What trust signals actually matter for AI citations? Backlinks feel less important than they used to be

Your hypothesis is supported by data. Let me share what the research shows:

Correlation with AI Overview visibility:

Key insight: Mentions are 3x more predictive than backlinks.

Why this makes sense:

AI systems are trying to answer: “What sources would a knowledgeable human recommend?”

• Backlinks show other websites find you useful to link to
• Mentions show people are talking about you as a legitimate player
• The second is a stronger “real world” trust signal

The “visibility cliff”:

Brands in the top 25% for web mentions get 10x more AI citations than the next quartile. Below median? Essentially zero mentions.

You need to cross a threshold of brand awareness before AI considers you citable.

Let me explain the entity identity component, which is often overlooked:

AI systems build knowledge graphs - representations of entities and their relationships. When your brand has a clear entity identity, AI can:

1. Recognize mentions of you across the web
2. Aggregate trust signals accurately
3. Cite you with confidence

How to build entity identity:

1. Organization Schema on your homepage:

{
  "@type": "Organization",
  "name": "Your Company",
  "sameAs": [
    "https://linkedin.com/company/yourcompany",
    "https://crunchbase.com/organization/yourcompany",
    "https://twitter.com/yourcompany"
  ]
}

2. Consistent naming everywhere - Same company name on all profiles, directories, mentions

3. Wikipedia presence - Even a stub establishes you in major knowledge bases

4. Google Knowledge Panel - Claim and verify your business panel

The problem with weak entity identity:

If AI sees “Acme Inc” in one place, “Acme Incorporated” in another, and “Acme Corp” in a third, it may treat these as three different entities. Your trust signals get fragmented.

Reddit deserves special attention for AI trust signals.

The data: Reddit appears in 68% of AI Overview results, with citations increasing 450% between March and June 2025.

Why Reddit matters so much:

1. It’s user-generated, authentic content
2. Upvotes = community validation
3. AI systems interpret Reddit mentions as “real people recommend this”

How to build Reddit presence authentically:

• Have team members who genuinely use Reddit participate in relevant subreddits
• Answer questions helpfully without being promotional
• Share genuine experiences and insights
• Let brand mentions happen naturally through reputation

What NOT to do:

• Astroturfing (will be detected and backfire)
• Promotional posting (gets downvoted)
• Creating accounts just for self-promotion

The goal isn’t “get mentioned on Reddit.” It’s “be genuinely helpful on Reddit and let mentions follow.”

E-E-A-T is getting more important for AI, not less. Here’s how to build each component:

Experience:

• First-person accounts (“In my 10 years as a developer…”)
• Original photos/videos of actual work
• Case studies with specific, verifiable details
• Genuine success and failure stories

Expertise:

• Detailed author bios with credentials
• Links to professional profiles (LinkedIn, industry associations)
• Published research or thought leadership
• Speaking engagements and certifications listed

Authoritativeness:

• Third-party mentions from trusted sources
• Being quoted as an expert in publications
• Awards and recognition
• Industry rankings and reviews

Trustworthiness:

• Transparent About Us and Contact pages
• Real business address and contact info
• Clear authorship on all content
• Sources cited for claims

The AI-specific angle: AI systems can verify these signals across the web. Fake credentials or unverifiable claims get discounted. Authentic, consistent signals across platforms get amplified.

This is reshaping my mental model. Let me try to synthesize:

Old model (backlink-focused):

• Get links from high-authority sites
• Anchor text optimization
• Link velocity and diversity

New model (trust signal-focused):

• Get mentioned by name in trusted contexts (linked or not)
• Build recognizable entity identity across platforms
• Establish author expertise visibly
• Participate authentically in communities
• Make verification easy for AI systems

The fundamental shift: From “convince algorithms you’re authoritative” to “actually be recognized as authoritative by real people and systems.”

Is that the right framing?

That’s exactly right. And here’s the uncomfortable truth:

You can’t fake AI trust signals the way you could game backlinks.

Link building has always had shortcuts - PBNs, paid links, link exchanges. They worked because Google was pattern-matching.

AI systems are more holistic. They’re essentially asking: “Would a knowledgeable human recognize and trust this source?”

• Fake mentions get detected by context analysis
• Astroturfing gets identified by behavioral patterns
• Inconsistent entity signals create confusion rather than trust

The strategy has to be: Actually build real authority, then make sure AI systems can see it.

Am I Cited is helpful here because it shows you exactly how AI systems perceive your brand - what they say about you, not just whether they mention you. That context reveals whether your trust building is working.

Practical actions we took to build AI trust signals:

Month 1-2: Entity foundation

• Claimed and optimized all major profiles (LinkedIn, Crunchbase, industry directories)
• Implemented Organization schema with sameAs links
• Standardized brand name across all platforms

Month 3-4: Mention building

• Pitched industry publications for expert commentary (not links, just quotes)
• Contributed to industry roundup articles
• Published original research that got picked up by others

Month 5-6: Community presence

• Key team members became active in relevant subreddits
• Participated in industry forums and Slack communities
• Started answering questions on Quora in our area

Results after 6 months:

• 40% increase in AI citations
• Moved from position 3-4 to position 1-2 for key queries
• Brand described more positively in AI responses

The timeline matters - this isn’t a quick fix. It’s building real recognition.

Technical trust signals that matter:

1. HTTPS everywhere - Basic but essential

2. Core Web Vitals - Fast, stable pages signal quality

3. Accessibility - Alt text, proper headings, good contrast

4. Clean HTML - Semantic markup, valid code

5. No dark patterns - Pop-ups, deceptive CTAs hurt trust

These might seem basic, but AI systems use them as quality filters. A site with poor technical signals may never get considered for citation regardless of content quality.

Think of it as table stakes - necessary but not sufficient for AI trust.

Amazing thread. Here’s my action plan:

Foundation (Immediate):

• Audit entity consistency across all platforms
• Implement Organization schema with sameAs
• Fix any technical trust issues (speed, HTTPS, accessibility)

Mention Building (Months 1-3):

• Shift PR focus from link building to mention earning
• Develop expert commentary opportunities
• Create citable original research

Community Presence (Ongoing):

• Identify team members who can authentically participate on Reddit
• Find relevant industry communities
• Start genuine participation, not promotion

Measurement:

• Track brand mentions (linked and unlinked) monthly
• Monitor AI citation frequency with Am I Cited
• Review context of citations (how are we described?)

Key mindset shift: Stop asking “how do we get more links?” and start asking “how do we become more genuinely recognized in our space?”

This is going to require cross-functional effort - PR, content, product, and executive involvement. But it’s clearly the direction AI is pushing us.

Perspective from someone who’s seen marketing cycles:

What you’re describing is actually marketing returning to fundamentals. Before the internet, trust was built through:

• Real reputation in industry
• Word of mouth
• Being quoted in trade publications
• Genuine expertise recognition

SEO created a shortcut game. AI is pushing back toward genuine authority signals.

The brands that will win in AI aren’t the best at optimization - they’re the ones that are actually trusted and recognized in their industries.

That’s good news if you focus on being genuinely excellent. Less good news if your strategy was mostly technical optimization.