Everyday low price agentic AI: Walmart launches “Sparky” shopping assistant

Walmart has launched Sparky, a generative AI-powered shopping assistant embedded in its mobile app that goes beyond traditional chatbots to function as an autonomous agent. The move represents Walmart’s most aggressive bet on agentic AI, positioning the retail giant to transform shopping from a search-driven experience into a service-based interaction while competing directly with Amazon and other retailers racing to deploy AI assistants.

What you should know: Sparky can currently summarize reviews, compare products, suggest items for specific occasions, and answer real-world questions about sports teams and events.

• Upcoming features will include reordering and scheduling services, visual understanding for image and video inputs, and personalized how-to guides linking products with specific tasks like fixing faucets or meal preparation.
• The AI assistant is designed to act autonomously—for a cookout, it can check weather, suggest menus, and help schedule delivery rather than just listing grill options.

The big picture: Walmart is building a comprehensive agent framework that extends beyond customer-facing bots to include Wally for merchants and internal systems for associates.

• This represents a top-to-bottom AI rollout that aims to transform the entire retail operation around AI-driven systems.
• The company is developing its own AI models rather than relying solely on third-party APIs like OpenAI or Google Gemini to ensure accuracy and control over hallucination risks.

What the data shows: Walmart’s “Retail Rewired 2025” report reveals consumers may be more ready for AI-powered shopping than expected.

• 27% of consumers now trust AI for shopping advice, surpassing the 24% who trust social media influencers—marking a clear shift from people to systems for shopping influence.
• 69% of customers cite quick solutions as the top reason they’d use AI in retail, with speed dominating adoption decisions.
• 47% of shoppers would let AI reorder household staples, but only 8% would trust AI to do their full shopping without oversight, while 46% say they’re unlikely to ever fully hand over control.

Why this matters: The retail industry has automated warehouse and logistics operations, but AI agents at the consumer-facing layer represent new territory where Sparky could become the first mainstream proof of concept.

• Unlike recommendation engines that match products to past clicks, Sparky aims to understand intent in context—inferring altitude, weather, travel dates, and baggage limits when someone says they need help packing for a ski trip.
• This requires multimodal AI capabilities including text, image, audio, and video understanding, with plans for features like photographing broken items to get replacement parts and DIY videos.

In plain English: Multimodal AI is like having a shopping assistant that can see, hear, read, and watch videos—not just understand text.

• Instead of typing “I need a replacement hinge,” you could snap a photo of your broken cabinet door and the AI would identify the exact part you need, show you how to install it, and arrange delivery.

Competitive landscape: Amazon, IKEA, and Lowe’s are also racing to launch AI assistants, but Walmart’s approach involves building a complete agent architecture rather than standalone chatbots.

• The company recently partnered with Wing to launch drone delivery in Dallas-Fort Worth, aiming to serve 75% of local customers in under 30 minutes.
• Walmart has simultaneously laid off 1,500 tech and corporate employees, signaling that automation is reshaping team structures as part of broader operational changes.

Potential risks: Greater AI autonomy introduces new challenges around accuracy and safety, with concerns about wrong product recommendations or misreading images for replacement parts.

• Walmart is implementing guardrails including human-in-the-loop confirmations, user approval on sensitive actions, and transparency around data usage.
• The company’s real-world performance with Sparky, not its launch capabilities, will determine whether customers trust it as a permanent shopping fixture.

What they’re saying: CTO Hari Vasudev emphasized the importance of internal AI model development to ensure “accuracy, alignment with retail-specific data, and stricter control over hallucination risks.”