#e4mXplains:  Will AI agents crash quick commerce's advertising party?

The diyas have barely cooled, the last mithai boxes are being disposed of, and India's commerce platforms and advertisers will be doing what comes naturally after Diwali: counting the money. Somewhere between the order volumes, advertising invoices and arguments over attribution, however, another question has entered the chat. What happens when the customer stops doing the shopping?

Across the Pacific, Amazon and Walmart are already confronting that problem. A Reuters analysis last week examined how autonomous shopping agents could threaten the retailers' advertising businesses, which together command an estimated 88% of America's $74.4-billion retail-media market.

Morgan Stanley estimates agents could handle 10–20% of US ecommerce spending by 2030. Amazon, whose advertising revenue grew 11% in Q3, insists outside agents have not materially affected its business. Nevertheless, it is restricting some third-party agents, while Walmart is opening its systems to them.

The dilemma is straightforward: retailers want the transactions, but would rather not surrender the lucrative advertising opportunities surrounding them.

In India, that question arrives just as commerce media is enjoying its post-festive glow.

Datum Intelligence estimates commerce and retail-media advertising at ₹11,500 crore during the extended 2026 festive season, including approximately ₹2,200 crore from quick commerce, up 50% year-on-year. Its full-year commerce advertising forecast stands at ₹30,091 crore, a 27% increase. These remain estimates, with the broader festive window running beyond Diwali.

 Read: From search to shortlist: AI enters India’s festive buying journey

And shoppers have certainly been busy. Unicommerce's analysis of over 138 million order items found ecommerce order volumes rising 21% between October 6 and November 4. Quick-commerce orders jumped 75%, although GMV grew just 38%. Across ecommerce, 85% of orders were valued below ₹1,000.

More orders, smaller purchases, and considerable competition to persuade shoppers to add just one more item.

For platforms, advertising helps make those economics attractive. Zepto's advertising revenue reached ₹1,636 crore in FY26, up 151%, accounting for approximately 7.8% of its net receivables value. In the September quarter, Blinkit's net order value climbed 79.3% to ₹20,940 crore, while Instamart's rose 34.8% to ₹6,594 crore.

The platforms are processing enormous amounts of purchase intent. And increasingly, brands are paying to intercept it.

 Read: Are AI platforms building advertising’s next walled gardens?

Now imagine an AI assistant being asked to organise a Diwali gathering for twelve people, within a ₹3,000 budget, with delivery before the guests arrive. It identifies refreshments, sweets, disposable plates and perhaps an emergency packet of batteries. It compares prices, checks availability across connected merchants, selects products and prepares the order for approval.

The consumer gets a shopping list. The retailer gets a transaction. But the familiar journey through search results, sponsored listings, promoted alternatives and checkout suggestions may never take place.

That is the agentic opportunity cost.

Retail media has built its commercial appeal around monetising the moments between intent and purchase. An advertiser pays to appear when somebody searches for biscuits, again when they browse alternatives, and perhaps once more when a complementary product is recommended.

An agent compressing that journey into a single shortlist could eliminate several of those opportunities. It may also reduce the browsing signals retailers collect and the chances brands have to influence unplanned purchases. The sale survives; the advertising inventory surrounding it becomes less certain.

Nor is agentic shopping entirely hypothetical in India. Swiggy has opened Instamart's catalogue of more than 40,000 products to compatible AI assistants, allowing them to search, assemble carts and facilitate purchases. Zepto has developed its own agent-compatible shopping interface, complete with personalised search, cart management and payment capabilities. 

BigBasket participated in an earlier OpenAI–NPCI–Razorpay pilot enabling purchases through ChatGPT, while Razorpay has demonstrated transactions across multiple commerce platforms through Claude, as well become a major partner with ChatGPT Ads. 

Read: From payments to media: What Razorpay’s ChatGPT ads play really changes

 The technology remains at varying stages of deployment, but the implications are already visible: India's quick-commerce platforms are preparing for transactions in which customers may never open their apps, much less admire their sponsored listings.

In other words, India's quick-commerce industry is already developing ways to let machines participate in shopping. The interface could increasingly be a conversation rather than a catalogue.

Read: Big Tech wants the shopping journey. What remains for brands?

The commercial distinction is who controls that conversation.

An agent operating inside a retailer's own app could preserve advertising opportunities through sponsored recommendations and promoted product suggestions. An independent agent comparing several retailers, however, may take the decision-making process outside their individual advertising environments.

That presents platforms closer home with something resembling the Amazon-Walmart dilemma: protect the interface or encourage agents to use the underlying commerce infrastructure and find new ways to charge for access to consumers.

For brands, meanwhile, the economics become even more interesting. The fight may shift from buying the first sponsored tile to getting onto an agent's shortlist. Product availability, unit pricing, structured catalogue information and delivery reliability could acquire greater importance alongside conventional paid visibility.

Of course, this cuts both ways. An agent could introduce unfamiliar brands, generate larger baskets and create transactions that might never otherwise have occurred. Sponsored recommendations could become valuable new inventory, although how they are selected and disclosed will matter.

There is also no published evidence yet that agents are materially eroding Indian retail-media revenue. A repeat purchase of an already chosen brand may not have generated much advertising opportunity anyway.

But that is precisely why the post-Diwali reckoning matters. Having spent heavily to win visibility during India's biggest shopping season, brands and platforms must now consider whether the next evolution of commerce will preserve the advertising opportunities they have been paying for.

For years, retailers have perfected the art of charging brands for a better position on the digital shelf.

The next problem is rather more awkward: how do you rent out shelf space when the customer has sent a machine to do the shopping?