Walmart's Sparky Is Now in ChatGPT and Gemini — Your Listing Data Decides If You Show Up

In Walmart's most recent earnings call, one product got more airtime than almost anything else: Sparky, the company's AI shopping assistant. The numbers explain why. Weekly active users more than doubled in a single quarter, units purchased through Sparky grew more than fourfold, and customers who use Sparky have an average order value about 35 percent higher than those who don't. Walmart also said its investments lifted Sparky's intelligence and response quality by around 40 percent this year, and it now works in stores, reorders repeat items automatically, and even speaks Spanish.

If you sell on Walmart Marketplace, those figures aren't a tech-press curiosity. They describe a new layer sitting between your product and the shopper — one that's deciding what gets recommended. The question every brand should be asking is whether their listings are optimized for that layer, or invisible to it.

Sparky doesn't shop like a keyword engine

Traditional Walmart search matches the words a shopper types against the words in your listing. Sparky works differently. Ask it "what's the best laptop for an art student" and it derives the intent and context, then returns a curated set of items with a short explanation of why each one fits that specific situation. Walmart's own research team has published work showing the AI was trained on millions of query-item pairs and reaches agreement with human evaluators on product relevance most of the time — meaning it judges whether your product actually meets a need, not just whether it contains a keyword.

That shift changes what wins. A keyword-stuffed title that ranked fine in old Walmart search can be passed over by Sparky if the listing doesn't give the AI enough structured information to understand who the product is for and why it matters.

The multiplier most sellers haven't registered

Here's the part that raises the stakes dramatically. Walmart has embedded Sparky into ChatGPT and Gemini. That means your Walmart.com product listings are now the source catalog for shopping recommendations across multiple platforms, reaching hundreds of millions of additional users. The same listing data either earns you recommendations across all of those surfaces — or, if it's incomplete, makes you invisible on several platforms instead of just one.

The logic is the same everywhere: Sparky on Walmart, Rufus on Amazon, and the AI assistants inside ChatGPT, Gemini, and Perplexity all match product data against shopper intent. The brand that provides the clearest, most complete structured data wins the recommendation. The brand with vague, generic listings doesn't underperform — it simply doesn't appear.

What Sparky actually needs from your listings

The optimization checklist is consistent and concrete. Fill in every available attribute field, because Sparky uses structured attributes for matching across all the platforms it reaches. Write for conversational queries — think about how a shopper would ask for your product in natural language ("a lightweight carry-on bag under $100"), not in catalog-speak. Build a clear value story, since Sparky favors products that communicate why they're worth the price. And keep listings complete and current as Walmart's data spec evolves, because incomplete or outdated listings quietly fall out of recommendations.

None of this is hard for a single product. Across thousands of SKUs, with new products arriving incomplete and Walmart's requirements changing, it becomes a continuous data operation that no manual team can keep up with.

Where SYNAPS comes in

This is exactly the gap SYNAPS closes. From minimal input — SKUs, EANs, titles — it enriches every product with the missing attributes Sparky relies on, structures the data into the dedicated fields the AI reads, validates each listing against Walmart's spec before submission, and monitors the catalog daily so nothing goes stale. The output is built to be dual-purpose: ready to publish on Walmart Marketplace and readable by the AI assistants — Sparky, and through it ChatGPT and Gemini — that increasingly decide what shoppers see.

The Sparky data is unambiguous: the best customers are already shopping through AI, and they spend more. The only open question is whether your listings are structured well enough to be the products it recommends.

Want to see whether your Walmart catalog is Sparky-ready? Book a 30-minute demo at cal.com/hugo-benloulou/30min.