← Back to Work
Analytics AI Integrations

Walmart AI Project

A sentiment analysis kiosk and AI-powered executive dashboard for Walmart's museum in Bentonville, AR. Visitors share their shopping feedback, and leadership gets real-time insight to act on it.

Walmart AI Kiosk

Client

Walmart

Industry

Retail / Enterprise

Location

Bentonville, AR

Project Type

Analytics, AI, Integrations

The Problem

Walmart's museum in Bentonville is where the company tells its story to customers, employees, and anyone curious about how a small-town shop became one of the largest retailers on the planet. At the end of that visit, they had an opportunity: ask visitors about their real Walmart shopping experiences and actually do something with the answers.

But a paper comment card wasn't going to cut it. Walmart wanted to capture something more than what visitors wrote down. They wanted to understand how visitors actually felt about shopping at Walmart, combining what people said with what their faces gave away. And they wanted that data to be immediately useful to the people making decisions at the top.

The ask: build a kiosk that collects visitor feedback on the Walmart shopping experience, then pipe that data into an AI-powered dashboard that executives could query to drive real improvements.

Tech Stack

  • Computer Vision (facial expression analysis)
  • NLP / Sentiment Analysis
  • AI Chat (custom-trained)
  • Kiosk UI

Key Decisions

  • Combined text + facial sentiment scoring
  • Hard-coded prompt library for executive ease-of-use
  • Open-ended AI chat for deeper exploration

What We Built

Two connected systems: a visitor-facing kiosk at the end of the museum experience, and an executive-facing AI dashboard where leadership could dig into what the data actually meant.

Visitor Kiosk

  • Touchscreen interface where visitors answer questions about shopping at Walmart: what they liked, what they didn't, and where there's room to improve
  • Built-in camera captures facial expressions throughout the interaction
  • Computer vision translates those expressions into emotional signals in real time
  • Text responses and facial sentiment are combined into a single unified data record per visitor

Sentiment Analysis Engine

  • Scores each visit across both dimensions: what they said and how they looked saying it
  • Flags discrepancies between stated and expressed sentiment (saying "great" while looking frustrated)
  • Aggregates data across all visitors over time

Executive AI Dashboard

  • Walmart executives access the full dataset through an AI chat interface
  • Hard-coded prompt library for the most common questions, no guesswork required
  • Open-ended chat for deeper, ad hoc analysis. Ask anything, get grounded answers.
  • AI trained specifically on the collected visitor data, not general knowledge

The Result

Walmart went from paper comment cards to a system that captures both the words and the body language behind them, then makes that data queryable by the people who actually need to act on it. Execs can ask the AI a real question and get a real answer grounded in actual visitor feedback on their shopping experience, not assumptions.

It's the kind of feedback loop that most enterprise companies dream about. A touchpoint at the end of every visit that doesn't just collect data, but turns it into something useful immediately.

Let's Work Together

Need to make sense of your data?

We build analytics pipelines, AI tools, and dashboards that turn raw data into decisions. Enterprise complexity, without the enterprise timeline.

Book a Discovery Call Send Us a Message