Spoke
Formatting for extraction
Six steps in a paragraph and six steps in a list carry the same information to a reader. To a machine they're nothing alike, and that difference decides whether you get quoted precisely or approximately.
That semantic structure is easier to parse than prose follows from how parsers work. No platform publishes a formatting specification, and we claim none.
Six steps written in a paragraph and the same six steps in a numbered list carry identical information to a human being.
To a parser they're nothing alike. The list says out loud that there are six discrete items, what order they're in, and where each one starts and stops. The paragraph says: here is a block of text.
That difference decides whether something can quote you precisely or has to approximate.
Why do lists and tables help with AI visibility?
Because they expose structure that prose only implies. A list states that these are discrete, parallel items in this order. A table states that these things differ along these named attributes. Anything quoting from either can take an exact item or an exact cell and attribute it correctly.
From running prose it has to infer the same structure — slower, lossier, and far more likely to produce a paraphrase that misstates you.
Pick the structure the information actually has
- A sequence — numbered list. Steps, phases, priority order.
- A set — bullets. Options, signals, symptoms.
- A comparison across attributes — a table. Anything where you'd otherwise write "whereas."
- A rule and its exception — a two-column table.
- An argument — prose. Reasoning doesn't survive being chopped into bullets.
That last one is the counterweight to this whole page. Converting everything to bullets destroys exactly the connective reasoning that makes content worth quoting in the first place. Match the structure to the content. Don't maximize bullets.
Before and after
Dense:
Core Web Vitals are assessed on three metrics, measured at the seventy-fifth percentile of real page loads rather than in a lab, with separate assessments for mobile and desktop, and a page passes only when all three are in the good range.
Extractable:
Core Web Vitals are assessed on three metrics. A page passes only when all three are in the good range.
>
- Measured on real visits, not lab tests. - Taken at the 75th percentile. - Assessed separately for mobile and desktop.
Same facts. The second version lets something take the rule and the three conditions as separate, quotable pieces.
Use the real elements
A visual table built out of styled div elements looks identical to a reader and means nothing to a parser. <table>, <th>, <caption>, <ol>, <dl>, <figure> and <figcaption> all carry relationships a machine can use.
The most common own goal here is the designed comparison graphic. A genuinely useful table, exported as a PNG, sitting on the page as an opaque rectangle. Everything that made it quotable got thrown away in the export.
Takeaway boxes
Every teaching page here opens with one. It serves three audiences at once: the reader who won't read the whole page, the reader deciding whether to, and whatever's looking for a compact statement of what the page claims.
It also disciplines the writing. If you can't state the core claims in four or five lines, the page is usually trying to do too much and wants splitting.
More in AEO
Answer-first writing
The first sentence under a question heading is the only part guaranteed to be read, quoted or extracted. Here is how to write one that works with nothing around it, and the before-and-after that shows the difference.
Question headings
A heading tells a reader what's coming and tells a machine what the passage is about. Those used to pull against each other. Mostly they don't anymore.
Snippets & voice
Before AI answers there were snippets and voice assistants, which also had to pick one answer instead of listing ten. Everything this section teaches was worked out against those.
Question markup
FAQ rich results were deprecated in May 2026. The FAQPage type wasn't. Knowing which of those is true saves you from the stale advice and from the overcorrection.