Walmart said customers using Sparky, its AI shopping assistant, had average order values about 35% higher than customers who did not use it. CEO John Furner reiterated the comparison on the company's May 21, 2026 earnings call.
That is a comparison between customer groups. It does not, by itself, establish that Sparky caused the difference or that another retailer should expect the same result. For merchants considering an assistant, the useful question is whether it helps customers assemble a basket that meets their needs.
Walmart describes Sparky as an assistant in its app that helps people compare products, build lists and plan for occasions, using product information and reviews. Those tasks offer a concrete starting point for a retail pilot: helping someone buy the items needed for a particular activity.
Consider an illustrative request to set a dinner table for eight people. The customer already owns eight dinner plates and needs glasses, napkins and a tablecloth for a specified table size. A useful assistant would respect what the customer owns, check pack quantities and help them choose a tablecloth that fits. Recommending another dinnerware set would increase the basket while ignoring the request.
The catalog needs to answer those practical questions. A glassware listing should distinguish a single glass from a set and state how many pieces are included. A tablecloth should give its actual dimensions and shape. Product images can show the setting, but the text should make clear which photographed accessories are sold separately.
Fulfillment also matters to a basket assembled around an occasion. Ask whether all the selected items can reach the customer's location before the event. If an item is unavailable, a replacement should preserve the important requirement, such as enough glasses for eight guests. The shopper should see any changed price, quantity or delivery promise before accepting it.
A retailer can use this dinner-table task to review an assistant before a wider rollout. Give the same stated needs to a merchandiser and check the proposed basket against the catalog. Look for duplicate items, incorrect pack sizes, products that do not fit and delivery dates that miss the occasion. Keep the shopper's budget visible throughout the review.
Judge the eventual orders alongside returns, cancellations and support questions. A larger basket is valuable when the customer receives useful products and the order remains profitable. Walmart's reported result is a reason to investigate that possibility, not a revenue forecast for an untested assistant.
Common questions
Did Walmart say Sparky increased every customer's spending by 35%?
No. Walmart reported average order values about 35% higher among Sparky users than nonusers. The statement does not establish an individual customer's increase or isolate the assistant's effect.
What should a retailer check before offering basket recommendations?
Check quantities, compatibility, the items the shopper already owns, total cost and delivery timing. A complete basket should satisfy the stated task without unnecessary additions.
Which result matters beyond order value?
Whether the customer receives the right items and keeps them. Review cancellations, returns and support issues alongside sales and the cost of serving the order.