How AI shopping agents will drive online sales

On Thursday, Business Standard reported that the government is working on building a national infrastructure for agentic payments. The proposed Unified Agents Protocol (UAP) is being designed to create an interoperable infrastructure with a common standard for AI agents and apps

MUMBAI: If you want to buy, say, a mattress, under ₹20,000, which is also good for your bad back, you no longer browse multiple e-commerce websites offering hundreds of products to choose from, compare them and then place the order. You probably already use AI Chatbots to do most of your work. They do the research, make comparisons and give you the best recommendations.

How AI shopping agents will drive online sales

However, very soon, your AI assistants will also be able to place the order and make the payment without you having to even open a website to see a mattress brand. This is agentic commerce, the future of online shopping, where an autonomous AI agent not only researches for a buyer but is also able to make the purchase. It’s started to happen in the US where Google, for instance, has launched Universal Commerce Protocol (UCP) that allows AI assistants like Gemini to discover products on some popular stores and complete transactions.

Back home, the Indian government seems to have an ambitious plan to make AI-assisted online shopping seamless. On Thursday, Business Standard reported that the government is working on building a national infrastructure for agentic payments. The proposed Unified Agents Protocol (UAP) is being designed to create an interoperable infrastructure with a common standard for AI agents and apps.

The development backs TruCommerce.ai co-founder Umang Chhaparia’s belief that Indians are very keen to use AI shopping assistants. Adobe’s June 2026 ‘AI and Digital Trends Report’ also said that 60% Indian consumers want personal AI agents to help them shop—the highest in Asia-Pacific. Chhaparia added that our infrastructure is 25% ready. “There are products like UPI Reserve Pay where the idea is to give a wallet to your agent, and authorization on the kind of transactions it can do,” he said.

He expects agentic commerce to boom here because shopping via personal AI agents is convenient. “They can evaluate products from a bigger spread across websites, make informed purchases and also because of the speed at which they can do it,” Chhaparia said.

Adarsh Menon, partner at Fireside Ventures which invests in consumer brands, said agentic commerce will flourish for two reasons: our young demographics and our comfort level with new technology. “Indians are quick on the uptake of understanding and using tech right from school children to retired people,” he said.

“Agentic commerce will be a massive disruption when it happens because the shopper may never see an ad or visit a website before a purchase. The AI shopping assistant will be the new layer between consumers, marketplaces and brands,” said Menon, adding that the value of data as a structural advantage in training and personalizing AI, cannot be emphasized enough.

McKinsey has projected agentic commerce to orchestrate $3 trillion to $5 trillion in global transactions by 2030.

Saket Dandotia, founder, Onetab.ai, said personal AI agents have led brands to build agentic AI in their storefronts. “We are not only in the business of agentic AI workflows and automation, but also offer solutions for agentic storefronts,” said Dandotia. He thinks AI shopping agents may benefit smaller, little-known brands as they comb the web more widely for better matches and relevant products.

TruCommerce is also building the infrastructure layer for agentic commerce and counts brands like ITC, Vodafone, Puma and Amazon among its clients. “We set up agentic storefronts for the brands so that agents can talk to them. We also map them against their competitors. AI agents shop differently. Earlier, brands had to worry about how to be on the top in a marketplace. Now they need to be within the purview of the AI agents,” Chhaparia said.

The impediments

But AI shopping agents will pose new challenges. For starters, traffic to e-commerce websites may fall, even though Menon said that ultimately the transactions will happen there. Dandotia said brands stand to lose the most as they will no longer connect directly with customers but with AI assistants. “So, they may not understand consumer behaviour as the engagement is happening on, say, a ChatGPT,” he said.

Menon said the real risk is brand hollowing. Though the identity of a brand is in its visual assets, if its data is poor and reviews are bad, the AI agents will take consumers away from it. “It is a risk for brands,” he said.

AI assistants may have their biases since a lot depends on the quality of data they have been trained on. “Say, if a brand has well-structured data and more reviews, it may get recommended over a genuinely superior but not so digitally mature brand,” Menon said. Consumers must understand that agentic AI is not neutral. It reflects the data quality of the ecosystem, he added.

There are concerns around safety too as the shopping assistant should not make unintended purchases or unauthorized payments. Going forward, agentic commerce will need more clarity on fulfilment of real-world commerce needs such as warranties, service installations etc.