Walmart — AI Assistant Sparky Increases Average Order Value by 35%

Company

Walmart — the world’s largest retailer.

Problem

Traditional e-commerce search was keyword-based and hierarchical rather than goal-oriented. Customers struggled to plan complex, multi-item shopping trips (e.g. “everything needed for camping”).

Solution

They built an AI shopping assistant named Sparky that uses Anthropic Claude and OpenAI models for natural language understanding and intent detection. It integrates directly with merchandising, inventory, and fulfillment systems.

Results

Customers using Sparky show 35% higher average order value (AOV). 50% of mobile app users are now engaging with the new AI assistant.

Source

• https://www.walmart.com/cp/sparky/5291783
• https://www.retaildive.com/news/walmart-launches-generative-ai-assistant-sparky/750300/

Replicable with OpenClaw?

Yes, with substantial retail domain adaptation. Requires intent-understanding agents, real-time inventory checking, and basket optimization logic.