How agentic AI is changing who influences online shopping decisions

As AI shopping agents evolve from assistants to decision-makers, retailers may need to rethink how they reach customers, present products and compete online

"Buy me the best 55-inch TV under ₹60,000. Wait until the price drops below ₹55,000, use my preferred credit card for the cashback offer, and have it delivered this weekend."

 

A few seconds later, an AI agent has compared products across retailers, tracked discounts, selected the best deal and placed the order without its user visiting a single shopping app or website.

 

It sounds like the next evolution of online shopping. For retailers, however, it could represent something far more disruptive.

 

For decades, brands have fought to attract consumers through advertisements, search rankings, app downloads and carefully designed storefronts. Agentic AI threatens to change that equation by placing an intelligent intermediary between retailers and customers. The consumer may still buy from Amazon, Flipkart or Nykaa, but the decision on what to buy, where to buy it from and when to complete the purchase could increasingly be made by AI.

 

According to an IDC report, this transition could become one of the biggest structural changes to retail since e-commerce. Rather than simply adding another sales channel, agentic AI can shift influence away from merchants and towards the companies building AI shopping agents.

From shopping to delegating

Today's online shopping journey is still largely consumer-driven. A shopper searches on Google, browses marketplaces, watches YouTube reviews, reads product comparisons, opens multiple tabs, checks prices and finally decides what to buy.

 

Agentic AI compresses much of that journey into a single conversation.

 

Instead of spending thirty minutes comparing smartphones, a consumer could simply instruct an AI assistant to find the best device within a certain budget, prioritise battery life and camera performance, wait for a discount, and complete the purchase once the price falls below a specified threshold.

Google has already outlined this direction through its Universal Commerce Protocol (UCP), Agentic Payments Protocol (AP2) and Universal Cart, which together aim to let AI agents discover products, monitor prices and complete purchases across retailers while keeping the merchant as the seller of record.

 

The same broader direction is visible across the AI industry. OpenAI has expanded shopping capabilities inside ChatGPT, while Amazon continues integrating AI deeper into its shopping experience. Rather than merely answering product questions, these systems are gradually becoming shopping intermediaries.

Why retailers should worry

Traditionally, brands have spent billions of dollars ensuring consumers notice their products. Shelf placement in supermarkets, sponsored listings on marketplaces, homepage banners, influencer marketing and search advertising all exist because visibility influences purchasing decisions.

 

Agentic AI changes who controls that visibility. Instead of a consumer scrolling through dozens of listings, the AI agent may evaluate thousands of products in seconds before presenting only a handful of recommendations. The retailer no longer controls the digital shelf.

 

IDC argues that this effectively replaces emotional and brand-driven purchasing with algorithmic decision making focused on measurable factors such as price, availability, delivery time and product specifications. That creates the risk of reducing shopping into a race to the lowest price, placing further pressure on retail margins.

 

The report also notes that merchants risk losing direct engagement with customers because the AI becomes the primary interface through which buying decisions are made. In other words, retailers could continue fulfilling orders while gradually losing influence over why consumers choose them.

Discovery is already changing

Interestingly, purchasing may not be the first part of commerce that AI transforms. Earlier this year, Sensor Tower found that AI-generated referrals have started appearing as a measurable source of website traffic, although they remain relatively small compared with search and social platforms. Shopping emerged as the largest category receiving AI-driven referrals globally, suggesting users are increasingly relying on AI systems to compare products and evaluate buying options.

 

The report also found that AI traffic currently accounts for less than one per cent of overall web traffic in most markets, indicating that the shift remains in its early stages.

 

This aligns closely with IDC's assessment. Its 2026 Worldwide Retail Consumer Sentiment Study stated that 84.7 per cent of consumers are willing to use digital assistants or chatbots while shopping. Yet only 30.7 per cent currently use AI tools such as ChatGPT to discover or research products, indicating that consumer trust has not fully developed across the complete buying journey.

According to the report, consumers appear comfortable experimenting with AI during product discovery, but remain more cautious about allowing AI to complete transactions on their behalf. However, that distinction may only be temporary as AI systems become more reliable and consumers grow accustomed to delegating increasingly complex tasks.

Retailers may have to sell to machines, not just people

An AI agent is unlikely to be influenced by banner advertisements, colourful product pages or limited-time promotional badges. Instead, it is more likely to compare structured information such as specifications, pricing, delivery timelines, return policies, seller ratings and inventory availability before making a recommendation.

 

That means retailers may need to optimise for two audiences at the same time. One is the human shopper who still values storytelling, branding and visual merchandising. The other is the AI agent that evaluates products using structured data and predefined user preferences.

 

IDC argues that merchants who fail to understand how agentic systems interact with product catalogues and digital storefronts could struggle during this transition. The report also notes that retailers remain uncertain whether agentic buying represents a new sales channel or simply the next evolution of e-commerce, making long-term investment decisions more difficult.

The platform companies gain another gateway

This also strengthens the position of technology companies that already control the interface consumers use every day.

 

For example, Google's vision of agentic commerce is built around open protocols that allow AI systems to move across retailers while remaining anchored within Google's own ecosystem. Consumers can discover products through Search or Gemini, monitor prices through Universal Cart and eventually complete purchases without manually repeating every step of the buying journey.

 

That mirrors what happened with search engines and mobile app stores over the past two decades. Retailers continued selling products, but increasingly relied on platforms they did not control to acquire customers.

 

Agentic AI could deepen that dependence because the AI does not simply direct traffic. It can actively decide which retailer, which product and even which alternative best satisfies a user's request.