AI assistants and agents influenced 262 billion dollars in online sales over the 2025 holidays, so sellers now have to be readable by machines, not just people.
During the 2025 holiday season, AI shopping tools influenced about 20 percent of all online retail sales worldwide. That came to 262 billion dollars, according to Salesforce data covering more than 1.5 billion shoppers. The number is not a forecast. It already happened.
Shopping used to mean searching, opening several tabs, comparing prices and then buying on a brand's site. Now a shopper can tell an AI assistant what they want, such as trail running shoes under a set budget, delivered by a certain date. The assistant searches and compares. Increasingly, it can also buy with little human input. This is what the industry calls agentic commerce: an AI agent shopping on your behalf.
McKinsey estimated in October 2025 that AI agents could handle 3 to 5 trillion dollars of global consumer sales by 2030. Salesforce also found that retailers using their own AI shopping agents grew sales nearly 60 percent faster than those that did not.
The tools are being built quickly. In January, Google announced the Universal Commerce Protocol, an open standard made with Shopify, Etsy, Wayfair, Target and Walmart. It lets agents find products, check stock and complete checkout inside Google's AI Mode and Gemini. OpenAI and Stripe released a similar standard earlier, and it powers shopping in ChatGPT.
In October 2025, India's National Payments Corporation, Razorpay and OpenAI began a pilot letting ChatGPT users pay through UPI without leaving the chat. UPI is India's real-time payment system, and it works on the user's consent. BigBasket and Vodafone Idea were the first merchants. By February, Claude joined, with Zomato, Swiggy and Zepto on board.
An AI agent does not respond to a striking photo or a clever tagline. It reads structured product data. It checks whether stock levels are accurate, whether the return policy on the website matches the product feed, and whether prices agree across channels. India's direct-to-consumer sector, valued at over 100 billion dollars by Mordor Intelligence, has put its effort into visual storytelling and influencers. Far fewer brands have made their data clean for machines.
That means a brand with loyal customers and strong reviews could still lose a sale over one inconsistency in its product feed. An agent might also fail to read a site well enough to find the reviews at all.
The source article names making a website machine-readable and trustworthy as the first priority, but the text available here cuts off before it lists further steps.
Audit your product listings so stock, prices and return policies agree everywhere, because AI shopping agents read data, not photos.
Mastercard describes AI tools that could buy things for you, and says trust is the hard part: who gets to spend your money, and under what rules.
Read more: Mastercard outlines how AI assistants could shop and pay for youWorld models, AI that learns physics from video, are being built into home robots and driverless cars, and your own videos may help train them.
Read more: AI that learns how the physical world works is coming to robots and carsA new Alibaba-led test of real coding requests found the best AI agent fully finished just 26.4 percent of tasks, so human review still matters.
Read more: AI coding helpers fully finish only about 1 in 4 real tasks
