A real-world case study exploring why Google AI Overview cited a product page about Bucchero instead of a more comprehensive article. What this may reveal about AI citations and content structure.
A few days ago I came across something that immediately caught my attention while testing Google AI Overview.
I searched for information about Bucchero, the famous black ceramic produced by the Etruscans. The AI-generated answer referenced three different sources, but what surprised me wasn't the topic itself. It was the order in which Google chose those sources.
The first citation pointed to a product page published by Molteni&C.
The second was Wikipedia.
Only further down the search results did I find what was arguably the most complete article on the subject, published by Krei.
At first glance, this doesn't seem to make much sense. If one article explains Bucchero in far greater detail, why would an AI-generated answer prefer a furniture company's product page?
Before trying to answer that question, it's worth looking at the three sources.
The first source is Molteni&C's page dedicated to the Bucchero table designed by Gio Ponti.
https://www.molteni.it/it/it/prodotto/bucchero
Although the primary purpose of the page is to present a product, it also includes a concise explanation of what Bucchero is, where it originated, why it has its characteristic black color and how this ancient ceramic technique inspired the design.
The second source is the Italian Wikipedia article.
https://it.wikipedia.org/wiki/Bucchero
As expected, Wikipedia provides a broad historical overview, describing the origins of Bucchero, its archaeological importance, production techniques and cultural context.
The third source is an in-depth article published by Krei.
https://www.krei.it/focus/bucchero-il-cotto-nero-degli-etruschi/
From a human reader's perspective, this is arguably the most comprehensive explanation. It explores the history, production process and significance of Bucchero in much greater detail than the other two pages.
So why wasn't it the first source used by AI Overview?
The honest answer is that we don't know. Google has never published the exact criteria used to select sources for AI Overview, and it almost certainly relies on many different signals rather than a single ranking factor.
However, this example suggests a few interesting possibilities.
The first is that AI systems may value information density more than article length. The Molteni page communicates several key facts in a relatively small amount of text. Every paragraph is directly related to Bucchero, its characteristics and its inspiration. There is very little surrounding content that could dilute the main topic.
The second possibility is entity trust. Molteni&C is a globally recognised design brand with a strong web presence. Even though the page is commercial, Google may consider it a reliable source for explaining the inspiration behind that specific product.
The third possibility is contextual relevance. The page doesn't try to explain every aspect of Etruscan history. Instead, it focuses on a small number of well-defined concepts that are closely connected to the object being presented. For an AI system generating a concise answer, that kind of focused context may be easier to reuse.
This case also highlights something that many content creators may need to reconsider.
For years, SEO discussions often revolved around writing the longest and most comprehensive article possible. That strategy can still be valuable, but AI-generated answers may sometimes prefer content that is easier to interpret, reconstruct and cite.
That doesn't necessarily mean shorter content is better. It simply suggests that clarity, structure and well-defined concepts may become increasingly important alongside completeness.
This single example doesn't prove a universal rule. It is only one observation, and many other searches may produce completely different results. But it raises an interesting question that deserves further investigation.
Perhaps the future of AI visibility won't belong only to the longest articles.
Perhaps it will belong to the pages that communicate knowledge in the clearest and most reusable way.
At NetContentSEO, we'll continue collecting real-world examples like this one. Rather than relying on assumptions, our goal is to understand how AI search engines actually choose, interpret and cite information across different types of content.