How AI Shopping Assistants Decide What to Recommend to Shoppers
AI shopping assistants like ChatGPT, Perplexity and Amazon’s Rufus decide what to recommend by reading structured product data, not by browsing your store the way a person would. If your product feed is thin, inconsistent or missing key fields, an assistant will simply skip your items and recommend a competitor’s instead.
AI-driven shopping is no longer a side channel
A growing share of purchases now start with a question typed into an AI assistant rather than a search box. These systems research options, compare prices, and increasingly complete checkout on the shopper’s behalf, acting less like a search engine and more like an always-available personal shopper.
AI-referred traffic to online stores nearly quadrupled in a year. Adobe Analytics tracked more than one trillion visits across over 130 major North American retailers and found AI-referred traffic grew 393% year over year in the first quarter of 2026, with those AI-driven shoppers converting 42% better than visitors arriving through other channels.
Source: Adobe Analytics, AI-Sourced Traffic Insights
A single AI shopping assistant now touches hundreds of millions of buyers. Amazon’s Q4 2025 earnings materials disclosed that its assistant, renamed Alexa for Shopping in 2026, had reached more than 300 million customers and was driving close to $12 billion a year in incremental sales.
Source: Amazon, Q4 2025 earnings release and SEC filing
Shoppers who arrive through an AI assistant spend more, not less. Shopify’s 2026 commerce data found that visitors referred by AI platforms convert nearly 50% higher than organic search visitors, and shoppers coming from Perplexity specifically place orders worth 57% more than those from other AI platforms.
Source: Shopify, AI Search Insights
Why most product catalogs are invisible to an agent
An AI agent does not scroll a page or admire a photo. It reads a feed: brand, GTIN, price, availability and shipping terms, matched against what the shopper asked for. When that data is missing, inconsistent with the live page, or written as marketing copy instead of plain facts, the agent has nothing reliable to compare. In one production audit of a mid-sized US store, AI shopping assistants ignored more than 40% of the catalog for exactly this reason, not because the products were wrong, but because the assistant could not confirm they existed.
How to get your products picked
Fill every core attribute, not just the required ones. Brand, GTIN or MPN, material, size, color and current stock status should be present and identical across your feed, your product page and your checkout. An agent cross-checks these fields, and a mismatch reads as unreliable data, not a small typo.
Write facts before adjectives. Lead with concrete specifications, then add persuasive copy after. Agents extract facts far more reliably than tone, and a description that answers what is this, exactly gets recommended more often than one that only sells a feeling.
Add Product, Offer and Review schema to every product page. Structured data in JSON-LD gives an assistant a machine-readable version of the same facts on the page itself, as a backup to your feed and a signal that your data is current.
Most store owners still optimize for a person scrolling a page. But a growing share of buying decisions now happen without a single page view, an agent reads your data and decides for the shopper. If your feed is thin or inconsistent, you are invisible to that decision no matter how good the product actually is.
About Patrik Vavrovič
Getting picked by assistants rests on the same foundations as classic search. For the full playbook, see our complete guide to ecommerce SEO.
Key takeaways
AI shopping assistants like ChatGPT, Perplexity and Amazon’s Alexa for Shopping now influence a fast-growing share of purchases, with AI-referred traffic to US retailers up 393% year over year in Q1 2026.
These assistants read structured product data, not page design, so a strong photo or clever copy does not help if the underlying feed is thin or inconsistent.
Shoppers referred by AI assistants convert better and spend more, not less, according to Shopify’s 2026 commerce data.
Filling core attributes, leading with facts, and adding Product schema are the fastest ways to become recommendable to an agent.