopping for products on other websites.
Design Implications
1. Design should support users in understanding complex information quickly using formatting or visuals that reduce cognitive load.
2. Design should enable users to compare granular information about different products/brands.
3. Design should remember past preferences and styles, including what they click on and purchase.
Design Idea #3: Web Application
An AI-powered website that assist users throughout their whole shopping process from initial product discovery and detailed product comparison all the way through to the final purchase
Design Implications
1. Design should enable users to customize if they want our site to do research for them, or make the final decision.
2. Design should enable users to compare granular information about different products/brands.
3. Design should suggest products from a variety of brands, making the recommendations relevant based on information users provide.
4. Design should learn from user feedback about which responses the user finds to be more accurate based on what they are asking for.
We conducted peer review sessions to evaluate our three designs where a set of standard questions were asked to evaluate interest in the designs. From these sessions, we chose the design concept that participants liked the most: Web Application.
Wireframes
We created wireframes across four key flows to address the design requirements and incorporate the user feedback we received from our sketches.
Onboarding
- Ensure more clarity/guidance about how AI can assist in the process
- Gain explicit permission about what the AI tracks
- Account for all genders in the onboarding process
- Allow users to upload a photo of themselves for more specific recommendations, if they want
Chat Page
- Implement voice controls, including both speech to text and image
Profile Page
- Allow users to save and compare products they like but don't want to purchase immediately (wishlist)
For You Page
- Include filters for finding specific products
This project is currently ongoing.
This page will be continuously updated.
@Jessica Chu 2025
made with love 
and lots of sunsets 