Walmart last week launched Sparky, a generative AI-powered shopping assistant integrated into its mobile app. The assistant is designed to help users compare products, summarize reviews, and surface suggestions for specific occasions or needs. Sparky is accessible through a smiley emoji icon placed between “My Items” and “Services” within the Walmart app.
The assistant responds to natural language inputs and is capable of handling a range of queries from identifying weather conditions in specific locations to suggesting gift ideas for Father’s Day. It can also retrieve recently searched items, prompt users with contextual questions, and display a limited set of related products. According to Walmart, Sparky will expand over time to include reordering, service scheduling, and support for image, audio, and video inputs.
Walmart’s Broader AI Strategy
Sparky is part of a larger shift underway at Walmart, which includes the development of proprietary AI models tuned for retail. According to CTO Hari Vasudev, using internal models allows the company to improve alignment with retail-specific data and reduce the risk of inaccurate outputs.
Walmart CEO Doug McMillon has previously noted that generative AI has already been used to update over 850 million product catalog entries, a process that, according to him, would have required nearly 100 times the current headcount without automation. This catalog work underpins much of the customer-facing digital experience.
In parallel, Walmart has deployed “Wally,” an internal tool that assists merchants in managing product listings and launching promotions through plain language prompts. The company has also partnered with Wing to expand drone delivery services across the Dallas-Fort Worth area, targeting delivery times under 30 minutes for a significant portion of local customers.
What Customers Are Willing to Automate
Walmart’s “Retail Rewired 2025” report outlines changing customer expectations around AI in shopping. The report notes that 27% of respondents said they trust AI for shopping advice, compared to 24% who trust social media influencers. Additionally, 47% said they would allow AI to reorder common household products. At the same time, only 8% were willing to delegate their full shopping activity to AI without oversight, and 46% said they were unlikely to do so in the future.
Walmart’s stated approach with Sparky reflects this balance. The assistant currently surfaces limited product lists in response to prompts like “What’s for dinner?” or “What do I need for a picnic?” In one instance, when asked about “taco night,” it showed seasoning and taco shell options, followed by a request for additional input. When prompted about weather in a specific location, Sparky displayed multi-day forecasts and UV index data, similar to what a user might expect from a dedicated weather service.
Sparky’s suggestions in some cases have been off-target. For example, when asked about the College World Series, the assistant displayed Major League Baseball merchandise. Walmart includes a disclaimer within the app that Sparky “may make mistakes.”
Integration Across the Organization
Sparky operates alongside other AI-driven efforts across the company. Walmart has incorporated AI into search, product review summarization, and personalized recommendations. Internally, the company is embedding natural language tools for merchant workflows and logistics.
These tools are not isolated. Walmart has described Sparky as part of a broader initiative to “seamlessly weave” AI into the way customers interact with the company. In a blog post, Desiree Gosby, Senior Vice President of Tech Strategy and Emerging Technology, stated that Sparky is “a foundation for what’s next,” and that it represents a step toward a more integrated AI-powered shopping experience.
Walmart’s AI deployments are also occurring alongside internal restructuring. In recent months, the company laid off approximately 1,500 tech and corporate employees. The company has not directly linked these layoffs to its AI initiatives, but they come at a time when automation is playing a larger role in both customer-facing and backend operations.
Current Limitations and Future Plans
As of now, Sparky’s outputs are constrained. Its responses typically include three to four product suggestions and often require the user to refine their query. It does not yet support full end-to-end tasks such as generating complete shopping lists or booking services, though Walmart has indicated that such features are part of its roadmap.
The company has stated that future iterations will include multimodal understanding, allowing customers to upload images or videos for example, a photo of a broken household item to receive relevant product suggestions or how-to content. Walmart has not shared a specific timeline for these updates. Sparky’s rollout arrives at a time when several competitors, including Amazon and Lowe’s, are also deploying generative AI-based assistants.
