Google's AI shopping features promise convenience: automated tracking, inventory calls, checkout. But they remove everyone between you and purchase—reviewers, influencers, store staff. What's lost when one company owns the entire shopping journey?
Cursor raised $2.3 billion at a $29.3 billion valuation while paying billions to the same AI companies now competing against it. The fastest-growing startup in tech history faces a choice: become a model company or accept structural disadvantage.
Chinese hackers automated 80-90% of cyber intrusions using Anthropic's Claude by simply telling it they were security testers. Four breaches succeeded. The jailbreak was embarrassingly simple, and now every AI company faces the same vulnerability.
Google's AI shopping features promise convenience: automated tracking, inventory calls, checkout. But they remove everyone between you and purchase—reviewers, influencers, store staff. What's lost when one company owns the entire shopping journey?
Google announced Thursday it would handle your holiday shopping from start to finish. Describe what you want, compare products, track prices, complete purchases. All without leaving Google's ecosystem. One tool calls local stores to check inventory. Another automatically buys items when they hit your target price.
The pitch sounds convenient. No more toggling between tabs, no more manual price tracking, no more phone calls to check if Target has that toy in stock.
But Google's shopping upgrade represents more than helpful automation. The company is attempting to capture the entire shopping journey while systematically removing the humans who currently populate it: influencers, review sites, comparison shoppers, even store employees who answer phones.
Key Takeaways
• Google's AI handles discovery through checkout: automated price tracking, store inventory calls via resurrected Duplex technology, and one-click purchases through Google Pay.
• Content creators and review sites face traffic losses as AI synthesizes their work into direct answers without attribution or compensation for human expertise.
• Small merchants face power imbalance: opt out of Duplex calls and lose visibility, or divert staff time to answering robots during peak retail periods.
• Shopping consolidation gives Google unprecedented purchasing data combining search intent, price sensitivity, brand preferences, and completed transactions to enhance advertising targeting.
Google's announcement centers on "agentic AI," a term proliferating across tech marketing this year. These features amount to automated workflows with decision-making capability. When you tell Google's AI Mode you need "a light sweater for my trip to Atlanta, timeless and wearable with jeans and dresses," the system pulls from Google's Shopping Graph of over 50 billion product listings, 2 billion updating hourly. It considers Atlanta's weather without prompting. It generates comparison tables if you're deciding between moisturizers.
The Gemini app now includes these shopping capabilities directly in your chat with the AI.
Google's "Let Google Call" feature revives Duplex technology, which first appeared in 2018 demos showing AI making restaurant reservations. That version never gained traction. Businesses proved reluctant to engage with robot callers, users didn't trust the system, and the feature quietly faded.
Now Duplex returns with what Google calls "a big Gemini model upgrade." Search for certain products near you, toys or electronics or health items, and you can let Google call stores. The AI phones local businesses during business hours, identifies itself as a robot, asks about stock and pricing. You get a text summary.
Track an item's price down to specific size and color. When it drops within your budget at an eligible merchant, Google notifies you and offers to complete the purchase using Google Pay. Initial partners include Wayfair, Chewy, Quince, and select Shopify merchants.
Together these features form a closed loop where Google handles discovery, comparison, inventory checking, price monitoring, and purchase completion.
Consider the current shopping ecosystem. Someone looking for winter skincare starts on TikTok, where creators demonstrate products. Maybe they check The Strategist for professional takes. Reddit for user experiences. Beauty influencers on Instagram. Price comparison on Honey or CamelCamelCamel. Humans at every step: creators testing products, writers researching comparisons, community members sharing what actually worked.
Google's system collapses that journey into a single conversation with an AI. During Google's press briefing, The Verge's Mia Sato noted the example query for moisturizers struck her as notable because she'd made the same search on TikTok the week prior. The AI Mode results pulled recommendations quickly, with comparison tables. She observed: "I could choose what to buy, track the price, and buy it, all from Google, consolidating the entire experience."
Sometimes the system incorporates content from TikTok or Reddit in its training data. But those human recommendations "eventually funnel up to Google's automated system," with attribution buried in an algorithm.
Content creators and product review sites built businesses around shopping recommendations. That beauty YouTuber with 500,000 subscribers? She earns when viewers click affiliate links. The Strategist pays writers who spend days researching "the best winter boots." Revenue depends on people clicking through to the actual content.
