Most B2B companies assume these signals are the same as Google's ranking factors. They're not. Domain Authority, which dominates traditional SEO thinking, explains only 4% of variance in AI citations (Princeton GEO study). The signals that matter for AI search are different, they vary by platform, and understanding them is the difference between being recommended and being invisible.

What signals do AI platforms use to rank businesses?

The research identifies six core signals that influence AI citation across all platforms. The weight each platform gives them differs, but the signals themselves are consistent.

Topical authority. How deeply a website covers its subject. This is the strongest single predictor of AI citation, with a correlation of r=0.41 (Princeton GEO study). A site with 30 articles covering AI search visibility in depth will be cited before a site with one generic page, even if the generic page has stronger backlinks. AI platforms reward depth, not domain strength.

Content freshness. How recently content was published or updated. 50% of all AI citations come from content under 13 weeks old. This is a hard constraint, not a preference. Content that was authoritative six months ago may already be invisible.

Structured data. Schema markup (JSON-LD) improves AI citation rates by 44-73%. Pages with three or more schema types are 2.8 times more likely to be cited (ZipTie.dev). FAQ schema alone increases citations by 44% and makes content three to five times more likely to be selected (BrightEdge). This is one of the highest-leverage technical changes a business can make.

Content extractability. AI platforms don't rank whole pages. They extract passages. Content structured with self-contained sections of 134-167 words, question-based headings, and explicit definitions ("X is...") gets extracted more reliably. 44.2% of ChatGPT citations come from the first 30% of content (Kevin Indig, 1.2 million responses). Front-loading matters.

Third-party validation. What others say about your business carries 6.5 times more weight than what you say about yourself (Muck Rack, one million-plus citations). 82% of AI citations originate from earned media. Directory listings, client reviews, industry publications, forum mentions. These are the sources AI platforms draw from when deciding whether to trust your business enough to recommend it.

E-E-A-T signals. Experience, Expertise, Authoritativeness, Trustworthiness. 96% of Google AI Overview citations come from sources with strong E-E-A-T signals. Named authors with credentials, published case studies, verifiable claims, and consistent business information across the web. Anonymous content with no attribution gets deprioritised.

How do signals differ between ChatGPT, Perplexity, and Google?

This is where it gets specific. Each platform weights these signals differently, which is why a business can be visible on one platform and invisible on another.

ChatGPT prioritises referring domains (32K threshold), Bing ranking (87% correlation), and content depth (2,900+ words). Citation decay is rapid outside 30 days. Top sources include Wikipedia, Reddit, and YouTube.

Perplexity weights freshness most heavily (~40% weight), applies an ML quality gate, and draws from curated domain lists. Citation decay is exponential after hours. Reddit accounts for 47% of its top sources, followed by Wikipedia and YouTube.

Google AI Overviews favour multi-modal content (r=0.92), semantic completeness (r=0.87), and E-E-A-T (r=0.81). The citation cycle runs on a 12-month window. Top sources are YouTube, Wikipedia, and Google.com.

Copilot relies on Bing index position, schema markup, and freshness, with 97% of citations decaying within 91 days. Claude favours primary sources and methodology transparency. Grok integrates web and X presence with a real-time bias and answer-first format preference.

The numbers that matter most:

• →
  Only 11% of domains appear in both ChatGPT and Perplexity results for the same query
• →
  ChatGPT correlates 87% with Bing's top 10 but only 56% with Google
• →
  80% of URLs cited by AI assistants don't rank in Google's top 100
• →
  Google AI Overviews and Google AI Mode share only 10.7% URL overlap with each other

A single content strategy cannot address all six platforms equally. Each piece of content should target two to three platforms based on its format and strengths.

How are AI ranking signals different from Google ranking signals?

Google and AI platforms share roughly 90% of the same quality signals (Lily Ray, 2025). Good content, clean technical setup, schema markup, E-E-A-T. The 10% difference is what catches companies out.

Domain Authority matters less. Google weights backlink profiles heavily. AI platforms don't. DA correlation with AI Overviews dropped from r=0.43 to r=0.18 between 2024 and 2026 (Wellows/Stackmatix). A strong backlink profile used to guarantee visibility. It no longer does, at least not for AI.

Freshness matters more. Google allows evergreen content to rank indefinitely. AI platforms actively prefer recent content. Perplexity gives a 38% citation boost to content updated within two hours. ChatGPT's top-cited pages are 76.4% updated within 30 days. The 13-week window is a hard constraint across platforms.

