# How AI assistants decide which sources to cite

> When ChatGPT, Claude, or Perplexity answer a question, only a handful of sources end up cited. The selection isn't random: AI assistants weight roughly five things (entity clarity, structured data, content shape, citation graph, recency), and most of them are within a publisher's control. Here is how each factor works, how answer engine optimization differs from SEO, and what we changed on our own site to be more citable.

URL: https://voicemoat.com/blog/how-ai-assistants-pick-sources

## Key facts
- Published: 2026-05-11
- Category: AI and Voice
- Read time: 13 minutes
- Author: VoiceMoat team
- Publisher: VoiceMoat
- Topics: Answer engine optimization; AI search citations; Structured data; Schema.org markup; AI writing tools

## Key points
- Entity clarity: can the AI tell what you are?
- Citation graph: who else points to you?
- Recency: when was this written, and when was it updated?
- Is answer engine optimization the same as SEO?
- Does the voice of your writing affect whether AI cites it?

## Related articles
- Answer engine optimization: a 2026 field guide: https://voicemoat.com/blog/answer-engine-optimization-2026
- Alt-text on X: the AEO move most creators skip, done in voice: https://voicemoat.com/blog/alt-text-twitter-aeo-and-voice
- The 10 signals of voice: what actually makes writing recognizable: https://voicemoat.com/blog/voice-dna-9-dimensions-canonical
- Can your audience tell you're using AI? An honest 2026 analysis: https://voicemoat.com/blog/can-audience-tell-youre-using-ai
- Twitter for lawyers: how to build authority on X without sounding like every other JD: https://voicemoat.com/blog/twitter-for-lawyers

## FAQ
Q: How do AI assistants decide which sources to cite?
A: They weight roughly five things: how clearly you identify yourself as an entity, whether your page carries structured data, the shape of your content, the citation graph pointing to you, and how recently the page was updated. Each is something a publisher can influence, and none is search engine optimization in the old sense. When only a handful of sources end up cited, these factors determine which pages make the list.

Q: Why does entity clarity matter for getting cited?
A: When an assistant cites a source it is really citing an entity, a person, organization, product, or article, not just a URL. If it cannot tell what entity a page represents, citing becomes risky because it might attribute the wrong claim to the wrong party, the kind of error these systems are trained to avoid. Consistent identity across the web, Schema.org markup with a stable @id, and knowledge-graph presence give the assistant confidence to attribute you cleanly.

Q: Does the voice of your writing affect whether AI cites it?
A: Indirectly yes, and the mechanism runs opposite to the intuition. Assistants cite a passage because it is groundable, specific, and attributable to a clear entity, not because it sounds human. Generic, voice-flat content is paradoxically harder to cite well because it reads as interchangeable with a hundred other pages making the same claim. Writing with a recognizable voice and first-hand claims gives the assistant something only your page says, which is exactly what a citation needs to point at.

Q: Is answer engine optimization the same as SEO?
A: Mostly no, and the difference is the unit of competition. SEO competes for a ranked link in a list of ten, while AEO competes for a sentence inside a synthesized answer with a citation next to it. The retrieval step is closer to retrieval-augmented generation than to classic ranking, so you optimize a passage to be the cleanest, most citable, most entity-clear source for one specific claim rather than optimizing a page to outrank nine others on a keyword.

Source: https://voicemoat.com/blog/how-ai-assistants-pick-sources
Site overview: https://voicemoat.com/llms.txt. AI usage policy: https://voicemoat.com/ai.txt.