tag
Nile Blog & Research articles filed under geo-aeo.
Your GEO report says mentions are up, but your orders are flat. Product data reaches AI assistants two ways, by crawl and by feed, and different teams own each one. A study of 548 million ChatGPT product offers, published in June 2026, shows which path GEO actually touches, and which one decides the order.
Shoppers ask an AI what to buy, and your product is either in the shortlist or it is not. Testing which is straightforward, but the answer is only as good as the searches you pick. How to draft 30 candidates for a product, cut them to 10 across five query types, and freeze them so the before-and-after means something.
A secondary analysis of 3,602 ChatGPT ad placements from a Penn/Haverford dataset. Paid advertisers appeared in the actual answer text only 8% of the time. After controlling for the question asked, the average lift from paying was −0.3 percentage points. Paying and getting recommended are separate systems.
Four AI agents said an in-stock product was backordered. Fixing the facts stopped the wrong answers, but it did not make the product recommended. A conversation with ecommerce practitioner Leo Nguyen on contested data, hallucinated policies, and the gap between being readable and being chosen.
Once a new commerce behavior becomes measurable, it stops being a future idea and becomes an operating question. Agentic commerce is crossing that line now.
Agentic commerce is not a novelty feature, and it should not be measured like one.
You've spent years optimizing your store for Google. Now someone tells you to worry about 'GEO' too. This one actually matters, and the mechanic is different enough to change what you do next week.
At some point a phrase started circulating: 'You need to get on ChatGPT.' There isn't one lever to pull. 'Being on ChatGPT' can mean three different things.
When a shopper asks ChatGPT for a product recommendation, the agent doesn't browse your store. It retrieves structured, contextual information. Here's what's missing from most pages.
If you've asked ChatGPT for a product recommendation and wondered why it named what it named, you're asking the right question. The answer is not random.
Every optimization decision you've made in the last decade pointed toward one goal: get a person to click. AI agents don't click through. They read, extract, synthesize, and recommend.