Format determines extractability. Google ranks pages. AI platforms extract passages. A page can rank well on Google without being extractable by AI. Comparative listicles account for 32.5% of all AI citations. Comparison tables are cited 2.5 times more often than prose. These formats don't necessarily help Google rankings, but they directly improve AI citation rates.

Off-site presence is weighted more heavily. Google evaluates your website primarily, with backlinks as a proxy for external validation. AI platforms directly read third-party sources (Reddit, review sites, directories, news publications) and use them as citation sources. Brand mentions across the web correlate at 0.334 with AI citations, stronger than backlinks at 0.218 (Ahrefs, 75,000 brands).

Brand entity consistency is a factor. Google is relatively tolerant of inconsistent business information across the web. AI platforms are not. Conflicting descriptions, outdated directory listings, and name variations reduce the AI's confidence in recommending your business. If Tracxn says you're a Dubai-based Web3 agency and your website says you're a UK AI consultancy, the AI doesn't know which to trust.

Which signals matter most for B2B companies?

Not all signals are equally actionable. For a B2B company starting from a low AI visibility baseline, the priority sequence is:

• 01.
  Technical accessIf AI crawlers can't read your site, your own content can't be a direct citation source. Check robots.txt. Allow GPTBot, ClaudeBot, PerplexityBot, Google-Extended. This is a five-minute fix.
• 02.
  Content depthTopical authority is the single strongest predictor (r=0.41). Most B2B sites have thin content: a few service pages, an about page, maybe a blog post. AI platforms need depth to build confidence in your expertise. 30 or more pages covering your core topics is the minimum for meaningful topical authority.
• 03.
  Schema markup44-73% citation improvement. Three or more schema types per page yields 2.8 times higher citation likelihood. Article, FAQ, Organization, Service, Person schema on the right pages. High impact, moderate effort.
• 04.
  Publishing cadenceThe 13-week freshness window means you need regular output. Weekly at minimum. During a sprint, daily. Each piece of content is a new surface for AI citation.
• 05.
  Third-party presenceThis takes longer to build but carries 6.5 times the weight of owned content. Industry directories, Clutch and G2 reviews, LinkedIn articles, Reddit contributions. These are the sources AI platforms actually trust.
• 06.
  E-E-A-T signalsNamed authors, credentials, case studies with specific results. 96% of Google AI Overview citations come from sources with strong E-E-A-T. This isn't optional.

The order matters. Fixing technical access before publishing content before building off-site presence. Each step enables the next.

How to strengthen your AI visibility signals

For each of the six signals, here's the specific action:

Topical authority. Publish 30 or more pieces of content covering your core subject in depth. Structure as hub-and-spoke clusters: one pillar page (3,000+ words) with 5-10 supporting articles (1,500 words each) linking back to it. Every article adds another surface the AI can cite.

Freshness. Publish weekly at minimum. Update pillar content quarterly (change the dateModified, add new data, refresh statistics). The 13-week clock resets with every meaningful update.

Structured data. Add Article + FAQ + BreadcrumbList + Person schema to every blog post. Add Service + FAQ schema to service pages. Add Organization schema to the homepage. Test with Google's Rich Results Test. Target three or more schema types per page.

Extractability. Structure content with self-contained answer units of 134-167 words. Start sections with clear definitions. Use question-format headings. Include a statistic with source attribution every 150-200 words. Add FAQ sections with schema to every substantial page.

Third-party presence. Get listed on Clutch, G2, and relevant industry directories. Request client reviews. Contribute to Reddit and LinkedIn discussions in your space. Ensure your business information is identical across every platform. Clean up outdated directory listings.

E-E-A-T. Add named authors with credentials to all content. Publish case studies with specific results and named clients (with permission). Display relevant qualifications and experience. Link to author LinkedIn profiles via Person schema.

What to do next

You now know the six signals. The question is which ones are weakest for your business.

Run a quick test: check your robots.txt for AI crawler access, count your indexed content pages, check your schema markup with Google's testing tool, and search for your business on Clutch, G2, and three industry directories.

If any of those checks come back empty, you've found your starting point. For a full signal assessment across all six AI platforms, 1DOT's free audit scores your business across seven categories and identifies exactly which signals need work first.