What we refuse to sell, and the evidence for each refusal.
Every item on this list is currently sold by someone in this category. We are not claiming they are dishonest people. We are saying that the evidence does not support the product, and that we checked.
1. Dofollow links
Google’s link spam policy names “links with optimized anchor text in articles or press releases distributed on other sites” as an unnatural link. Selling them transfers manual-action risk onto your domain.
Sourced Google Search Central. Full policy.
2. Guaranteed AI citation
No provider controls what an engine returns. Answers vary by query, by model version, by account, by date. A guarantee here is either a misunderstanding or a lie, and we cannot tell which from the outside.
Vendor claim, no methodology EIN Presswire markets releases as “ingested by prominent AI and LLM infrastructures” with no platform named and no report delivered. ACCESS Newswire sells an “LLM Citation Score” with no published rubric. Both read from their live pages, 2026-09-05.
3. A place in “best X” AI answers
Press-release domains were measured at 0% across 100 category queries in 10 verticals on Perplexity, ChatGPT and Gemini. That is why we sell editorial coverage in trade publications, not a page on a wire, and we still do not promise a place on that list.
Sourced Loganix, March 2026. The full evidence.
4. llms.txt as a paid deliverable
Google’s own documentation states “Google Search itself doesn’t use them” and that keeping one “will neither harm nor help.” An Ahrefs study of 137,210 domains found 97% received zero requests during May 2026.
We publish one on this site because it is free and coding agents read it. We do not bill for it and we do not list it as a visibility lever.
Sourced The full evidence.
5. Schema markup sold as an AI-citation lever
The best-controlled study, 1,885 pages that added JSON-LD, measured difference-in-differences against ~4,000 controls, found no meaningful effect. Every rank-controlled study finds null or slightly negative.
We do implement structured data, because Google documents that it helps with disambiguation and rich-result eligibility. Those are the reasons we give. They are not AI-citation reasons.
Sourced The full evidence.
6. A Google Knowledge Panel
Google: knowledge panel creation “happens automatically and isn’t something we can or would influence.” There is no documented request process and no structured-data property that triggers one.
The sameAs property is frequently sold for this. The Google feature it refers to was
deprecated on 2019-06-25 and its documentation page now returns a 404.
Sourced Google Knowledge Panel help documentation.
7. Releases we would not want on our own domain
Unsubstantiated medical or financial claims. Announcements with no actual news event. Anything we believe to be false. We decline these and return the money, because a domain that will publish anything is worth nothing to publish on.
If we are wrong about any of this
Every item above carries a source you can check. If you have a primary source that contradicts one of them, send it to hello@aivisibilitywire.com. We will read it, and if it holds we will change the page and say what changed. A refusal list that cannot be updated by evidence is dogma, not method.