Who Gets Cited by AI?

Key Findings

1. Overall Citation Rate

88.1% of AI responses cited at least one specific brand. (141/160)

AI platforms are actively recommending brands in the vast majority of queries — businesses not appearing in these responses are invisible to a growing segment of buyers.

2. Platform Citation Rates

Platform	Citation Rate	Avg Brands Per Response
ChatGPT	85.0%	2.5
Perplexity	87.5%	2.6
Claude	97.5%	2.9
Gemini	82.5%	2.4

Claude cited brands most frequently at 97.5% of responses, followed by Perplexity at 87.5%. However, Perplexity was the only platform to consistently provide source URLs with its citations — making it the only AI engine currently driving direct referral traffic. Perplexity's architecture (search-augmented generation with live web retrieval) means it pulls from current, structured web content rather than training data alone. Businesses seeking AI-driven referral traffic should prioritize Perplexity visibility, while those seeking brand recognition should optimize across all four platforms.

3. Vertical Comparison

Vertical	Citation Rate
HVAC	85.0%
Plumbing	87.5%
Legal	85.0%
B2B SaaS	95.0%

Because B2B SaaS brands were cited at 95.0% — higher than local service verticals — this confirms that brands with comprehensive online presence, structured data, and extensive review profiles are significantly more likely to appear in AI responses. Local service businesses (HVAC at 85.0%, Plumbing at 87.5%) had lower citation rates, with most citations going to directory platforms (Yelp, Google Maps, Angi) rather than individual business websites.

4. Cross-Platform Agreement

All 4 AI platforms agreed on the #1 brand in only 2.5% of queries. (1/40)

This means each AI platform has its own citation preferences and data sources. A business visible on one platform may be invisible on another. This fragmentation creates both risk and opportunity: you can't optimize for just one AI engine.

5. Answer Format Breakdown

Format	Share
List	62.5%
Step-by-step	16.9%
Comparison table	15.0%
Paragraph	5.6%

The dominance of list-based and step-by-step formats means AI engines prefer structured, extractable content. Businesses with content organized in lists, tables, and clear hierarchies are more likely to have their information parsed and cited.

6. Schema Markup Is Near-Universal Among Cited Brands

We audited the websites of 110 cited brands for Schema.org structured data. The results were definitive:

Has Schema Markup	Cited Brand Count	Share
Yes	107	97.3%
No	3	2.7%

Brands with Schema markup were cited 35.67x more often than those without.

The 3 brands lacking schema were small local businesses cited only by Perplexity through live web search — not through training data. Because 97.3% of all cited brands had structured data markup, Schema.org implementation is effectively a prerequisite for AI citation. This is the single strongest technical signal in our dataset.

7. Reviews Are a Near-Mandatory Signal

Of 110 cited brands with auditable websites, 109 had customer reviews or testimonials — either on their own site or on major review platforms:

Has Reviews/Testimonials	Count	Share
Yes	109	99.1%
No	1	0.9%

Reviews correlated with citation at 109x lift.

The single brand without reviews was a small law firm cited once by Perplexity. Every other cited brand — from HubSpot to Yelp to local HVAC contractors — had some form of review presence. Because AI models use review signals as trust indicators, businesses without reviews are effectively invisible to AI answer engines.

8. FAQ Pages Provide a 1.75x Lift

Has FAQ Page	Count	Share
Yes	70	63.6%
No	40	36.4%

Brands with FAQ pages were cited 1.75x more often than those without.

While not as dominant as schema or reviews, FAQ presence correlates with higher AI citation rates. This aligns with finding #5 — AI engines generate list-based, Q&A-style responses, and FAQ pages provide pre-structured content that maps directly to that format.

9. Content Type Breakdown: Directories Dominate

We classified every cited brand by the type of content AI engines were referencing:

Content Type	Count	Share
Directory listing	53	48.2%
Service page	50	45.5%
Review site	7	6.4%

Directory listings were the #1 content type cited (48.2%).

For local service queries (HVAC, Plumbing, Legal), AI engines overwhelmingly recommended where to find businesses (Yelp, Google Maps, Angi, Avvo) rather than recommending specific businesses directly. Service pages dominated in the B2B SaaS vertical, where brands like HubSpot and Salesforce were cited directly.

The most frequently cited directory platforms:

• Yelp: cited 38 times
• Google Reviews: cited 20 times
• Google Maps: cited 15 times
• Avvo: cited 15 times
• Angie's List: cited 14 times
• Martindale-Hubbell: cited 11 times
• Angi: cited 10 times
• Better Business Bureau: cited 5 times

The implication: your presence on these platforms (with complete profiles, reviews, and structured data) is critical for AI visibility in local service verticals.