If Google answers the question directly, why would anyone visit the original sources? The comparison tables look authoritative because they draw on human labor. But the humans who produced that knowledge don't participate in the transaction.
Google showed ads in AI Mode shopping results during demonstrations. But those sponsored listings compete with organic AI recommendations that already incorporate shopping wisdom from across the web. A small beauty blogger who carefully tests moisturizers has no mechanism to ensure her conclusions get weighted appropriately, or that readers know her work informed the AI's answer.
Small businesses face their own trade-offs. Google positions "Let Google Call" as driving foot traffic to local stores. Someone searches for a toy, Google calls nearby shops, confirms availability. The shopper drives over. Store makes a sale it might have missed otherwise.
That logic assumes several things work in merchants' favor. The AI accurately represents inventory. Stores have staff available to answer calls from a robot. Merchants can afford to have employees field these calls during busy retail periods.
Businesses can opt out of receiving Duplex calls. But that option exists within a power imbalance. Google controls how people discover local businesses. A massive share. Opting out of AI calls might mean missing potential customers. Staying opted in means diverting employee time to conversations with machines.
Google notes the AI will call during business hours and observe "reasonable cooldown" periods. What constitutes reasonable remains undefined. During holiday shopping season, when inventory moves quickly, a store might receive multiple AI inquiries about the same hot toy within hours. Each call requires human attention to answer questions, confirm stock, explain pricing.
Small businesses increasingly face a choice: keep inventory updated in Google's Shopping Graph or risk invisibility. That requires technical capability. Constant maintenance. Time most small retailers don't have. The incentive structure pushes merchants toward feeding Google's ecosystem rather than cultivating their own customer relationships.
The broader pattern involves data. Search for products? Google learns your preferences. Compare items? Google understands your decision-making criteria. Track prices? Google sees your budget constraints and purchasing timeline. Authorize a purchase through Google Pay? The company closes the loop on the entire transaction.
This isn't incidental. Shopping data provides signals that improve ad targeting across Google's properties. Knowing someone searched for portable humidifiers, compared models based on room size, and set a price alert at $89.99 offers far more valuable targeting information than knowing they clicked an ad for humidifiers.
Google has long operated a shopping comparison engine. But these AI features transform the relationship. Previously you'd search, see listings, click through to merchant sites. Google facilitated discovery but didn't own the full journey. Now the company positions itself as your shopping consultant, decision aide, and purchasing agent.
By describing these features as AI agents acting on your behalf, Google frames itself as your representative rather than an intermediary platform. The language suggests the AI serves your interests. But the system optimization points elsewhere: keep users in Google's ecosystem, gather more behavioral data, position Google as the essential layer between consumer desire and product acquisition.
Tech companies have adopted "agentic AI" as preferred terminology for systems that take actions rather than just providing information. The marketing usage often describes simpler automation dressed in aspirational language.
Google's price tracking that automatically completes a purchase when conditions are met? An if-then statement with a payment API attached. Calling it an "agent" implies more sophisticated reasoning than actually occurs.
Duplex calling has more claim to agency. It engages in dynamic conversation, adjusts questions based on responses, summarizes information. Still, tightly bounded parameters.
The gap between "agentic" as marketing and "agentic" as technical capability creates space for misunderstanding. When Google says its AI will "buy items for you," some users hear an intelligent system making informed decisions. The reality involves more modest automation: execute a pre-approved purchase when a price condition triggers.
This matters because autonomy implies trust. Believe an AI agent actively represents your interests? You'll delegate decisions to it. That delegation is the point. It reinforces Google's position as essential infrastructure for shopping.
Several dynamics strain as Google consolidates shopping. Creator economy faces direct pressure. AI responses absorb countless creators' work without attribution or compensation. Product review sites that depend on affiliate revenue watch traffic diverted to Google's comparison tables. Independent shopping communities become training data rather than destinations.
Price competition mechanisms also shift. Right now, merchants compete across platforms. When Google becomes the primary interface, optimization increasingly means optimization for Google's systems rather than for actual best prices or quality. Merchants who integrate most deeply with Google's Shopping Graph, who process Duplex calls efficiently, who enable agentic checkout gain visibility advantages.
