writing
How answer engines read websites
Each post answers one question our own audits raise, and cites the primary documentation it rests on. Where the evidence is thin, it says so.
- 9
- posts
- 27
- sources
all posts
- 01
Does llms.txt actually do anything?
No assistant has confirmed it reads llms.txt, and the measured request data is thin. Here is what the evidence supports and what it does not.
- 02
Who does an AI assistant think you are?
Assistants answer about entities, not URLs. If your name, domain and profiles do not agree, you are several weak entities instead of one.
- 03
Do publication dates and sitemaps change whether AI assistants cite you?
Dates and sitemaps are how a crawler knows what changed. They are cheap, checkable, and routinely wrong on real sites.
- 04
How do I write a page an AI assistant can quote?
Answer in the first sentence, put the question in the heading, and keep each answer self-contained. Retrieval works on passages, not pages.
- 05
What robots.txt mistakes quietly block AI search?
Groups do not inherit, matching is by prefix, and the longest rule wins. Each of those trips up a rule that looks correct.
- 06
Does structured data help AI assistants quote my pages?
Structured data states plainly what a page is about. It is not a ranking trick, and what it can and cannot do is documented.
- 07
What is answer engine optimisation, and how is it different from SEO?
AEO is about being quotable inside an answer, not ranked beside one. The mechanics differ, and so does what you should measure.
- 08
Which AI crawlers can actually reach my site, and how do I check?
robots.txt states an intention. The status code your server returns is the fact. They disagree more often than owners expect.
- 09
Why does my site look empty to AI assistants when it works fine in a browser?
Most AI crawlers do not run JavaScript. If your text arrives via the browser, they see an empty page — and that is measurable in seconds.