10. Pricing Pages Are Not Required — But Pricing Data Is

Has Pricing Page	Count	Share
Yes	33	30.0%
No	77	70.0%

Most cited brands (70%) did NOT have public pricing pages. This is because directories (Yelp, Angi, BBB) and manufacturers (Carrier, Trane, Lennox) — which account for the majority of citations — don't publish pricing. However, 49.4% of AI responses included pricing data anyway, pulling from blog posts, comparison articles, and industry guides. The takeaway: businesses don't need a dedicated pricing page to appear in AI responses, but those that do publish pricing provide extractable data that AI engines actively surface.

11. Domain Authority Distribution

Domain Authority	Count	Share
High	101	91.8%
Medium	6	5.5%
Low	3	2.7%

91.8% of cited brands had high domain authority. AI engines overwhelmingly cite established, well-known brands and platforms. The 3 low-DA citations were local businesses surfaced only by Perplexity through live web search. This confirms that for ChatGPT, Claude, and Gemini — which rely more on training data — brand authority is a dominant citation signal.

12. B2B SaaS Brands Dominate Named Citations

In contrast to local services, B2B SaaS queries produced abundant named brand citations:

• HubSpot CRM: cited 12 times across platforms
• HubSpot: cited 9 times across platforms
• Salesforce: cited 9 times across platforms
• Carrier: cited 7 times across platforms
• Trane: cited 6 times across platforms
• G2: cited 6 times across platforms
• Capterra: cited 6 times across platforms
• Lennox: cited 5 times across platforms
• Zoho: cited 5 times across platforms
• LegalZoom: cited 4 times across platforms

These brands share a common profile: Schema markup (100%), review presence (100%), FAQ pages (majority), high domain authority, and rich, structured content libraries.

13. Zero-Citation Rate

11.9% of queries returned generic advice with zero brand recommendations. (19/160)

Very few responses were entirely generic — AI platforms are actively citing brands in most queries. The zero-citation responses were concentrated in local service verticals where AI engines gave procedural advice ("here's how to find a plumber") rather than naming specific businesses.

14. URL Provision Rate

Only Perplexity consistently provided source URLs with its citations.

ChatGPT, Claude, and Gemini cited brands by name but did not link to specific web pages. This means:

• Perplexity is the only platform currently driving direct referral traffic from AI citations (100% of its citations included source URLs)
• For ChatGPT, Claude, and Gemini, brand recognition (not clicks) is the primary value of being cited
• Businesses should track brand mention volume across all platforms, not just click-through rates

What Cited Brands Have in Common

Our site audit of 110 cited brands revealed a clear profile. Brands that appeared in AI responses shared these measurable traits:

The data tells a clear story:

1. Schema markup is table stakes — 97.3% of cited brands had structured data. The 35.67x lift means brands without Schema.org markup are nearly invisible to AI engines.
2. Reviews are non-negotiable — 99.1% of cited brands had review presence. This is the closest thing to a hard requirement for AI citation.
3. FAQ pages provide a meaningful edge — At 1.75x lift, FAQ-structured content maps directly to how AI engines generate responses (62.5% list format, 16.9% step-by-step).
4. Domain authority matters for training-data models — 91.8% of cited brands had high DA. ChatGPT, Claude, and Gemini primarily cite brands from their training data, favoring established entities.
5. Pricing pages are optional — Most cited brands (70%) lacked pricing pages. However, AI engines still surfaced pricing data from other sources in 49.4% of responses.

What Non-Cited Brands Are Missing

Conversely, the brands absent from AI responses — overwhelmingly local service businesses — lacked:

1. Structured data markup — The 3 cited brands without Schema were local businesses. Non-cited local businesses almost universally lack JSON-LD and Schema.org markup, removing them from AI engines' structured data pipeline entirely.
2. Review platform presence — No profiles or minimal reviews on Yelp, Google, Angi, or industry-specific directories. Without reviews, AI engines have no trust signal to cite.
3. FAQ-structured content — Content organized around brand messaging ("Why choose us") rather than the questions buyers ask ("How much does HVAC maintenance cost"). AI engines generated Q&A-style responses 79.4% of the time (list + step-by-step) — content that doesn't match this format won't be extracted.
4. Entity consistency — Cited brands used the same name everywhere. Non-cited brands often had inconsistent naming across directories, their own site, and social profiles — making entity resolution harder for AI models.

