E-E-A-T for AI Overviews: Trust Signals That Get You Cited
We checked 158 AI Overview citations for Australian service questions. Here's which trust signals the cited pages had, and how to add them.
TL;DR
E-E-A-T isn't a ranking factor in itself, but the pages Google's AI Overviews cite carry its signals. In our check of 158 AI Overview citations for Australian service questions, 94% of cited pages were dated and 81% had a named author. And 61% of citations went to the service providers' own sites, so a local firm can earn the citation directly.
We pulled every page Google's AI Overviews cited for 21 Australian questions about mortgage brokers, buyers agents, lawyers, accountants, dentists, builders and SEO. Then we checked each one for the E-E-A-T signals that every checklist recommends.
Most E-E-A-T SEO advice is a checklist with no evidence behind it. This post shows which signals the cited pages actually had, then turns them into a 30-minute check you can run on your own page.
Google E-E-A-T: what it means and what it isn't
E-E-A-T stands for Experience, Expertise, Authoritativeness and Trustworthiness, the qualities Google uses to judge whether content deserves to be believed. Google's guide to creating helpful, reliable, people-first content is plain about two things. "Of these aspects, trust is most important." And "E-E-A-T itself isn't a specific ranking factor."
So there's no E-E-A-T score to raise. It still matters for AI search. Google says the same SEO best practices apply to its AI features as to regular search, so the qualities E-E-A-T describes are likely to shape which pages it picks as sources.
What we found in 158 AI Overview citations
We looked at two things: who Google cited and what those cited pages had in common. The first table shows where the citations came from. The second compares trust signals across all cited web pages and provider sites specifically.
Who AI Overviews cited.
| Source type | Share of 158 citations |
|---|---|
| The service provider's own website (brokers, law firms, accountants, dentists) | 61% |
| Publishers, comparison sites and banks (Canstar, Savings.com.au, OpenAgent, Westpac) | 19% |
| Government, regulators and industry bodies (Moneysmart, NSW Government, Healthdirect, MFAA) | 10% |
| Forums (Reddit, Facebook groups) | 7% |
| YouTube | 3% |
These signals aren't requirements for being cited. They show what was commonly present on pages that Google chose to cite, with provider sites separated to show whether the pattern holds for businesses citing themselves.
| Trust signal on cited pages | All cited webpages (116) | Service provider webpages (82) |
|---|---|---|
| A publish or update date on the page | 94% | 98% |
| A named author (byline, author markup or author meta) | 81% | 88% |
| Structured data (JSON-LD) of any kind | 92% | 96% |
| Specific figures ($ amounts or percentages) | 79% | 80% |
| Dated 2025 or 2026 in the page markup | 70% | 77% |
| At least one link to a government, regulator or industry body | 53% | 54% |
| Author marked up as a Person in schema | 49% | 51% |
| FAQPage schema | 26% | 30% |
| A "reviewed by" or "fact-checked by" line | 15% | 16% |
Every figure comes from Google AI Overviews in Australia, checked in September 2026. ChatGPT, Perplexity and AI Mode may choose their sources differently, so the numbers aren't applied to them.
How we measured. We analysed 21 Australian questions across seven service categories with an AI Overview, using Ahrefs SERP data for Google Australia, last crawled September 2026. There were 158 citations across 147 unique URLs. Forums, YouTube and PDFs were counted in the citation mix but not checked for on-page signals. Sixteen pages blocked our crawler, leaving 116 web pages. Signals were detected from HTML and spot-checked manually.
Providers win most citations. Contrary to the idea that AI mostly cites third parties, 61% of citations went to service providers' own websites. Dental cost questions were the clearest example: 13 of 16 citations came from practices, while 12 of 17 accountant fee citations came from accounting firms.
Dates and authors are common. 94% of cited pages had a publish or update date and 81% named an author. On provider sites, the figures were 98% and 88%. A page with neither is missing two things most cited pages had.
Reviewed-by lines and FAQ schema are less common. Only 15% of cited pages had a "reviewed by" line and 26% had FAQPage schema. None of these signals is required. Only 41% of cited pages had all three of a named author, a date and an authoritative outbound link.
This shows what cited pages have in common, not what caused the citations. The comparison with ranked-but-uncited pages is covered in why your competitors appear in AI search and you don't. This is also one engine, country and snapshot: Google AI Overviews in Australia, September 2026. It shouldn't be treated as a rule for ChatGPT or Perplexity.
Where E-E-A-T signals live
Trust gets read in three places. Your page, your business as an entity, and what the rest of the web says about your category.
| Signal | Where it lives | What AI can check | How to check yours |
|---|---|---|---|
| Named author with a bio | On the page | A real, qualified person stands behind the advice | Is there a byline that links to a bio with credentials? |
| Publish and update dates | On the page | Whether the information is current | Is a date visible, and does it match the last real edit? |
| Specific figures | On the page | Claims it can quote and compare (fees, rates, timeframes) | Is there at least one number in the first few paragraphs? |
| Link to the regulator or government source | On the page | Whether your claims line up with the official source | Do you link to the body that sets the rules in your field? |
| First-hand detail | On the page | Detail only the firm would know | Is there anything a generic article couldn't say? |
| Consistent business details | On the entity | That the business on the page is the one it finds elsewhere | Do your name, services and location match across your site, Google Business Profile, LinkedIn and directories? |
| Organization and Person markup | On the entity | A machine-readable link between content, author and business | Does the markup match what's visible on the page? |
| An About page that names who is accountable | On the entity | Who answers for the advice | Does it name your people and their qualifications? |
| Forums, YouTube, comparison sites and government pages | Off the site | What else AI reads about your category | Search your category question on Reddit and YouTube |
Schema is one row here, and it belongs in the entity layer. 92% of cited pages had some JSON-LD, but only half marked the author as a Person. It's common hygiene, not the deciding factor.
