Google Gemini tests e-commerce shopping in India: AI agents evolve from helping you search to helping you place orders

Google is transforming its AI assistant from an “information入口” into a “transaction入口.” According to TechCrunch, Google has begun testing a new feature in India: users can directly place orders on Gemini and AI Mode Walmart’s e-commerce platform Flipkart Purchase products without ever leaving the AI interface. This means AI agents are evolving from “helping you search” to “helping you buy,” and e-commerce’s transaction loop is for the first time truly integrated into large language models.

What happened: A “Buy” button stitched AI and e-commerce together.

According to TechCrunch’s reporting and hands-on testing, users in the beta test see a “Buy” buttonon select Flipkart product listings when searching for items in Gemini and Google’s AI Mode.

Clicking this button takes users directly into Flipkart’s checkout flow—no need to exit the AI interface, no need to open Flipkart’s app or website.

AI Mode: Google Search’s “AI-powered” answers

To understand the significance of this test, first clarify what AI Mode is. Simply put, it’s an AI-enhanced mode of Google Search; after switching in the search bar, users receive not traditional blue-link lists but instead structured, AI-generated responses—complete with images, comparisons, and recommendations. It shares the same underlying model as Gemini Gemini chatbot Gemini chatbot but differs in emphasizing the fusion of “search results + AI generation.”

Google has previously attempted various “transactional” features in AI Mode—such as displaying product cards, comparing prices, and recommending purchase channels—but these were all “traffic referrals,” ultimately redirecting users back to e-commerce sites. This time, embedding the “Buy” button directly on Flipkart brings the transaction itself into the AI response.

Why India? Why Flipkart?

Google chose India as the first test market for practical reasons:India is one of Google’s fastest-growing user marketsand also the most competitive arena for AI search. Flipkart holds a unique position—it’s the e-commerce giant owned by Walmart, competing head-to-head with Amazon in India’s e-commerce market.

• Walmart backingWalmart, Flipkart’s parent company, is deploying AI shopping capabilities globally; this collaboration is mutually beneficial.
• High e-commerce penetrationIndian users have extremely high acceptance of “one-stop shopping on mobile,” making India an ideal market to test AI-powered transactional closed loops.
• Avoid antitrust-sensitive areasGoogle faces antitrust pressure in the US and Europe; launching first in a relatively lenient yet massive market reduces risk.

What this means for users: Three real changes brought by AI shopping

Though still in internal testing, the experiential shifts this feature heralds are already clear:

• One fewer stepThe shopping journey shortens significantly—from “search → navigate → purchase” to “search → purchase.”
• AI assists your decision-makingGemini can screen options based on price, reviews, and specifications, then directly advance you to checkout—eliminating manual price comparisons.
• Trust is criticalEntrusting AI with real money raises users’ expectations for recommendation accuracy and security; a single erroneous recommendation could destroy trust.

Impact on the e-commerce landscape: The battle for the transaction entry point has just begun

Underlying this trial lies a broader competitive logic:Whoever controls AI’s “transaction entry point” controls the next generation of e-commerce traffic.If Google can make AI Mode the default starting point for shopping, e-commerce platforms will devolve from “destinations” into “back-end fulfillment providers,” and their profit structures will be reshaped accordingly. This is why Amazon—as well as domestic players Douyin E-commerce and Pinduoduo—are aggressively investing resources into AI shopping. Meanwhile, Google’s own Gemini monthly active users have already surpassed 1 billion—such a massive user base gives Google a strong hand in the “transaction entry point” race.

For ordinary users, AI shopping will initially roll out in standardized categories (e.g., smartphones and home appliances—products with clearly defined specs), while experience-heavy categories like apparel and beauty will continue exposing AI’s limitations for some time.

Notably, Gemini’s ecosystem is expanding rapidly: from integration with iPhone’s Siri, to @Adobe commands for Adobe Creative Cloud suite, and now directly placing orders—Google clearly aims to turn Gemini into a super gateway covering “querying—creating—transacting.”

Frequently Asked Questions (FAQ)

Is Google’s AI shopping feature available now?

Currently, it is in limited beta testing only in India for select users, and covers only a subset of Flipkart’s products. Users in China and the U.S. cannot access it yet; Google has not announced a specific rollout timeline.

Who receives payment when buying via Gemini?

Payment goes directly to Flipkart. Google provides only the “gateway” and “recommendations”; transaction processing, payment handling, logistics, and after-sales service remain fully managed by Flipkart, with users completing checkout within Flipkart’s flow.

Is AI-assisted ordering secure? Could it order the wrong item?

This is the biggest risk in AI shopping. The current beta version still requires users to manually confirm both product selection and payment—the AI handles filtering and recommendations only, not fully automated purchasing. Google also emphasizes that users must personally confirm checkout.

How does this differ from shopping in a conventional e-commerce app?

The core difference lies in the “decision-making” step: traditional e-commerce requires you to search, compare prices, and filter manually, whereas AI Mode can directly recommend products and proceed to checkout based on your query (e.g., “best battery-life smartphone under ¥5,000”), eliminating intermediate price-comparison and navigation steps.

Want to be the first to know about every evolution of AI tools? Check out our AI Model Library and Tool Comparison Engineor continue reading:Google Gemini 4 Launch Preview · Gemini Monthly Active Users Exceed 1 Billion · Google I/O 2026 Full Coverage · Gemini 3.8 Live Review.