New Way Now: Morrisons serves up stress-free grocery shopping with the help of Gemini

*Featured in this video:* Peter Laflin, Director of Data Analytics and AI at Morrisons *Executive summary:* Morrisons, one of the UK’s largest supermarkets, is serving up faster, more convenient shopping experiences to millions with Google Cloud. Using BigQuery, Vertex AI, and Google’s Gemini models, Morrisons developed Product Finder — an in-app AI tool that helps customers instantly find the items they are looking for in stores. Now, shoppers can simply type in a product name and receive real-time information about its precise aisle and location, even if the shelves have been rearranged that day. The Product Finder is already becoming a go-to helper in Morrisons stores: Customers racked up 50,000 hits a day over Easter 2025 alone, with usage growing steadily as shoppers continue using it for return trips. *Challenge:* The quest for specific items — particularly smaller, specialty products like tahini or cranberry sauce — can quickly turn a routine trip to the grocery store into a time-consuming search. To address this common frustration, Morrisons wanted to bring the same intuitive, effortless search experiences shoppers enjoy online, directly into its aisles. *Solution:* Morrisons had already partnered with Google Cloud to migrate all of its data to BigQuery and adopted Looker to enable real-time visibility across store transactions, inventory, logistics, product information, and customer feedback. Having this strong data foundation enabled Morrisons to embrace Google Cloud AI technologies to create the Product Finder — an in-app AI tool that helps customers quickly find the products they’re looking for, right from their phones. Through Vertex AI, Morrisons harnessed the advanced capabilities of Google’s multimodal Gemini models to interpret natural language queries and translate them into corresponding product IDs. This information is then used to retrieve the exact product locations from BigQuery and deliver it back to customers. Crucially, BigQuery’s tigh