Resources · Guide · Free download
HOW AI ASSISTANTS PICK THEIR SOURCES
Why some sites get cited by AI assistants and others never appear. Observed patterns, labelled honestly.
What the answer engines appear to reward, structure, specifics, authority and access, and where the honest limits of that knowledge sit. Patterns, not published rules, and they shift.
If you would rather have this done with you than to you, this is the work I do. See services
Observed patterns, labelled as observations
Nobody outside the model providers knows the ranking rules, and anyone claiming certainty is selling something. This guide sets out what is repeatedly observable across assistants: the preference for content that can be read without JavaScript, the weighting of corroboration across independent sources, and the effect of a clean, consistent entity. Where something is inference rather than observation, it says so.
Why your own blog is not enough
A claim that appears only on your own site is a claim from an interested party. The same claim on a publication, a client's site, a directory and a conference listing is corroborated. Most businesses over-invest in the first kind and wonder why they are never cited, which this post covers in full.
COMMON QUESTIONS
How do AI assistants choose which sources to cite?
They favour pages they can read without executing JavaScript, claims corroborated across independent sources, and entities that resolve cleanly to one identifiable business. Exact weighting is not public.
Does publishing more content get me cited?
Only up to a point. Volume on your own domain establishes what you say about yourself. Being mentioned somewhere you do not control is what makes an assistant treat the claim as verified.
Can I pay to appear in AI answers?
Not in the organic answer itself. Some products are introducing advertising alongside answers, which is a separate placement. Anyone selling guaranteed inclusion in the answer is describing something that does not exist.