In announcing the update, Google said it wants shopping queries to feel more conversational. The company noted that users can describe what they want “just as you’d say it to a friend,” and the system will pull from what Google describes as a constantly refreshed Shopping Graph containing “more than 50 billion product listings, 2 billion of which are updated every hour.”
The blog explained that when users define their budget and product specifications, “down to the specific size, color and amount you want to spend”, they’ll get a notification when the price falls within budget.
Google also said that if a merchant supports it, users may opt into having the system place the order on their behalf, though the company stresses that the assistant will request confirmation first and will only buy after the user approves shipping and payment details.
Product Discovery and Store Calls
A new feature is the ability for Google’s AI agent to call nearby stores. Google said that when users perform a “near me” search, they may see a “Let Google Call” option, which triggers a call to the retailer to ask about pricing, availability, or promotions, and then delivers a transcript of what the agent asked, and the store responded. Each call requires user authorization, which Google described explicitly.
Google also shifted product research into a conversational thread rather than discrete search sessions. A user can ask for “gift ideas under fifty dollars for a science-loving child,” then refine to “now in blue packaging,” without starting over.
For checkout automation, Google lets users set rules such as maximum spend, preferred delivery or product specs. Once those are active, the agent monitors inventory changes and price movements, then executes the purchase after user confirmation. Google said the system logs each step so users can review what actions the assistant took.
How Google’s Approach Differs From ChatGPT Checkout
The rollout arrives as OpenAI’s “Instant Checkout” feature expands inside ChatGPT. That service permits users to purchase items from sellers like Etsy and Shopify directly within ChatGPT, powered by an open-standard Agentic Commerce Protocol developed with Stripe.
Google’s model diverges in two ways. First, it relies on its Shopping Graph and routes transactions to merchant sites rather than embedding checkout entirely inside a single chat interface. Second, it moves beyond conversational ordering toward rule-based automation and operational tasks like store calls.
Meanwhile, as reported by PYMNTS, Walmart customers will soon complete purchases in ChatGPT through the retailer’s integration of Instant Checkout and the agentic commerce model. These differences highlight how platforms are choosing distinct entry points into agentic shopping.
Where Perplexity Shopping Fits
The answer-engine approach from Perplexity Shopping remains focused on sourcing and precision rather than automation. Its “Buy With Pro” feature lets U.S. Pro-subscribers research and purchase items directly from its interface, but it does not call stores or execute rule-based purchases in the same way.
Google’s new updates combine research, verification, plus autonomous execution in one flow.
According to Bain, customers are increasingly using AI as a starting point, as 17% of unique online shoppers said they will begin their holiday shopping with an AI platform such as ChatGPT or Perplexity, and 30% plan to start with Google search.