Consumer choice architecture changes when one company controls the shopping conversation. Google's AI Mode shows you options based on its understanding of your query. Its reading of 50 billion product listings. But what gets weighted higher? Products from merchants who pay Google more? Items with better margins? Products that integrate smoothly with checkout? The company hasn't detailed those ranking factors.
The parallel to Google's web search feels relevant. For years, the company maintained that search results aimed to show the most useful information. But countless studies have documented how commercial incentives shape results, how Google properties appear prominently in searches that could link to external sites, how the line between ads and organic results has blurred.
Shopping introduces even more direct financial incentives. Google processes the payment, facilitates the transaction, sees the revenue.
Google's shopping features will prove useful. Automated price tracking saves time, maybe money. Conversational search beats clicking through filters. Having local inventory checked eliminates the frustration of driving to stores for sold-out items.
Convenience has value. Real value. The question is what you trade for it.
When shopping involved visiting multiple stores, comparing circulars, calling around for availability, people accepted that purchasing required effort. The internet reduced that friction through centralized information. But it preserved plurality: multiple retailers, multiple review sites, multiple sources of recommendation.
Google's vision pushes toward singular integration. One company mediates discovery, comparison, purchasing. One AI synthesizes all shopping knowledge. One system tracks your preferences, budgets, and transactions.
That centralization offers efficiency. It also creates dependencies. If Google becomes the default layer between consumer desire and product acquisition, what happens when Google's incentives diverge from yours? When ranking factors favor merchants who pay more? When AI recommendations optimize for engagement rather than satisfaction?
These aren't hypothetical concerns. They're questions worth asking before delegating shopping decisions to the same company that provides your search, email, maps, calendar, and now wants to handle your purchases.
Q: Why did Google's Duplex fail when it launched in 2018?
A: Businesses refused to engage with robot callers making reservations, and users didn't trust the system to represent them accurately. The feature quietly disappeared from Google Assistant after minimal adoption. Google now claims a "big Gemini model upgrade" makes Duplex more capable, though the fundamental friction of businesses talking to robots remains.
Q: Which retailers support Google's automatic checkout feature?
A: Google's agentic checkout currently works with Wayfair, Chewy, Quince, and select Shopify merchants. The feature is rolling out gradually to U.S. users. You must use Google Pay for the automated purchases, and you'll approve each transaction before Google completes it on the merchant's website.
Q: Can my business opt out of receiving AI calls from Google?
A: Yes, merchants can opt out of Duplex calls. However, this creates a dilemma: Google controls how most people discover local businesses, so opting out might mean missing potential customers. Google says the AI will call during business hours and observe "reasonable cooldown" periods, but hasn't defined what that means.
Q: How big is Google's Shopping Graph that powers these features?
A: Google's Shopping Graph contains over 50 billion product listings, with 2 billion of those updating every hour. This gives Google real-time inventory, pricing, and availability data across retailers. The system combines this product data with Gemini AI models to power conversational shopping, price tracking, and automated purchasing.
Q: What happens if Google's AI buys the wrong item automatically?
A: Google requires your approval before completing any automated purchase. You'll receive a notification when your tracked item hits your target price, then review the product details, shipping, and payment method before Google proceeds. The AI doesn't make autonomous purchases without this final confirmation step, though the system is marketed as "agentic."
Marcus Schuler
Marcus Schuler
Maria Garcia
Microsoft led AI infrastructure with 60% market share in 2023, then froze 3.5 gigawatts of planned capacity. OpenAI signed over $400 billion with Oracle instead. Nadella calls it strategic repositioning. SemiAnalysis data suggests Microsoft blinked first.
OpenAI released GPT-5.1 without publishing benchmarks, leading instead with "warmer" personality controls. The shift reveals how 800 million users push the company toward engagement metrics over technical capability, mirroring social media's path.
SoftBank just posted record profits driven by OpenAI's soaring valuation. So why is Masayoshi Son selling $15 billion in assets, expanding margin loans, and converting paper gains into cash? The answer reveals uncomfortable truths about AI's peak.
Meta's chief AI scientist Yann LeCun reports to a 28-year-old after twelve years building FAIR. Now he's leaving to raise billions for the exact research Meta couldn't afford. The death of corporate AI labs and the VC arbitrage replacing them.
Get the 5-minute Silicon Valley AI briefing, every weekday morning — free.
