Products SEO & GEO Agent Paid Ads Agent Website Agent Conversion Agent
How it works Pricing FAQ Blog Book a call

Trust signals for AI search: experts, sources and authors

Updated · 7 min read

Short answer AI answers lean on pages that show who knows the topic and where each fact comes from. Google says its AI features run on its core quality systems, which value clear authorship and sourcing; Microsoft weighs original, attributed content. Name a real expert on each advice page, cite primary sources, add first-hand facts, and never fake credentials.

Want help applying this to your business? Talk it through on a free 30-minute call.

Book a call

What trust signals do AI answers look for?

Three kinds: who stands behind the page, where its facts come from, and proof only you could have. No AI company publishes a trust score. Google and Microsoft do describe what their systems look for, and both point the same way.

Google says its AI answers run on the same machinery as its normal results:

"our generative AI features on Google Search are rooted in our core Search ranking and quality systems." (Google Search Central: Optimizing for generative AI features)

Google's self-check for site owners then spells out what trust looks like:

"Does the content present information in a way that makes you want to trust it, such as clear sourcing, evidence of the expertise involved, background about the author or the site that publishes it, such as through links to an author page or a site's About page?" (Google Search Central: Creating helpful content)
SignalOn a service business pageWho says it matters
Named authorA byline that links to a bio: role, license, years in the tradeGoogle asks whether "bylines lead to further information about the author"
Expert reviewA "reviewed by" line with name, credential and dateGoogle asks if content is "written or reviewed by an expert or enthusiast who demonstrably knows the topic well"
Cited sourcesThe regulator, code, manufacturer or study named beside the claimMicrosoft: "Examples, data, and cited sources help build trust when content is reused in AI-generated answers"
First-hand factsYour own job counts, prices, photos, findingsMicrosoft asks whether a site reports "first-hand information" or republishes others' work
Outside reputationThe expert and the business show up on other trusted sitesGoogle asks what someone who "researched the site" would conclude; Bing asks "What types of other websites link to the site?"

A Microsoft Bing product manager adds that Copilot splits pages into small pieces "that can be evaluated for authority and relevance." So keep each claim's proof right next to it.

What E-E-A-T means, and which topics face a higher bar: E-E-A-T for a small business.

Do AI tools check who wrote the page?

They check what their search systems check, and authorship is part of it. Make each expert easy to confirm:

  1. Byline on advice pages. The real person who knows the topic. If the owner or a licensed staff member checked a draft someone else wrote, say "Reviewed by", with their name and the date.
  2. One author page per person. A real photo, role, license type and number where your trade issues one, years of practice, and the work they do.
  3. Links to outside proof. From the author page, link the licensing board's lookup, the association directory and the person's professional profile. These confirm the credential on sites you don't control.
  4. One name, spelled one way. The same name and credential on your site, Google Business Profile, directories and social profiles.
  5. Author markup, if your site makes it easy. Google's article markup guide says: "To help Google better understand who the author is, we strongly recommend using the type and url (or sameAs) properties." Google also says structured data "isn't required for generative AI search." How to add it: local business schema markup.

Service, pricing and booking pages don't need a personal byline. Google asks for one "where one might be expected": articles that give advice.

Does citing sources help you get cited?

In the best-known test, yes. Researchers from Princeton and IIT Delhi (GEO: Generative Engine Optimization, KDD 2024) rewrote web pages several ways and measured how much of each page showed up in AI answers. Three changes did best: citations to credible sources, quotations, and statistics. They raised visibility by 30–40% on one of the study's two measures and 15–30% on the other. Keyword stuffing did not help; on Perplexity it scored about 10% below the unchanged page.

Two findings matter to a small business:

  • Lower-ranked pages gained more. The page ranked fifth on Google gained 115% from citing sources, while the top-ranked page lost 30%.
  • Sounding confident is not proof. A rewrite in a more "persuasive and authoritative" tone gained about 10%. Adding quotations gained about 40%.

The limits: answers came from GPT-3.5 in a test engine the researchers built, and from Perplexity reading uploaded page text, not live websites. Treat the numbers as direction, not a promise.

How to cite on a service page:

  • Use the primary source. The regulator, the building code, the manufacturer's manual, the tax agency, the study itself. Not a blog that summarized it.
  • Name it in the sentence. "The manufacturer's install manual requires...", then link it.
  • Quote exact words when a rule decides the answer.
  • Keep each number next to its source and date. Microsoft's advice for AI answers: "anchor claims in measurable facts."
  • Check the source says what you say. Google's self-check asks: "Does the content have any easily-verified factual errors?"

