Picture this: you tell an app you want a specific board game in stock nearby, and within minutes an AI agent has phoned three stores, asked about price and availability, and texted you a summary, all without you dialing a single number. That is not a concept demo. It is live in the United States right now, and it marks one of the clearest signs yet that AI shopping agents have moved from novelty to daily utility.

Google’s rollout of “Let Google Call” and its companion agentic checkout feature is the latest proof that autonomous commerce is arriving faster than most retailers expected. For anyone building with or adopting AI agents, this is a preview of where customer-facing automation is headed next: agents that do not just recommend, but act, negotiate, and transact on a person’s behalf.

In this article, you will learn how Google’s new shopping agents actually work, what the adoption data says about where agentic commerce stands today, and what practical steps businesses and shoppers should take as AI shopping agents become a normal part of buying things online and offline.

How AI Shopping Agents Work: Google’s Agentic Checkout Explained

Google built its new capability on two connected pieces. The first, called “Let Google Call,” appears when someone searches for a product “near me.” After a short back and forth about brand, size, or budget, Google’s AI, powered by its long-running Duplex voice technology and Gemini, places real phone calls to nearby stores to check stock, pricing, and current promotions. The shopper then gets a plain text or email summary of what each store said, no hold music required.

The second piece, agentic checkout, goes a step further. A shopper can ask the AI agent to track a product’s price and buy it automatically once it drops below a set budget. When that condition is met, Google confirms shipping and payment details and completes the purchase through Google Pay, with the shopper’s prior approval baked into the setup.

Both features, detailed in Google’s own announcement, are rolling out first for categories like toys, electronics, and health and beauty in the United States, with Google requiring the AI caller to identify itself as automated and giving stores the option to opt out. That transparency layer matters because it addresses one of the biggest open questions in agentic commerce: how much a business or consumer actually trusts an agent to act on their behalf without a human double-checking every step.

Why Agentic Commerce Is Growing Faster Than Retailers Expected

The numbers behind this shift are striking. AI-driven shopping traffic grew 693% year over year in the most recent holiday season, and AI agents already influence roughly $262 billion of global online spend, close to 20% of total e-commerce activity. Morgan Stanley’s research puts large language model adoption for shopping tasks near 50% among US consumers, with AI agents capturing an estimated $190 billion to $385 billion in e-commerce value.

Consumer sentiment backs up the growth. Roughly 39% of shoppers say they have already used AI for online shopping, and 85% of that group reports the experience actually improved how they shop. A third of consumers expect at least 10% of their purchases to be AI-driven within a year.

Retailers see the same trend from the other side, and it is making many of them nervous. Nearly three quarters of merchants agree that consumers will adopt agent-led shopping faster than most businesses are prepared to support. That gap between shopper enthusiasm and merchant readiness is exactly why Google, alongside a broader push like the Universal Commerce Protocol launched earlier this year, is racing to standardize how AI agents discover products, compare prices, and complete transactions. It is the same standardization pressure showing up in Mastercard’s push toward autonomous AI agent payments, where machine-to-machine transactions are being built on shared rails rather than one-off integrations.

What This Means for Businesses and Shoppers Right Now

If you run a business with any local or e-commerce footprint, the practical takeaway is to check whether your store can even be found and understood by an AI agent today. That means clean, accurate, and machine-readable inventory data, current pricing, and a phone or chat channel that can handle an automated caller identifying itself honestly. Businesses that ignore this risk becoming invisible to a growing share of shoppers who let an agent do the comparison shopping for them.

For individual shoppers curious about how AI shopping agents work in practice, start small. Try a price-tracking agentic checkout for a single, non-urgent purchase before trusting an agent with anything high stakes. Read the transparency disclosures Google and other platforms provide, and keep a clear budget ceiling so the agent cannot act outside boundaries you actually intended.

For developers and entrepreneurs building in this space, the opportunity is in the gap merchants openly admit they have. Tools that help smaller retailers expose accurate, structured product and inventory data to AI agents, or that help businesses set guardrails around automated calls and purchases, are likely to see real demand in the next twelve months. This mirrors what is already happening across AI agents transforming retail and e-commerce more broadly, where personalization and checkout automation are moving in lockstep.

The Road Ahead for AI Shopping Agents

Google is not alone here. OpenAI, Perplexity, and a growing list of commerce platforms are all racing toward the same destination: a world where an AI agent can discover, negotiate, and pay for something with minimal human input. The Universal Commerce Protocol suggests the industry already sees the need for shared standards rather than a dozen incompatible agent ecosystems.

The more nuanced view is that trust, not technology, is the real bottleneck. MIT’s Initiative on the Digital Economy and Forrester research both point to the same friction: only about 24% of US online adults currently trust an AI agent to handle a routine purchase on its own. Consumers are willing to let an agent check prices and even make small purchases, but full autonomy for large or sensitive transactions will take longer to earn. That trust gap is also why clear AI agent governance frameworks matter as much for consumer-facing shopping agents as they do for enterprise deployments. Expect gradual expansion into more product categories, tighter opt-out and disclosure rules for businesses, and continued debate over who is liable when an autonomous purchase goes wrong.

Key Takeaways

AI shopping agents have moved past the demo stage. Google’s Let Google Call and agentic checkout features show autonomous agents already phoning stores and completing purchases with real money, and adoption data suggests this is not a fringe behavior but a fast-growing habit among nearly 40% of online shoppers.

Businesses that want to stay visible to this new class of shopper need to get their inventory, pricing, and contact channels agent-ready now, while individual shoppers should start with low-stakes purchases to build comfort with the technology.

Want to keep up with how AI agents are reshaping business and everyday life? Explore more tools, trends, and deep dives at BigAIAgent. What is the first purchase you would trust an AI agent to make on your behalf without asking you first?