ChatGPT Shopping and the new logic of e-commerce: what will change in 2026

ChatGPT, Perplexity, and Google AI Overviews no longer just answer questions. They shape demand, select products, and initiate purchases. For e-commerce, this means one thing: the funnel is narrowing, and most of the decision-making now takes place before the user even visits the website.

But this is not the “death of SEO” or the disappearance of online stores. It is a change in roles – and this is worth discussing in more detail.

How the buyer’s journey has changed

Previously, the journey looked like this: search query → list of sites → comparison → selection.

Now: dialogue with AI → ready recommendation → confirmation or correction.

Instead of “where to buy iPhone 15 Pro,” the user asks “recommend a smartphone for photos under $1,000” – and gets 1–3 options instead of 20 links. It is at this moment that the AI assistant becomes the first seller.

This pattern is especially noticeable in the categories of electronics, household appliances, and home goods. A query that would be considered informational in classic SEO becomes highly convertible in the AI interface.

Three platforms – three different roles in the funnel

It is a mistake to perceive all AI tools as one channel.

ChatGPT Shopping forms preferences and shortlists through dialogue with clarifying questions. Google AI Overviews provides an initial guide directly in the search results. Perplexity Instant Buy shortens the path to transaction by combining AI search with sources and payment.

Each platform intercepts the user at its own stage. A store that is not on any of them simply falls out of this chain.

An alliance that confirmed the direction

Ahead of Black Friday 2025, Perplexity announced a partnership with PayPal with technological support from OpenAI. The result is purchases directly from AI search, without going to the merchant’s website.

Why is this important? PayPal serves over 430 million active accounts, and more than 75% of transactions occur with already saved user identification. Accelerated checkout increases conversion by 15–30%, and the rate of abandoned carts decreases by 20–25%. In AI interfaces, this effect is even greater: the time from request to purchase is reduced by 2–3 times.

This is not a separate feature of Perplexity. It is a demonstration of a new architecture: AI interface + payment provider + store data = a single chain without unnecessary steps.

From SEO to GEO: the new logic of visibility

Classic SEO competed for positions. Generative systems do not “rank” – they select sources that can be trusted.

This is the transition to GEO (Generative Engine Optimization). For AI, it is not important how well a page is optimized for a keyword, but how structured the data is, how relevant it is, and how stable the source is.

The practical consequence: stores with less traffic but well-designed product cards are more likely to appear in AI recommendations than large sites with “creative” but poorly structured content.

Brand visibility is now determined not by position in search results, but by appearing in AI responses.

What AI considers when making recommendations

Four critical signals that determine whether a product will be included in a recommendation:

• Structured data. Schema.org microdata (Product, Offer, Review) is the basic threshold. Without it, AI either ignores the source or uses it fragmentarily.
• Completeness of attributes. The fewer clarifications AI needs, the higher the chance of a recommendation. Important characteristics, usage scenarios, restrictions, delivery terms.
• Reviews and external mentions. AI checks reputation outside the site – marketplaces, reviews, thematic resources. This reduces the risk of false recommendations.
• Relevance of prices and availability. Outdated data is one of the strongest negative signals. Contradictory information about price or availability excludes the store from recommendations even if the content is of high quality.

Product card: a source of data, not just a tool for persuasion

In the AI shopping model, the product card solves two tasks at once: it persuades the buyer and “feeds” the AI with data.

The ideal card for AI contains: a brief description indicating the type of product, target audience, and usage scenario; structured characteristics without marketing wording; a block of scenarios (“suitable for apartments up to 60 m²,” “optimal for travel”); reviews with an aggregated rating; FAQ in machine-readable format.

Usage scenarios are the most underrated block. They determine whether a product will appear in response to a query “what to choose for…”.

Technical minimum for participation in the AI funnel

Element	What’s Needed	Why It Matters
Structured Data	Product, Offer, Review	AI clearly understands what is being sold
Product Feed	Up-to-date merchant feed	Eliminates errors and outdated data
Data Updates	Price and availability every 1–2 hours	Outdated data leads to loss of trust
Consistent Attributes	Same values everywhere	AI dislikes inconsistent formats
Clean HTML	Minimal hidden JavaScript	AI reads dynamic content worse

Without this minimum, SEO and content simply do not participate in the new funnel, regardless of brand or budget.

New metrics instead of CTR and positions

In the AI funnel, classic metrics lose their informativeness. They are replaced by:

• Coverage – how fully the assortment is represented in AI responses
• Trust signals – data stability, reviews, payment partners
• Conversion efficiency – number of steps to purchase
• Revenue impact – contribution of AI channels to revenue, not traffic

The focus shifts from “how many clicks we got” to “how quickly the user made a decision.”

There is also a “gray area” – things that the market is not yet able to measure: the frequency of brand exposure in AI responses, accurate attribution of instant purchases, and reasons for choosing one seller over another. This requires indirect analytics and testing.

Conclusion by MIM:AGENCY

AI purchases are not just another tool in digital marketing. They represent a change in the very logic of decision-making.

The winners are not those with more traffic, but those with better data and trust infrastructure. SEO is no longer an isolated channel but becomes the foundation for AI recommendations, shortened funnels, and new conversion points.