Once a page can be crawled and understood, search engines and AI assistants still have to decide whether it deserves to rank or be quoted. This guide covers the qualities that make that decision go your way: new information, visible experience, evidence, clear sources and trust. Two sections describe Brilad working concepts rather than industry standards, and they are marked as such.
Information gain
Information gain is the amount of new information a page adds compared with other pages on the same topic. A Google patent granted in 2022, on the contextual estimation of information gain, describes scoring documents by how much they add beyond what a user has already seen. A patent does not prove a ranking system uses it, but Google’s own guidance for creating helpful content asks the same question in plain words: does the page provide original information, reporting, research or analysis?
For AI retrieval the effect is easy to see. When ten pages say the same thing, the assistant can merge them into one sentence and credit any of them. When one page contains a fact the others lack, that page becomes the source for that fact.
Ways to add information gain: your own figures, prices or benchmarks; results from projects; a clear method others can follow; an honest statement of limitations; a comparison nobody else has made in a table.
E-E-A-T
E-E-A-T stands for experience, expertise, authoritativeness and trustworthiness. It comes from Google’s Search Quality Rater Guidelines, the document Google gives to the people who rate search results for quality. Google added “experience” in December 2022 and describes trust as the most important of the four.
E-E-A-T is not a single ranking factor that can be measured. Raters do not change rankings directly; their ratings help Google test its systems. What the guidelines describe is what Google’s systems try to reward, so the practical steps follow from them:
- Name the author of each article and link to a page about them with their real role and background
- Show first-hand experience in the content: what you did, what you saw, what went wrong
- Make it easy to find who runs the site, how to contact them and where the business is
- Keep facts about the company identical across your site and your profiles elsewhere
First-party evidence and original research
First-party evidence is material that only you can publish because it comes from your own work: data from your projects, survey results, test results, screenshots of real dashboards, anonymised client outcomes. Original research goes a step further with a defined question, a method and a dataset.
The 2023 GEO research paper by Aggarwal and colleagues found that adding statistics and quotations from sources to a page improved its visibility in AI-generated answers in their tests. The benefit depends on the numbers being real and verifiable. Invented or unsourced figures damage trust when someone checks them, and AI assistants repeating a wrong number attributed to you is worse than not being cited. See the guide to first-party evidence for how to publish your own data properly.
Proof of work
Claims like “we are experts in technical SEO” carry little weight with readers or ranking systems. Showing the work does: a before-and-after crawl report, the robots.txt change you made and its effect in server logs, the schema you wrote, a screenshot of an AI answer that now cites a client. On this site, the home page shows the live llms.txt file the site generates, for the same reason.
Source attribution and content provenance
When a page states a fact that came from elsewhere, it should name and link the source. This helps readers check the claim, and it helps AI systems tell your own findings apart from repeated facts.
Content provenance refers to recording where a piece of content came from and how it was changed. The C2PA standard, backed by Adobe, Microsoft, Google and others, attaches signed “content credentials” to images and media for this purpose. It is mainly relevant today for publishers of photos and video, but it shows the direction: systems increasingly want to know who made something.
Trust signals
Google’s rater guidelines treat pages about money, health, safety and major decisions (called “Your Money or Your Life” topics) with the highest standard of trust. Business services pages are judged by similar common-sense signals:
- A real company name, address and contact details
- Named people responsible for the business
- Clear information on what the service includes and, where possible, what drives the price
- Honest statements about what cannot be promised
- Reviews and references on sites you do not control
The trust signals guide covers these in detail, including honest risk reversal.
Quotability (a Brilad working concept)
We use the term quotability for how easily a single passage can be lifted out of a page and still be correct and complete. It is our own checklist, not a metric published by any search engine. A quotable passage names its subject instead of using “it”, states one claim, includes the number or condition that makes the claim true, and does not rely on the sentence before it.
Compare “It usually takes a few weeks” with “Google typically recrawls a changed page within a few days to a few weeks, depending on how often it crawls the site”. Only the second can be quoted on its own. The full six-question check is in the quotability guide.
Brand mentions and co-occurrence
Language models learn associations from the text they are trained on and retrieve. If your company name appears often next to a topic on other websites, in articles, forums, directories and reviews, a model is more likely to connect the two. This includes mentions that do not link to you.
No search engine or AI company publishes how much weight such mentions carry, so this section describes our view of how the systems work rather than a documented rule. The practical conclusion is ordinary PR: be mentioned in the right places, by name, in connection with the work you want to be known for. More in the brand mentions guide.
Conversion-ready pages (a Brilad working concept)
Being cited is only useful if the person who clicks through can act. We check that each page an assistant is likely to cite gives the visitor the next step without friction: what the service costs or how pricing works, what happens after they get in touch, and a contact route that does not require a long form. This is our own working checklist rather than a search ranking factor, but it decides whether visibility turns into enquiries. See the full conversion-ready checklist.
In-depth guides
- Information gain: why original content gets cited
- E-E-A-T for AI search
- First-party evidence and original research
- Trust signals and risk reversal
- Quotability (a Brilad concept)
- Brand mentions and co-occurrence
- Conversion-ready pages (a Brilad checklist)