Learn with the research

The Gap Between Being Cited and Being Heard

1. The Concept, Plainly

Getting cited by an AI and having your words shape its answer are two different things. When an AI lists your page as a source, that is citation selection. When it actually draws from your content to write its answer, that is answer absorption. A 2026 study by Zhang et al. analysed 21,143 citations across three major AI platforms and found that the two events have different causes and different levers. Being cited does not mean being absorbed.


2. Why This Matters Right Now

AI answers are now a primary way readers get information. If you publish content and want your words to reach readers through AI, citation is not the end goal — absorption is. The gap between the two is measurable. The AI Answer Absorption Analyser at psytable.com scores your content against six structural dimensions that distinguish high-absorption pages from low-absorption ones. Understanding how the two-stage mechanism actually works will make your score interpretable — not just a number, but a map of what to change.


3. The Mechanism

Every time an AI generates an answer to a question, it runs two distinct processes. The first is citation selection: the AI decides which sources are relevant enough to list. The second is answer absorption: which of those sources actually shape the words, claims, and structure of the response. Many sources get selected. Fewer get absorbed.

The Zhang et al. (2026) study examined 21,143 citations across ChatGPT, Google AI Overviews, and Perplexity. It classified pages into high-influence and low-influence groups based on whether the AI answer visibly drew from them — not just listed them. The structural profile that separated the two groups is where the research gets concrete.

High-influence pages were on average 11.44 times longer than low-influence pages. They had 12.5 times more headings. They had 5.69 times more paragraphs. These are not small differences. They describe structurally different documents. A piece of content that is short, minimally headed, and paragraph-sparse is structurally invisible to absorption — the AI may cite it, but has little to draw from.

Content properties mattered too. Pages with strong definitional content showed approximately 57% higher absorption. Definitional content means sentences that explain what something is: statements like "X is a process of..." or "X refers to..." or "X is defined as...". Comparative content — sentences using "unlike," "compared to," or "whereas" — was associated with approximately 55% higher absorption. Statistics presence correlated with approximately 61% higher absorption in the same study. Having statistics in your content — specific numbers, percentages, or quantified findings — is associated with higher absorption than content without them.

One finding cuts against common optimisation advice. Q&A formatted sections showed approximately 5.74% lower relative absorption compared to narrative structure for ChatGPT. FAQ formatting may help with search visibility, but the Zhang et al. data points in the opposite direction for absorption.

The practical picture: citation selection responds to relevance — is your page about the right topic? Answer absorption responds to structural richness, definitional clarity, and comparative depth — does your page give the AI enough well-formed content to draw from? You can influence both. They require different content decisions.


4. Try It Now

You can see the gap between citation and absorption directly. This exercise uses two prompts in sequence.


5. The One Thing to Remember

Citation selection and answer absorption are two separate events with different causes. Getting into the sources list does not mean your words shaped the answer.


6. Go Deeper

The Field Notes post on the Zhang et al. 2026 study covers the full methodology, the platform-by-platform variance in findings, and the preprint status caveats: There Is a Difference Between Getting Cited and Getting Used.

See where your content lands on absorption.

The AI Answer Absorption Analyser scores your content against six structural properties from Zhang et al. 2026 — word count, headings, paragraphs, definitional language, comparative language, and statistics. No signup.