Case studyWalmart
A more intelligent purchasing experience
LastMile AI: preserving order value when items are out of stock.
- Built for
- Walmart
- Practice
- AI Products & Services
- Built by
- Velocity by CourtAvenue

In brief
Velocity by CourtAvenue built LastMile AI for Walmart to fix the substitution moment in online orders. When a requested item is unavailable, LastMile AI recommends the best alternative from customer preferences, confirms it with the customer in a natural language conversation and sends the approved choice back to the shopper in real time, protecting order value and satisfaction.
Walmart · The challenge
Reimagining the substitution experience.
Shoppers are not product experts.
When an item is unavailable, shoppers are often left guessing which alternative best matches the customer's needs and preferences.
Communication delays disrupt the experience.
Customers may miss messages, respond late or be unavailable while shopping decisions need to be made in real time.
Language and communication barriers create friction.
Differences in language, context and communication style can make it hard to quickly identify the right replacement product.
Substitution failures reduce satisfaction and order value.
Without an effective way to identify acceptable alternatives, customers receive less than they wanted and retailers lose revenue.
Walmart · What we did
LastMile AI: preserving order value at scale.
Identify and recommend.
Built an intelligence layer that understands customer preferences and recommends the best alternative when requested items are unavailable.
Communicate and confirm.
Automated customer communication through natural language conversations, helping shoppers and customers quickly agree on acceptable substitutions.
Update and fulfill.
Delivered approved substitutions back to shoppers in real time, reducing delays and keeping orders moving toward completion.
Protect order value.
Preserved revenue by reducing abandoned substitutions and helping customers find acceptable alternatives.
Walmart · The output
A more intelligent purchasing experience.
Intelligent substitution engine.
An AI recommendation engine that understands customer preferences, shopping behavior and product relationships to recommend the best available alternatives.
Automated customer engagement.
A conversational experience that communicates substitutions, gathers feedback and helps customers quickly approve or refine recommendations.
Continuous learning platform.
Captures preferences, substitution outcomes and decision patterns to keep improving recommendations and fulfillment performance.
Building yours
How we'd build it for you.
Map the substitution moment.
Find where unavailable items stall orders today and what your customers accept as a good replacement.
Connect preferences and catalog.
Give the recommendation layer your product relationships and each customer's preference history.
Put the conversation in the loop.
Let customers approve or refine a substitution in plain language while the order is still being picked.
Learn from every outcome.
Feed accepted and rejected substitutions back so recommendations improve with every order.
