AI Search & Commerce News
Timely industry and technology developments across ChatGPT, Perplexity, Google AI, Shopify, and cross-border commerce—framed for GEO operators.
87 briefings
AI shopping assistants now use return and warranty terms as decision filters, not fine print
Assistants are surfacing return windows and warranty facts inside shortlists, making policy wording a competitive citation surface for merchants.
Read briefingAssistants Are Naming a Confidence Level Next to Price — and It Comes From Merchant Page Text
Price is no longer reported as a bare number. Assistants increasingly attach a confidence qualifier to it, and whether they call your price current or approximate is decided by how your page states it.
Read briefingDelivery Windows and Order-By Dates Are Becoming Deciding Facts in AI Shopping Answers
Assistants increasingly attach delivery timing to recommendations, and they lift it from merchant page text when it is stated as a commitment rather than a disclaimer.
Read briefingShopping Agents Are Dropping Products They Cannot Confirm, Not Products That Are Out of Stock
Agents increasingly filter candidate lists by data confidence rather than availability — and the merchants losing placements are the ones whose remaining unknowns sit in unstructured text.
Read briefingAssistants Are Adding Caveats — and Those Caveats Come From Merchant Page Text
Assistant recommendations increasingly carry a qualifier naming when the advice stops holding. Those boundary conditions are retrieved from page text, which makes them a new kind of product fact to publish.
Read briefingAssistants Are Answering Category Questions With Shortlists — and There Is No Page Two
Assistants increasingly answer category questions with two or three named candidates rather than a page of links. For merchants, the cost of not being named is absence, not a lower rank.
Read briefingAssistants Are Answering Fit Questions Directly — and Image-Only Size Charts Are Invisible to Them
Fit is the one attribute buyers cannot judge from a photo, and the one most often published only as an image. Assistants do not read pixels, so an unstated measurement is an unanswerable question.
Read briefingPlatforms Are Rebuilding Commerce Around Agent APIs — and That Moves Differentiation to the Text Layer
Agent-callable catalog and checkout APIs normalize product presentation. Uniform records mean tiebreakers shift to the text an agent reads around the API row.
Read briefingShopping Agents Are Now Judged on Return Rates — and Who Gets Cited When They Explain Why
Agentic shopping pilots now weigh post-purchase outcomes, not just conversion. That turns your returns and warranty pages into citation sources.
Read briefingAI Assistants Now Favour the Most Consistently Stated Product Facts
When price, shipping, or warranty wording differs across channels, assistants hedge or pick arbitrarily — consistency is becoming a citation trust signal.
Read briefingWant playbooks instead of headlines? Visit the GEO Knowledge Hub