Tech Explained: Here’s a simplified explanation of the latest technology update around Tech Explained: Adding AI to the shopping cart in Simple Termsand what it means for users..
What if AI can do this in a few minutes and all you need to do is select the platform, check the ingredients and pay?
This is what Bengaluru-based startup Zave is building–an AI-powered shopping site that can discover products and offer the best pricing to customers.
Users can find the product best for them across multiple platforms without having to check each one of them out.
“These platforms are very good at finding the product you want at the price you want. They have nailed the discovery part of commerce using AI,” said one user, a 31-year-old techie in Bengaluru.
He’s been using AI fashion discovery platforms in the past year. A recent research report from ChatGPT noted that shopping accounted for about 2% of the 2.5 billion prompts it handles per day.
In addition, according to market reports, the global online advertising market is expected to reach $1.3 billion by 2033, growing at a CAGR of 13%. US online holiday sales hit a record $257.8 billion between November 1 to December 31 2025, up 6.8 % year-over-year. This growth was bolstered in part by elevated use of AI tools and generative-AI traffic, according to published estimates by Adobe.
AI-driven traffic to US retail sites increased by 693 % year-over-year during the holiday period, showing how shoppers increasingly used generative AI assistants for product discovery, comparison, and deal hunting. This is the opportunity that large technology and AI companies along startups want to tap into.
Global bet
Companies are betting heavily on this segment worldwide. OpenAI and Google have introduced Agentic Commerce Protocol and Universal Commerce Protocol (UCP), respectively, partnering with retail giants such as Walmart and payment providers such as Stripe to develop and launch these in the US.
Google is seeing increased adoption of UCP in the US, according to a spokesperson. Before that, the company had launched Try On, a tool that allows people to see how clothes look on them virtually, in July 2023. As much as 70% of users found the experience fun, it said.
The company recently launched Business Agent, allowing users to chat with brands through voice, and is used by retailers such as Lowe’s, Michael’s and Reebok. Google is rolling out shopping features in AI mode for India, a key market, as the search giant explores the expansion of UCP beyond the US market.
Perplexity AI is seeing a significant number of questions related to commerce and shopping.
“(For us) shopping is important, not in the sense that we are trying to be a shopping platform, but to make sure that we have the best information,” said Raman Malik, VP, growth, Perplexity AI told ET. The company is working on a product roadmap for shopping, which is expected soon.
AI shopping assistants
“Customers are not just looking for ‘the cheapest option’ anymore, they want the right book for a specific exam pattern, the best phone for content creation, or a weekly grocery plan that fits their diet and budget,” said Kishore Thota, director, shopping experience, emerging countries, Amazon.
Millions of customers engage with the platform through Rufus, Amazon’s AI assistant, he said. During the November earnings call, Amazon CEO Andy Jassy said Rufus is expected to generate $10 billion more in annualised sales for the company.
About 250 million people would use the AI assistant in 2025 to do their shopping, he had said. The company has also launched an AI shopping assistant for third-party companies, Buy for Me, which can purchase products on the user’s behalf from such sources.
Shopping has become central to AI innovation, where billions of ad dollars are at stake, said Glance AI chief operating officer Mansi Jain. “With markets pegging that 50% of shopping can happen via agentic commerce with consumer behaviour changing, companies want to tap into this,” she said. Flipkart too recently launched SLAP, an AI shopping assistant that users can buy from directly using prompts.
Startups are tapping into the momentum too. US firms DayDream got $50 million in funding for an AI shopping assistant and Phia, started by Bill Gates’ daughter Phoebe Gates, has raised $35 million. Flash AI founder and former Flipkart executive Ranjith Boyanapalli said the company is betting that commerce will be research-led.
Users will make purchases based on information thus gleaned across categories such as electronics and skin care. The company has built Flash AI, which can be appended to the website one’s buying from as well as a browser plugin and an app. The custom-built AI model will help users choose, for instance, sunscreen brands by getting information on ingredients and how they work on different skin types and complexions besides suggesting alternatives.
The company is also looking to take it to the next level. AI will provide information on the efficacy of products across categories such as electronics, fashion and skin care, the three biggest for the platform based on queries.
AI-powered fashion shopping assistant ALT is looking to solve the discovery problem, said cofounder Rashida Bohra.
“The biggest challenge right now is that discovery, where there are thousands of products across marketplaces, brands and D2C platforms and the current model is broken as ecommerce platforms go heavy on ads,” she said.
The company has fine-tuned a small language model of 400 million parameters on proprietary fashion data. It works with over 300 brands and marketplaces, helping users find products using prompts.
Bohra said that it has over 500,000 customers on the platform.
Monetisation still a drag
While the opportunity is substantial, the challenges that these platforms grapple with include retention of users and monetisation. The user cited above stopped using the platform once he found the right set of clothes. In the case of big tech, it is still unclear if adoption will pick up as the space continues to evolve.
A founder pointed out that the technology is here to stay and it is important to evolve with it in whatever shape or form it takes. While there is momentum, platforms face some challenges. The biggest of these is retention, said Vivek Madani, co-founder, Firi, an AI-led discovery and ecommerce platform for skin care products. Madani had started the firm in 2025 as an AI-led discovery for skin care.
“We had great traction in the beginning as customers used the platform to discover products. But the issue was they never came back,” he said. He explained that monetisation of AI-led shopping is hard and does not scale as there is no instant gratification. “Ultimately you need to buy the product and for that you go on to brands and online commerce platforms. So we pivoted to an AI-led discovery and quick commerce model late last year,” he said.
The industry has also seen companies such as Alle, backed by Elevation Capital, shut down last year due to lack of a viable business model.
Several companies that ET spoke to are yet to decide on a business model for retention as customer behaviour is continuously evolving. Zave co-founder Hiren Patel said while there is a product market fit, they are focusing on providing value before revenue, where they could see newer models emerging. “We need to show that we can retain customers and have strong discovery. The focus is now on developing a viable business model,” ALT’s Bohra said.
Other challenges
Multiple founders told ET that adoption is yet to pick up in India as users prefer clicks to conversational AI platforms.
“Adoption is moderate on the platform,” said Rajiv Aggarwal, founder of MyStore Genie, an AI shopping assistant for the ONDC marketplace.
Glance AI’s Jain said that while the platform was first launched in India, its focus market is the US, where it has 5-6 million users. Adoption is yet to pick up back home. This is because it’s still too early for the Indian market, where there is a trust deficit and one needs to solve for that before adoption picks up, Zave’s Patel said.
In addition, it is unclear how the landscape will evolve as models continue to become more intelligent and retailers adopt agentic AI.
At the end of the day, AI shopping is still nascent. “Everyone is building what they think the future of commerce will look like,” said one of the founders.