Dive Brief:
- Walmart is beta testing a generative AI-powered shopping assistant in its mobile app, the company announced last week. The assistant will show up for select customers in categories such as home products and electronics.
- Shoppers can ask for advice using natural language phrases like “Which TV is best for watching soccer?” Walmart plans to add more advanced features over time, such as the ability for the assistant to respond to queries with follow-up questions for clarification.
- The average U.S. consumer spends six hours per week searching for items, but generative AI holds the potential to simplify the shopping journey, Anshu Bhardwaj, SVP and COO at Walmart Global Technology and Walmart Commerce Technologies, previously told CX Dive.
Dive Insight:
Walmart’s assistant can help customers find just the right product in its large digital selection, but experts note that any generative AI rollout needs careful maintenance to keep data safe and results relevant.
The retailer began adding generative AI-powered enhancements in late 2023 with an AI-powered search function for the Walmart app, letting shoppers seek out items using natural language queries.
One of the key benefits of generative AI assistants is narrowing down purchases to exactly what a specific customer wants and needs, according to Jill Standish, senior managing director and global lead of retail at Accenture.
“Think of it as a ‘sales advisor’ that is present throughout the purchasing journey — providing recommendations based on what a customer cares most about in the moment from pricing and nutritional information, to whether an item has been ethically sourced,” Standish told CX Dive in an email.
For retailers with large sites, and an array of content and products, generative AI can improve the customer experience and journey in ways that are not possible with traditional site navigation, according to Brad Jashinsky, director analyst at Gartner.
However, no technology is perfect. The technology is still in its early days, and many consumers don’t trust companies with their data.
Poorly trained generative AI models can offer bad suggestions that may cause customers to switch retailers, Jashinsky told CIO Dive via email.
“To add value for customers, GenAI models will require both robust training to ensure that results are appropriate to searches and continual testing to be sure positive performance is maintained,” Jashinsky said. “Retailers should also allow customers to easily turn off generative AI assistants if they prefer the traditional search experience.”
