When AI Does the Shopping, Who Gets Paid? Ads, Amazon and Google at a Turning Point
Nearly 80 million people used JD.com's AI shopping assistant in China, while United Airlines moved to restrict outside AI agents. If AI picks the product, ad models for Amazon, Google and brands may change.
The 20-second read
1. The shopper has changed
"Find me a washing machine that's quiet and fits on a narrow balcony."
Shoppers used to type keywords into a search bar, open dozens of listings and compare reviews. Now they describe their needs to an AI once and get a recommendation. The AI compares prices and specifications, and can carry the purchase through.
China's Singles' Day shopping festival put numbers on the shift. JD.com said about 80 million people used its AI shopping assistant, Dongdong, more than triple the number a year earlier.
- About 80 million: users of JD.com's AI shopping assistant
- More than 3x: user base versus a year earlier
The point is less that more people visited the store. It is that more consumers are handing the work of comparing products to an AI.
2. What if the purchase is decided before the ad is seen?
Online shopping used to run from search to a results list, a product page, a cart and checkout. Search ads, ranking position and product pages were the main places where money was made.
On Douyin, China's short-video platform, one private analysis found that about 5% of October purchases came through AI chatbot links. For some products where comparison matters, such as dietary supplements, the share topped 20%.
That does not mean shopping as a whole is booming. Early tallies put Singles' Day sales growth at 2.4%. AI's first effect may be less about growing total spending than about changing which products consumers choose.
For brands, the new contest may be moving from appearing on the first page of search results to making it onto an AI's shortlist.
3. Meanwhile, a US airline put up a gate
On the same day, news pointed the other way in the US. United Airlines said it would restrict unauthorized third-party personal AI agents from accessing its services. Indirect access through travel agencies and price-comparison sites is covered too.
The airline's concerns are account information exposure, unintended payments or mileage use, and regulatory issues in the check-in process. An AI that finds flights well is one thing. An airline granting that AI actual booking authority is another.
It is not a ban on all AI transactions. Travel platforms such as Expedia and Booking are working with OpenAI. A new path for shopping is opening, and companies are starting to manage the doors.
The winner in AI shopping may not be only the smartest recommendation engine. It may be the service authorized to complete payments safely.
4. What changes for Amazon and Google?
Amazon (AMZN) has purchase history, logistics and payments in one place. That is a base from which an AI could link recommendation to delivery. But if shoppers see fewer sponsored product listings, Amazon has to find a new balance between selling convenience and ad revenue.
Alphabet (GOOGL) faces a shift if the habit of finding products through search moves to AI recommendations: the place of search ads and shopping ads changes too. What matters is less search volume itself than how large a role Alphabet secures in AI-driven purchase decisions.
Brands change as well. Eye-catching ads matter less than making product specifications, prices, quality and purchase terms something an AI can read and compare accurately. Advertising may well evolve from buying "human clicks" to earning a place on an "AI's recommendation list."
With Amazon and Alphabet, the thing to check is less the polish of the AI features than where AI actually routes purchases, and who earns money from those transactions.
5. Three signals for investors
- Real purchase conversion at US big tech: whether recommendations lead to completed payments, not just whether AI shopping features launch.
- Changes in ad revenue: how search and shopping ads hold or grow revenue in an AI-recommendation environment.
- Platform partnerships and blocks: where AI agents are officially allowed to connect and transact.
In an era when AI chooses, does the advantage go to the brand that spends the most on ads, or to the product that best fits the consumer's conditions? If the answer shifts, the revenue model of shopping platforms, and the companies investors need to watch, could shift with it.
- China's JD.com said its AI shopping assistant, Dongdong, has about 80 million users.
- That is more than triple the number from a year earlier.
- A private analysis found that about 5% of purchases on China's Douyin in October came through AI chatbot links.
- For some products where comparison matters, such as health supplements, the share topped 20%.
- It is hard to say so yet. Early tallies put Singles' Day sales growth at 2.4%.
- AI's first effect may be to change which products consumers choose, rather than the total amount they spend.
- It said it will restrict access by unauthorized third-party personal AI agents.
- The concern is the exposure of account information, unintended payments or mileage use, and check-in rule problems.
- It is likely to shift from buying human clicks to getting into the AI's list of recommended candidates.
- If so, making sure AI reads product specs, prices and quality accurately could become more important.
What to watch
Insight Times Editorial Desk
