Dive Brief:
- Walmart introduced a customer-facing generative AI assistant named Sparky on Friday to summarize and answer questions about reviews, offer recommendations and help shoppers plan purchases.
- The release follows the March debut of its generative AI assistant for merchants, Wally, and builds on Walmart’s broader plans to integrate AI into its operations and customer experience. Walmart is creating agentic AI tools tailored to retail-specific tasks and company workflows using its internal data and language learning models, the company said recently.
- Sparky, which is available in the Walmart app, will soon expand to include reordering and service booking while understanding text, image, audio and video inputs.
Dive Insight:
Early testing from Walmart found that AI agents work best when deployed for highly specific tasks.
Walmart is training its generative AI shopping assistant with retail-specific large language models to task agents with item comparisons, deep personalization, shopping journey completion and more. Sparky stitches together the results from those tasks to solve complex queries.
Shoppers are becoming more comfortable with AI suggestions. Recent research from Walmart found 27% of respondents prefer AI recommendations, while 24% trust social media influencers.
Customers are using AI to compare prices, shipping times and availability, to receive alerts on price drops and to narrow down options based on historical preferences, according to the report, released last week.
Shoppers aren’t fully surrendering to AI shopping assistants just yet, data from Walmart’s Retail Rewired report found. Almost half of respondents said they were either somewhat unlikely or very unlikely to use an agent to handle an entire shopping trip.
Even as it invests in AI capabilities, Walmart recently cut 1,500 jobs in its U.S. retail and global tech teams. The cuts were done to “remove layers and complexity, speed up decision-making, and help associates innovate rapidly.”