The off-site row shows when AI gets its answer from third-party sources. For example, all three AI Overview citations for a buyers agent question came from Reddit threads. Reddit also accounted for 5 of 12 citations in the buyers agents category. If this happens for your category, check what those sources say about your business and whether they mention you.
The four letters, applied to AI search
Here's what each letter looks like in a form an AI engine can pick up.
Experience. Show you've done the work. Put detail from real client work on the page. Actual fee ranges you charge, typical timeframes, the problems that come up on real jobs. That fits with what we saw on dental and accountant cost questions, where provider cost pages took most of the citations. The AI needed a number, and the practice had one.
Expertise. Put a qualified name on it. A byline linked to a bio, with the qualification that matters in your field. 88% of cited provider pages named an author. An unsigned page asks the AI to trust nobody in particular.
Authoritativeness. Be the name others use. This is built mostly off your site. Industry bodies, directories, comparison sites and forums decide whether your name comes up when someone else talks about the category. The MFAA and Healthdirect appeared among cited sources, and both are the bodies their categories treat as the reference.
Trust. Make it checkable. In our data, that meant a date, a named person and a link to the regulator. The first two were on more than four in five cited pages. The regulator link was on just over half. Each one gives the AI something it can verify rather than take on faith.
Why government pages get cited with no author
Government pages break the pattern. None of the seven government pages we could crawl had a named author. NSW Government, Healthdirect and Legal Aid Victoria pages were still cited, which suggests the organisation itself can carry the trust.
You can't borrow that institutional authority, but you can link to these sources and quote them accurately. That's what 53% of cited pages did. For example, a mortgage broker linking to Moneysmart on a fees page puts an official source alongside its own advice, giving AI another source to check against.
A 30-minute trust check
Pick the page you most want cited, usually your main service or cost page, and use these as checks:
- Is a publish or update date visible, and does it reflect the last real change?
- Is there a named author, with a bio showing their credentials?
- Does the page include a specific figure, such as a fee, rate or timeframe, near the top?
- Does the page link to a relevant regulator or government source?
- Does it include a detail from your own client work that a generic article couldn't include?
- Are your business name, services and location consistent across your site, Google Business Profile, LinkedIn and relevant directories?
- Does any Organization or Person markup match the visible page content?
- Have you searched your main category question on Reddit and YouTube to see what other sources AI might use?
These aren't requirements for getting cited. They're practical checks based on patterns we found across cited pages. Start with the areas that are easiest to improve on your page.
Trust has to be checkable
In service categories, the firm's own page can earn the citation. The cited pages in our check mostly gave the AI something it could check: 94% had a date, 81% a named author and 79% a specific figure.
For the mechanics behind this, see how AI search works, information gain SEO, first-party vs generic content and why most brands are invisible in AI search.
If you want to know how your pages compare with the ones AI cites in your category, start with a free AI visibility audit. Senior-led, two business days, no obligation.
Frequently asked questions
Is E-E-A-T a ranking factor?
No, not directly. Google says "E-E-A-T itself isn't a specific ranking factor." Google uses other signals to assess content quality and trust, so E-E-A-T describes qualities those signals can reflect rather than a score you can improve.
Do AI engines like ChatGPT use E-E-A-T?
E-E-A-T is Google's framework, so other AI engines don't necessarily use the term or the same system. They still need to assess information from different sources. Our data covers Google AI Overviews only, so we can't say which signals ChatGPT or Perplexity use.
What are E-E-A-T signals?
They're features that can help show who created the content, what they're qualified to say and how the information can be checked. In our AI Overview sample, the most common were a date (94%), structured data (92%), a named author (81%) and specific figures (79%).
Which trust signals matter most for getting cited by AI?
Our data shows that dates and named authors were common on cited pages, followed by specific figures and links to authoritative sources. That doesn't show that any one signal causes citations. "Reviewed by" lines and FAQ schema were less common in our sample.
Does an author bio help you get cited in AI Overviews?
Named authors were common on cited pages. 81% of cited web pages had a named author, rising to 88% for provider pages. That's a correlation, not proof that an author bio leads to citations, but it shows that authorship was common among the pages we checked.
Can a small business build E-E-A-T without press coverage?
Yes. 61% of the citations in our check went to service providers' own websites. A business can make its own pages easier to assess by showing who wrote them, keeping information current, including relevant figures and linking to authoritative sources.
Does AI-generated content hurt E-E-A-T?
Not simply because AI was used to create it. Google says its focus is on the quality and usefulness of the content rather than the tool used to produce it. Generic AI copy can still be a problem if it lacks first-hand detail, useful figures or information specific to the business.
How long does it take for trust signals to show up in AI answers?
Changes such as adding a date or byline can be made quickly, but there's no fixed timeframe for when they might appear in AI Overviews. Google needs to recrawl and process the updated page, and off-site signals such as mentions on forums or industry sites depend on factors outside your control.