What counts as first-hand proof?

Anything a summary of other websites can't contain. Google's AI guide draws the line:

"a first-hand review provides a unique perspective based on personal experience, whereas a summary of existing content simply restates information already available elsewhere." (Google Search Central: Optimizing for generative AI features)

Microsoft's ranking principles say the same for Bing: "A first-hand account published on a personal blog could have more authority than unsourced content."

First-hand proofOn the page
Counts from your own workHow many of this job you did last year, and what you found most often
Your real pricesThe range you actually quoted, and what moved it
Photos from real jobsWith the town and the month
How your licensed people decideThe checks they run and why
Named staff quotesFull name, credential, one specific point

Outside experts. In the US, when a page advertising your service presents someone as an expert who endorses it, the FTC's Endorsement Guides apply: "the endorser's qualifications must in fact give the endorser the expertise that the endorser is represented as possessing" (16 CFR 255.3). A "doctor approved" line needs a doctor whose field fits the claim.

Which trust signals backfire?

TacticWhy it fails
Made-up authors, AI headshots, borrowed credentialsGoogle: "Fabricating creator profiles ... is a form of deception", which makes a page "untrustworthy to both users and our automated quality systems"
Planting mentions of yourself on forums or review sitesGoogle: "seeking inauthentic 'mentions' across the web isn't as helpful as it might seem"
Author markup that doesn't match the bylineGoogle wants structured data that "matches the visible text on the page"
Sources that don't back the claimGoogle's self-check counts "easily-verified factual errors" against trust

How do you check what AI tools know about your experts?

Ask them. In ChatGPT, Gemini, Copilot and Perplexity, ask "Who is [name], [trade] in [city]?" and one question your best article answers. Note three things: does the answer name the person, get the credential right, and cite your page?

If a credential is wrong, find and fix the old bio or listing it came from. For Copilot and Bing's AI answers, Bing Webmaster Tools' AI Performance report, in public preview since February 2026, shows which of your pages are cited. A fuller method: the 20-question ChatGPT test.

Example

Illustrative cases, round numbers.

A dental practice's implant cost page (US)

Before: no author, "implants cost a few thousand dollars", no source. After: written by the dentist, whose author page links the state dental board's license lookup. The price range comes from the practice's own last 12 months: most single implants $4,000–$5,000. Insurance limits are quoted from the plan documents patients bring in.

An electrician's panel upgrade guide

Before: a generic list copied from other sites. After: "Reviewed by" the master electrician with license number and date, the city permit office's page named and quoted, and photos from 6 real upgrades with town and month. The page also says what made 3 of last year's 40 jobs cost more.

A Toronto bookkeeper's HST deadline article (Canada)

Before: a deadline list with no author. After: "Reviewed by" the firm's CPA with the date, each deadline linked to the Canada Revenue Agency page it comes from, and the 3 filing mistakes the firm saw most in 2025.

FAQ

Do AI tools care if AI wrote the article?

Google says AI-assisted work must meet its Search Essentials and spam policies. Put the expert who checked it for accuracy on the byline.

Won't linking to other sites send AI to them instead?

Google and Microsoft both treat cited sources as a trust signal, and citing sources raised visibility in the GEO study. Link the primary source.

Is author schema required for AI search?

No. Google says structured data isn't required for its AI features. If you add article markup, link each author to their page.

How fast do AI answers notice new bylines and sources?

Google says crawling a changed page can take from several days to several months. Recheck your test questions monthly.

Different for you?

Keep learning

NextConversational search queries: optimize for how people ask AIPeople ask AI long questions full of conditions. How Google and Copilot match them to pages, and the plain facts a service page needs to get picked.

Do it yourself

Does ChatGPT recommend your business? A 20-question testA free test any owner can run to see whether ChatGPT names their business, which competitors it names instead, and why.

Ask us one question

One line is enough. We reply by email.

Sources

  1. Google Search Central: Optimizing your website for generative AI features on Google Search
  2. Google Search Central: Creating helpful, reliable, people-first content
  3. Google Search Central: AI features and your website
  4. Google Search Central: Article structured data, author markup best practices
  5. Microsoft Support: How Bing delivers search results (March 2025)
  6. Microsoft Advertising Blog: Optimizing your content for inclusion in AI search answers (October 2025)
  7. Bing Webmaster Blog: Introducing AI Performance in Bing Webmaster Tools Public Preview (February 2026)
  8. Aggarwal et al.: GEO: Generative Engine Optimization (KDD 2024)
  9. eCFR: 16 CFR 255.3, Expert endorsements (FTC Endorsement Guides)