Plain-language commitmentWe separate what we observed from what a company claims, name our commercial interest when Quoted First is included, and do not invent tests, ratings, customers or guarantees.

Who publishes Quoted First research?

Quoted First, Inc. publishes the guides and comparisons on quotedfirst.com. The organization operates an LLM search optimization service, so it has a commercial interest in this category. Pages use an organizational byline—Quoted First or the Quoted First research desk—rather than inventing an individual author. Questions about a page can be sent to hello@quotedfirst.com.

How do we disclose commercial interests?

When a comparison includes Quoted First, the disclosure appears before the comparison itself. Quoted First currently uses no affiliate links in its buyer guides and receives no compensation from the vendors included in them. If either fact changes, the relevant page must say so where a reader can see it before relying on the recommendation.

A Quoted First recommendation is framed as a fit for a defined use case, not as an independent universal ranking. Managed services, monitoring tools, content systems and enterprise platforms solve different operating problems and should not be collapsed into one unexplained score.

Which sources do we use?

Current first-party evidence is the starting point for product capabilities: official documentation, product pages, legal product descriptions, help centers and pricing pages. Reputable third-party reporting may be used for independently testable context, but it is not substituted for a vendor's current description of its own product. Every comparison page should identify its review date and link readers to the primary references used.

We distinguish four evidence labels in the copy:

  1. Vendor-stated: a capability described by the company in a current public source.
  2. Observed: something Quoted First directly checked under stated conditions.
  3. Reported: a claim attributed to a named third-party source.
  4. Judgment: a best-fit recommendation based on the page's disclosed criteria.

How are comparison options selected?

Options are selected to represent materially different buyer needs and operating models, not to produce the longest possible list. The standard evaluation questions are: what job the option performs, who must operate it, whether it preserves prompt and source evidence, whether it helps beyond reporting, which implementation work remains with the buyer, and whether the public information supports the description.

Quoted First does not label a product “tested” unless the page states the test conditions and the product was actually used under those conditions. Public-source research is described as public-source research. A numerical score is not assigned when the compared offerings are not meaningfully interchangeable.

How are pages reviewed and updated?

A comparison is checked against the linked primary sources before publication. The publication or review date appears in the byline and structured data. Material updates change the visible reviewed date and the page's machine-readable modification date. A minor copy edit should not be presented as a new product review.

Search-result observations are dated because search corpora, query rewriting and rankings change. Quoted First's ChatGPT web-search experiment separates corpus inclusion, retrieval and citation, and reports the tool and queries used. It is not represented as disclosure of a private ranking algorithm.

How do corrections work?

Send a correction request to hello@quotedfirst.com with the page URL, disputed statement and supporting source. We review evidence from the affected company and independent primary records. Material factual errors are corrected in the page; changes to a vendor's offering are treated as updates rather than silently framed as an earlier error.

What we will not claim

  • That an agency or tool can guarantee inclusion, ranking, citation or recommendation by ChatGPT or another independent answer engine.
  • That one generated answer proves stable visibility across users, models, locations or time.
  • That structured data, an llms.txt file or repeated deployments are ranking directives.
  • That a paid placement, fabricated review or synthetic community post is independent corroboration.
  • That a public product-page review is a hands-on test.
Disclosed comparison

Compare 10 optimizer options

Research method

See how source selection is tested