This page exists so you can judge our work without trusting us. It covers where our facts come from, how we check them, what we label, and how we correct mistakes.
Sourcing
We start from primary material: Google Search Central, WordPress and browser engine dev notes, Bing and IndexNow announcements, schema.org releases, crawler and registry documentation, and published datasets. Every article links the source it argues with.
We deliberately avoid syndicated news. A story reprinted by two hundred sites adds nothing, and rewriting it adds less.
How we treat vendor research
Most published data about AI search comes from companies selling AI-visibility tools or consulting. We use it, but only when the study states all seven of the following. If it does not, the number does not appear here as a fact:
- which model
- consumer interface or API
- date of collection
- region
- the prompt set
- number of repetitions
- observed stochastic variation
Verification
Claims that can be checked, we check: response headers, robots directives, markup validity, page weight, crawl behaviour, index status. When we quote a number we say where it came from and on what date. When we measured it ourselves we say what we measured, on what, and how many times.
What our labels mean
- Update — something changed; what, when, and what to do about it.
- Explainer — how a mechanism works, for people who have to act on it.
- Deep dive — a long piece with its own research or its own table.
- Data study — our own measurement, with method stated.
- Teardown — a specific page, site or feature taken apart.
- Opinion — an argument, clearly labelled as one.
Updates and corrections
Search changes weekly, so articles are updated rather than duplicated, and the update date is shown on the page. Corrections are made in place and noted at the bottom of the article. We do not quietly edit a mistake away.
Found an error? Write to hello@seoafterai.com with the article link and the source you think we got wrong. If you are right, we fix it and say so on the page.
AI disclosure
Drafting is assisted by language models. Every article is edited, fact-checked against its source and approved by a human before publishing. Nothing is published unread. Where a model produced a claim we could not verify, the claim is cut rather than softened.
Independence
We publish no sponsored posts, paid links or paid reviews. Nobody pays for coverage or for a mention. If that ever changes, sponsored material will be labelled as such on the page itself, not only in a policy.