How brands get cited in ChatGPT, Perplexity, and Google AI
Updated: Aug 18
90% of brands have zero AI citations.
Not low visibility. Zero.
That figure comes from a Search Engine Journal analysis of 177 brands across healthcare, SaaS, and financial services, published May 2026. These are businesses with real websites, real content, and real SEO investment. When someone asks an AI engine about their category, they simply do not appear.
The question is why. And more importantly: what does it actually take to get cited?
Is citation a side effect of good SEO?
No. Most businesses assume AI tools will pick them up if they rank well on Google. The assumption is wrong.
Research published in 2026 found that 60% of AI-cited sources are not in Google's top 10 organic results. By early 2026, only 38% of Google AI Overview citations came from top-10 results, down from 76% in mid-2025 (Ahrefs). BrightEdge's parallel research put that figure as low as 17%.
AI citation operates on different logic to organic ranking. It has different signals, different source preferences, and dramatically different outcomes by platform. Understanding that gap is where citation strategy starts.
Where do AI citations actually come from?
5WPR's AI Platform Citation Source Index 2026 synthesised 680 million citations across ChatGPT, Google AI Overviews, Perplexity, Gemini, and Claude. The source hierarchy, ordered by citation frequency:
Reddit: the number one source across every major AI engine, cited at roughly 40% frequency
Wikipedia: near-foundational for ChatGPT, accounting for 26 to 48% of its top-10 citation share
LinkedIn: the most-cited domain for professional queries across AI Overviews, AI Mode, ChatGPT, Copilot, and Perplexity, appearing in 11% of all AI responses (Profound, March 2026)
Major news media: Forbes, Business Insider, and national publications cited regularly across all engines
Industry authority sites: including NIH/PubMed for health queries and sector-specific publications
Review platforms: Trustpilot, Google Business Profile, and equivalent third-party review sites carry increasing citation weight
Your own website: present, but lower in the hierarchy than most businesses expect
The top 15 domains capture 68% of all AI citation share (5WPR, 2026). That is concentration more extreme than anything Google PageRank produced.
This is not a comfortable picture for most brands. Most of the citation landscape is off-site and outside direct control. But several of those signals, review presence, LinkedIn publishing, content freshness, are entirely actionable.
Does ChatGPT cite the same sources as Perplexity or Google?
No. An analysis of 680 million citations found that only 11% of domains are cited by both ChatGPT and Perplexity (Growth Memo, May 2026). Each platform runs on fundamentally different citation logic.
In a study of 34,234 AI responses, brand citation rates varied by a factor of 46 across platforms:
ChatGPT: 0.59% brand citation rate; draws primarily from Wikipedia, Reddit, and established media
Perplexity: 13.05% brand citation rate; prioritises primary sources, named authority, and recency, citing content published within the last 30 days at an 82% rate
Grok: 27% brand citation rate; broader source scope
Google AI Overviews: citation from top-10 organic results has declined sharply; now weights semantic completeness, structured data, and E-E-A-T signals independently of rank
A brand optimising only for Google is likely invisible across the rest. Getting cited at scale means understanding what each platform draws from and being present there.
What trust signals does AI actually look for?
One of the most counterintuitive findings from 2026 research involves review profiles.
Brands with no Trustpilot presence have a median AI citation rate of 1%. Brands with just 1 to 13 reviews jump to 53.5%, a 52-percentage-point increase (SEJ, May 2026). The same principle applies to Google Business Profile and sector-specific platforms: the signal is third-party validation, not the platform itself.
AI models are not evaluating your website in isolation. They pull in external signals to confirm whether a brand is credible before recommending it. Review platforms, third-party directories, press mentions, and professional networks all feed the same trust layer.
Recency sends its own signal, particularly to Perplexity. Keeping cornerstone content visibly maintained, updating timestamps, refreshing data, and adding to knowledge hubs, indicates an active source worth citing. Practitioners tracking this on product listing platforms have found citation rates increase significantly when content is treated as a maintained document rather than a static page.
For LinkedIn specifically, the most-cited posts share a consistent profile (Semrush, March 2026):
54 to 64% of cited posts focus on sharing knowledge or practical advice
Long-form articles (500 to 2,000 words) and mid-length posts (50 to 299 words) account for the largest citation share
Most-cited authors post consistently, five or more times per month
Typical cited engagement: moderate rather than viral, around 15 to 25 reactions
Follower count matters less than relevance and consistency.
Does your website still matter once you're being cited?
Yes, more than ever.
External signals get you cited. Your website has to convert whoever arrives.
Getting cited puts you in front of a potential client. Converting that visit is a separate challenge, and most brands are focused on one while neglecting the other. The page a recommendation leads to needs to honour what the AI promised:
Structured, extractable content: clear headers, clean markup, specifically defined services
Entity signals: your brand name connected unambiguously to what you do and where
Decision-stage content: answering the questions the AI has likely already summarised for the user
Trust confirmation that matches the signal that prompted the citation
Both layers need to work. Most brands are only building one of them.
So where should you actually start?
Most brands asking why they are not being cited are looking inward at their website. The data suggests they should look outward first.
Getting cited requires at least three layers working simultaneously: your on-site content structure, your external signal presence, and your content freshness strategy. Focusing on only one will still leave two-thirds of the challenge unsolved.
That is what a proper generative engine optimisation (GEO) audit covers, not just whether your technical markup is correct, but whether the full set of signals AI uses to decide who to recommend are working in your favour. If you are unsure where you stand, Xyra Foundry's AI Readiness Audit covers all three layers and shows you exactly where the gaps are.
Practical note
The fastest single lever for most service businesses is establishing a verified third-party review presence. Google Business Profile is the most universally applicable and among the most heavily indexed by AI engines. Trustpilot carries additional weight in professional services. Sector-specific platforms such as Doctify for healthcare also contribute to the external signal layer. Start there, then work back through your LinkedIn publishing cadence and on-site content structure.
Sources
Search Engine Journal, May 2026: 177-brand citation analysis across healthcare, SaaS, financial services
5WPR AI Platform Citation Source Index 2026: 680 million citations across ChatGPT, Google AI Overviews, Perplexity, Gemini, Claude
Growth Memo, May 2026: platform citation overlap (11% domain overlap between ChatGPT and Perplexity)
Profound, March 2026: LinkedIn citation frequency across AI platforms
Semrush, March 2026: LinkedIn cited post characteristics
Ahrefs, early 2026: Google AI Overview citation source shift (38% from top-10 results)
BrightEdge, 2026: Google AI Overview (17% from top-10 results)
Leapd.ai, April 2026: 34,234 AI response brand citation rate comparison by platform
Quora/Business, 2026: 60% of AI-cited sources not in Google top-10
Comments