Let AI do the hard parts of your holiday shopping (blog.google)

Google announced a major holiday shopping update that layers its Gemini models, agentic capabilities and the Shopping Graph into Search and the Gemini app to automate discovery, inventory checks and even purchases. In AI Mode you can shop conversationally—describe what you want and get tailored, visual results with prices, reviews, inventory and comparison tables. The Gemini app in the U.S. now surfaces shoppable listings, side‑by‑side comparisons and Shopping Graph–sourced prices. For local availability, a Duplex + Gemini upgrade can call nearby stores on your behalf, summarize responses and return inventory and promo details. And agentic checkout can monitor a tracked item’s price and, with your permission, complete the purchase via Google Pay for eligible merchants (Wayfair, Chewy, Quince and select Shopify sellers), starting in the U.S. For the AI/ML community this is a notable step toward production-scale agentic systems: the Shopping Graph underpins it with >50 billion product listings (2 billion updated hourly), while large Gemini models drive natural language shopping queries, call handling and summarization. Technical implications include real‑time data ingestion, cross‑modal ranking (text, images, inventory), automated action pipelines (calls and purchases) and tighter model–commerce integration—raising opportunities for personalization and efficiency but also questions about trust, consent, robustness, and merchant/consumer safety in agentic commerce workflows.