Go back
Walmart AI Chat Assistant
Project Overview
Walmart – My Assistant is an internal AI-powered chatbot designed to help store associates complete daily tasks more efficiently, from checking product availability to finalizing customer orders. The goal of this project was to redesign the assistant experience to be more intuitive, accessible, and aligned with real-world workflows inside Walmart stores. Through stakeholder interviews, user journey mapping, and iterative prototyping, I helped identify friction points in the current system and led the design of a more seamless, voice-and-text-enabled interface. A key focus was supporting offline order finalization, enabling associates to complete purchases even when systems were down, saving time and reducing dependency on manual processes.
The Challenge
When I joined Walmart, the field teams were struggling. Store associates had to juggle multiple tools and workflows just to get basic answers or complete everyday tasks. It was slowing them down, leading to miscommunication and delays. The vision? Build My Assistant, an AI-powered, conversational experience that lives inside the Me@Campus app. The goal was to give associates a single point of access to ask questions, solve problems, and move fast. But it had to feel human, seamless, and efficient, just like chatting with a knowledgeable teammate.
My Role
I led the UX design for My Assistant, owning everything from initial research and concept sketches to high-fidelity designs and final handoff. I partnered with researchers, engineers, and AI teams to ensure the experience felt smart, helpful, and easy to use.
This project lived at the intersection of AI, enterprise UX, and mobile product design, which meant everything had to be intentional and scalable.
Discovery & Research
We kicked off by listening. I conducted interviews with store associates, leads, and support teams to understand how they currently get help. The friction points were loud and clear:
Too many systems with inconsistent answers.
Simple questions took too long to resolve.
New hires had no idea where to begin.
We mapped these insights into key user journeys, like looking up PTO policies, resolving device issues, or understanding task assignments, and started shaping how an assistant could guide them better.
Designing the Assistant
This wasn’t about building a chatbot for the sake of it, it had to be truly useful.
I collaborated with our AI team to design conversational flows for different scenarios. I defined:
The tone of voice (friendly but professional)
Message types (text, buttons, quick actions)
Edge cases and fallback behavior
We focused heavily on mobile-first design, knowing that most users were on-the-go. I designed scalable UI components for greeting messages, follow-ups, and feedback, making sure everything worked within Walmart’s existing design system.
Solution
The design solution prioritized clarity, speed, and context-aware interactions to support high-volume retail operations. We introduced A simplified mobile interface for in-aisle task execution. A scalable web dashboard for managing bulk operations. Offline order finalization flows to help associates complete transactions even during system interruptions. A conversational UI tuned for both voice and text input, enhancing flexibility and ease of use. This holistic redesign ensured greater productivity and smoother user experience across devices.
Testing & Iteration
Once we had working prototypes, we tested them with real users.
We ran multiple usability sessions and A/B tests to find what clicked. One key insight? Associates wanted confirmation and confidence, not just smart answers. We added typing indicators, feedback prompts, and clear fallback responses to help users trust the system.
The result? A more transparent, guided, and human-feeling experience.
Impact
The pilot launch spoke for itself:
43% reduction in time-to-answer for frequent questions
2.6x increase in usage compared to the previous help system
Consistently positive feedback around usability and tone
It’s now being rolled out across more locations, with new capabilities being layered in (like deeper integration into task management and policy lookups).
What I Learned
This project reminded me that AI is only as good as the UX that surrounds it. Designing My Assistant taught me how to create clarity in uncertainty, and how to translate natural human behavior into structured, machine-driven interactions.
It also deepened my passion for working on products that make real people’s lives easier—especially in fast-paced environments like retail.