Recommendations

Based on these findings, local service businesses and B2B companies should prioritize these actions in order of measured impact:

1. Implement Schema.org markup immediately — Because 97.3% of cited brands had structured data (35.67x lift), this is the highest-impact technical change. Add Organization, LocalBusiness, or SoftwareApplication schema with JSON-LD. This is the single most actionable finding in our dataset.
2. Build review volume across platforms — 99.1% of cited brands had reviews. Claim profiles on Google Business, Yelp, Angi (for local services) or G2, Capterra, TrustRadius (for SaaS). Actively solicit reviews from customers. AI models treat review presence as a near-mandatory trust signal.
3. Add FAQ pages with Schema markup — FAQ-structured content provided a 1.75x citation lift. Create FAQ pages that answer the exact queries your customers type into AI engines. Mark them up with FAQPage schema so AI models can extract Q&A pairs directly.
4. Claim and optimize directory profiles — Because 48.2% of all citations were directory listings, your presence on Yelp, Google Maps, Angi, Avvo, and industry-specific directories is critical. AI engines recommend where to find businesses more than they recommend specific businesses for local queries.
5. Structure content as lists and tables — 94.4% of AI responses used list, step-by-step, or table formats. Restructure service pages and blog posts to match: use numbered lists, comparison tables, and clear H2/H3 hierarchies rather than paragraph-heavy marketing copy.
6. Optimize for Perplexity specifically — As the only platform providing source URLs (100% of Perplexity citations included links), it drives actual referral traffic. Ensure your site is crawlable, fast, and answers specific queries in the first 2 sentences of each section.

15. Live Validation: What Happens to Bot Traffic After AEO Implementation

The findings above are drawn from analyzing other brands' citation patterns. This section documents what happened to AEOfix.com itself after implementing the same signals we identified as critical: Schema markup, structured content, FAQ pages, and a correctly configured robots.txt.

We tracked every bot visit using AEOfix's own AI Access — a 1x1 pixel that logs bot name, category, page, and timestamp to a live database. The data below covers the 30 days following a content and Schema push in late January / early February 2026.

The growth is entirely attributable to structural AEO changes — new content, Schema markup, internal linking, and robots.txt signals. The site went from functionally invisible to crawled by 47 bots in the same window.

AI Systems Account for 54% of All Bot Traffic

Of 2,662 total visits, the majority came from AI systems — not traditional search crawlers:

This is the core proof point for the AEO thesis: a site built for AI engine visibility is, in fact, being prioritized by AI crawlers. Traditional search crawlers (Search Index category) account for only 21% of visits.

The Crawl Inflection: February 22

Bot discovery did not grow gradually — it arrived as a step-change event tied directly to the Schema and content deployment:

The Feb 22 spike is when AI crawlers, search bots, and social bots discovered the site simultaneously — consistent with a sitemap resubmission and new pages becoming indexable after the Schema markup pass. Google impressions followed ~10 days later (Mar 3–4), which is the typical lag between bot crawl discovery and index registration.

Every Major AI Engine Is Represented

Meta AI Search (729 visits) crawled the site 4× more than all Googlebot variants combined as a single crawler. This is unusual and suggests Meta's AI products are aggressively building their retrieval corpus. Every major AI engine — Meta, OpenAI, Perplexity, Groq, Anthropic, Amazon, Apple, Google — was active within 30 days of the content push.

ChatGPT User: Verified AI Referral Traffic

The ChatGPT-User user-agent identifies visits from humans who clicked a link in a ChatGPT response. These are verified AI-to-human referrals — the end state every AEO implementation is working toward. Pages they landed on:

The /optimize-for-chatgpt page pulling 14 visits is the strategy validating itself — people asking ChatGPT "how do I optimize for ChatGPT" are being sent to AEOfix. This is direct proof that the AEO signals identified in this study (Schema, FAQ structure, entity consistency) produce measurable AI citations within 30 days of implementation.

What This Data Proves

1. AEO signals trigger AI crawler discovery — The site went from 1 bot visit to 2,662 in one month, with the inflection point directly tied to the Schema and content deployment. No paid links. No PR campaign.
2. Bot crawl precedes indexing by ~10 days — The Feb 22 bot spike preceded the March 3–4 Google impression spike. Monitoring bot crawl activity is an early warning signal for ranking changes — weeks before GSC shows movement.
3. AI crawlers are the majority of bot traffic on AEO-optimized sites — 54% AI systems vs. 21% traditional search index. Sites not built for AI visibility are receiving the opposite ratio.
4. AI referral traffic is measurable today — 78 ChatGPT-referred human visits in 30 days from a site with no paid traffic and no backlink campaign. The pipeline from AEO implementation → bot crawl → AI citation → human referral is confirmed and repeatable.
5. 80% revisit rate confirms content quality signals — Bots returning to re-index pages at 80% rate indicates the content is being prioritized in their crawl queues. Re-indexing frequency is a precursor to citation frequency.

Full Data

The complete dataset of 160 AI responses is available in our raw data repository. The CSV tracker with extracted metrics is available in ai-visibility-tracker.csv.

This study was conducted by William Bouch at AEOfix.com on February 2, 2026. For methodology questions or to request the raw dataset, contact AEOfix.com@gmail.com.