So Google is testing shopping inside Gemini and AI Mode in India. The partner? Flipkart — Walmart’s crown jewel in the subcontinent. The pitch is exactly what you’d expect: ask the assistant, find a product, buy it without ever leaving the chat window. TechCrunch reported the limited test covers select products and users, with a broader rollout planned for later in October. Cue the triumphant music.
I’ve been covering this space since before the metaverse was a punchline, and I’ve learned one thing: the gap between a demo and a habit is where most of these announcements die. So let’s pull this apart. What is Google actually shipping? Why Flipkart? And does any of this matter for the virtual worlds and AI-mediated commerce we keep being promised?
The pitch: shopping without the store
On paper, it’s clean. You open Gemini. You type something like « wireless earbuds under 3,000 rupees with good noise cancellation. » The assistant parses intent, pulls live listings from Flipkart, surfaces options, and lets you check out right there. No app switch. No tab juggling. No comparison-site rabbit hole.
That last part is the real prize. Google has spent two decades training us to start product searches in its search bar, then hand us off to retailers. Every handoff is a leak. Attention leaks, data leaks, ad revenue leaks. If Gemini can keep the transaction inside Google’s surface, the company claws back control of the funnel it created in the first place.
Flipkart makes sense as a first partner. India is the market where Google needs a win in AI, and Flipkart is the local giant with the inventory, logistics and payment rails already in place. Walmart bought it in 2018 and has been pouring capital into it ever since. For Flipkart, being the default commerce layer inside Gemini is a distribution gift. For Google, it’s a live-fire test of whether an assistant can actually close a sale.
Why India, and why now
India is not a side quest for Google. It’s the main stage for the next billion users, most of whom will never own a desktop. They live on phones, they use voice, and they’re comfortable with chat as an interface. That’s the exact profile agentic commerce needs.
There’s also a competitive clock ticking. OpenAI has been inking commerce deals. Amazon is stuffing Rufus into everything. Perplexity wants to be the answer engine that also sells you things. If Google sits still, the assistant layer becomes someone else’s storefront.
And let’s be honest about the timing. A limited test in September, broader rollout in October — that’s not a product launch, that’s a land grab dressed as a beta. Google wants the headline before the holiday quarter and before its rivals can claim the same narrative.
The part everyone skips: trust
Here’s where I get skeptical. Buying something is not the same as searching for something. Search is low stakes. If the result is wrong, you refine the query. Commerce is high stakes. Money leaves your account. Returns get messy. Customer service becomes a maze.
When an AI assistant recommends a product, who is responsible if it’s a dud? Google will say Flipkart. Flipkart will say the seller. The seller will say the algorithm. And the user is left holding a broken gadget and a chat log.
Then there’s the ranking question. If Gemini surfaces three earbuds, why those three? Is it relevance? Is it margin? Is it a sponsored slot? Google has a long, uncomfortable history of blurring the line between organic results and ads. Inside a chat window, that line becomes almost invisible. Unless Google is brutally transparent about why a product appears, users will eventually feel manipulated — and they’ll be right.
I’m not saying this can’t work. I’m saying the hard part isn’t the checkout button. The hard part is trust, and trust doesn’t ship in a sprint.
What this means for the metaverse crowd
Regular readers know I don’t let the metaverse off the hook either. For years we were sold a vision of shopping in virtual malls with avatars trying on digital sneakers. Most of it was theater. The actual future of immersive commerce looks a lot more boring: a text box, a smart assistant, and a logistics network humming in the background.
That’s the irony. The metaverse promised presence. What we’re getting is convenience. And convenience wins almost every time.
If Google and Flipkart can make assistant-led checkout feel effortless, they’ll have built the real spatial commerce layer — not a 3D mall, but a conversation that ends in a delivery. Web3’s dream of decentralized storefronts looks even more like a niche hobby from here. The rails are being laid by the same big platforms everyone said we’d escape.
The questions Google hasn’t answered
The TechCrunch summary is thin, and that’s telling. Limited products. Limited users. Broader rollout later. Fine. But the details that matter are missing.
- How does Gemini handle returns, refunds and disputes inside the chat?
- Are recommendations disclosed as ads, or does Google blur the line again?
- What data does Google keep from the transaction, and what does Flipkart get?
- Does the assistant compare prices across retailers, or only Flipkart?
That last one is the tell. If Gemini only ever shows Flipkart listings, it isn’t a shopping assistant. It’s a storefront with a chat skin. Nothing wrong with that — but let’s call it what it is.
My read
I think this is a smart, calculated move by Google. India is the right beachhead. Flipkart is the right partner. The assistant-as-storefront is where commerce is heading, whether we like it or not.
But I’ve watched too many « revolutionary » commerce integrations quietly die in a settings menu. The test will succeed or fail on three things: trust, transparency and whether people actually change their buying habits. That last one is the hardest. We are creatures of routine. We open the Flipkart app because we always have. Convincing us to buy through a chatbot is a behavior change, not a feature launch.
So no, this isn’t a game-changer. It’s a bet. A good one, maybe. But a bet. And the house — as always — is Google.
Original source: read the full article