Google has an internal manual — public, downloadable — that defines what it considers authoritative when it picks the sources for AI answers. Most business owners don't even know it exists, while the more prepared competitors use it as an operating guide. Worse still: Gemini already lives inside your clients' Gmail inboxes and Drives, and influences them every day. Whoever understands how to move within that ecosystem has a concrete advantage the others don't even see.
Professionals use Gemini in Google Workspace to prepare client responses. If your brand sits in their Google Drive files as a shared asset, the AI cites you from inside their workflows. It isn’t a theoretical possibility: it’s what is already happening inside consulting firms, tour operator agencies, and procurement offices that have integrated Gemini as an operational assistant. And while you think the game to show up in AI answers is played only on public engines, a significant part of it is played inside the Google Workspaces of your business partners.
In this scenario there is one document that becomes more important than it seems: the Google Quality Rater Guidelines. Let me explain.
What the Quality Rater Guidelines are (and why they concern you even if you don’t do SEO)
The Quality Rater Guidelines are the public manual that Google distributes to its human quality raters — the people who assess the quality of search results in order to train the algorithms. The document is downloadable, runs over 170 pages, and contains the criteria Google uses to decide whether a page is “high quality”, “medium quality” or “low quality”.
The criteria are summed up in the acronym E-E-A-T: Experience (first-hand experience), Expertise (competence), Authoritativeness (authority), Trust (trustworthiness). These are not direct ranking factors: they are evaluation principles that train the systems to recognize what is worth showing.
So far we’re talking about traditional search. But here comes the point that changes everything for visibility in AI answers.
The principle: same referee, same league
Google has a single quality-evaluation system. The quality raters who judge search results are the same ones who judge the answers from AI Overviews and from Gemini. The public criteria are the same. The signals that train one system are the signals that train the other.
There is no academic paper that publicly certifies this — it’s an inference, and I’m stating it as such. From the way Google has structured its public documents (the QRG were updated in parallel with the rollout of AI Overviews) and from the way AI Overviews cite sources that match the E-E-A-T profile described in the guidelines, it logically follows that the quality filter for selecting AI sources is derived from the same framework. Same referee, same league.
Translated into practice for you: if you want to be cited by Google’s AI answers, the Quality Rater Guidelines are the most complete optimization manual you have available. Free, public, and maintained by Google itself.
Confusing quantity of content with demonstrated expertise.
What changes for those inside Google Workspace
Let’s go back to the opening scenario. A tour operator in Genoa has integrated Gemini in Workspace to prepare proposals for clients. When one of its clients asks “where can I stay around Savona for a romantic weekend”, Gemini doesn’t answer only with public data: it also draws from the documents shared in the tour operator’s Drive.
If your B&B on the Riviera delle Palme has shared a structured dossier with that tour operator — with description, services, reviews, accreditations, photos, contacts — Gemini cites you from inside that workflow. If you’ve never sent anything, or you sent a generic PDF, you don’t exist in that conversation.
And here the Quality Rater Guidelines stop being abstract and become an operational checklist.
Including the CIR code, VAT number, a verifiable physical address, and links to third-party reviews may seem like bureaucracy, but for Google’s evaluation systems it is the strongest signal of trustworthiness.
Real case: a boutique B&B in Savona
Let me tell you about a case I followed. A boutique B&B in Savona, six rooms, contemporary design feel, run by a couple with a decade of hospitality experience. The problem: stable organic traffic, excellent Booking reviews, but zero presence in AI answers when someone asked for advice on stays along the Riviera delle Palme.
The intervention was built exactly on the four letters of E-E-A-T:
Experience. They prepared a 12-page PDF dossier recounting the stay experience in the first person: breakfast, suggested routes between Noli and Varigotti, the partnership with the local winery, room details. Not a brochure: an insider document with dated and named photos.
Expertise. In the dossier they included concrete credentials: years of management, Liguria Region certifications, the owner’s hospitality training, inclusion in specialized print guides.
Authority. They added links to external sources that talked about them: an article in Il Secolo XIX, a mention on a travel blog, the Chamber of Commerce page with the property’s official data.
Trust. Verifiable contact details, a visible VAT number, the Liguria CIR code, a direct link to the Booking and Google Business Profile reviews, honest indicative prices.
They shared that dossier as a Google Drive asset with seven tour operators along the western Ligurian coast, from the Savona area to the Imperia area. Three months later, when the tour operators’ clients asked Gemini for advice inside the agency’s Workspace, the B&B appeared cited in a structured way in 4 out of 7 monitored conversations. An indicative test, a small sample, but the pattern was clear: dossiers built with E-E-A-T criteria got picked up, the generic ones did not.
The Quality Rater Guidelines as a grid for every shared asset
If you want to apply the same approach, take any document you share with business partners (PDF, presentation, product sheet, dossier) and run it through four filters.
Visible Experience. Does the document show direct, first-person experience, with details only someone who lived it could write? Or is it neutral, brochure-like description?
Declared Expertise. Are the person’s or company’s credentials explicit, with years, certifications, and verifiable qualifications? Or are they generic?
External Authority. Are there links or references to third-party sources that talk about you (publications, institutions, associations)? Or is the document self-referential?
Demonstrated Trust. Are contact details, legal references, honest prices, and verifiable links present? Or does the reader have to take it on faith?
If you answer “no” to two or more, the document is probably invisible to Gemini when it’s pulled inside a partner’s Workspace.
The mistakes I see most often
Thinking the QRG are only for people doing SEO. You apply them to a PDF sent to a tour operator, to a product sheet on a B2B marketplace, to a press dossier: anywhere there’s a digital asset that can end up in a Gemini workflow, the QRG become your grid.
Confusing quantity of content with demonstrated expertise. A 30-page dossier with no declared credentials is worth less than an 8-page one in which you explain who you are, how many years you’ve been operating, and which certifications you hold.
Neglecting the Trust part. Including the CIR code, VAT number, a verifiable physical address, and links to third-party reviews may seem like bureaucracy, but for Google’s evaluation systems it is the strongest signal of trustworthiness.
Not updating shared assets. A dossier sent two years ago and never updated loses trust value over time. Quality raters also assess how fresh the data is.
What you can do this week
Open Google Drive and search for all the documents you’ve shared with business partners in the past year. Take three of them. For each one, run the E-E-A-T grid and mark a binary yes/no for each of the four filters.
Then check your presence in Google Business Profile and on Wikidata: these are the two places where Google looks for external confirmation of your credentials. If you don’t have a Wikidata node that represents you, we covered it in detail in the article on how to get into Google’s Knowledge Graph.
Finally, if you want to understand how the authority perceived by AI systems works at a more systemic level, read what I wrote on E-E-A-T applied to AI and on how AI recognizes authors through Author Entity Recognition.
The Quality Rater Guidelines are not a magic factor. Downloading the PDF and applying it won’t guarantee you AI citations. But it’s the clearest manual you have for understanding the lens through which Google’s systems look at a piece of content, and therefore also the lens through which Gemini decides whether to pick you up or ignore you inside a partner’s Workspace. Visibility in AI answers passes through there too, not just through the public engines.
In the next articles in the series we’ll cover in detail how Gemini handles sources inside Workspace and how to structure a B2B dossier that genuinely gets retrieved by AI systems. The real analysis of your presence requires professional tools, but you can make a first honest diagnosis yourself with just the E-E-A-T grid applied to your shared assets.
