AI Search Has a Tiny Universal Core and a Massive Specialist Long Tail: How Citation Authority Fragments Across 283 Commercial Markets

Answer Capsule

AI citation authority is not distributed like one universal web. It is structured as a very small broad core surrounded by an enormous market-specific and platform-specific long tail.

Across the strict 283-market benchmark, 308,807 distinct registrable domains appeared in the retained citation data. Of those, 192,753 domains, 62.42%, appeared in only one market. At the opposite extreme, just five domains appeared somewhere in all 283 markets: reddit.com, youtube.com, google.com, facebook.com, and wikipedia.org.

The long tail remained large even after the analysis was restricted to the strongest source portfolios. Across 1,655 eligible market-platform Top-100 portfolios, 36,273 distinct domains appeared. 24,381 of them, 67.22%, reached a Top 100 in only one market. Among the 2,743 domains that reached a Top 10, 1,873, 68.28%, did so in only one market.

That does not mean niche domains dominate total citation activity. The 192,753 single-market domains generated 600,437 resolved citation occurrences, or 6.89% of the 8,718,629 occurrences in the strict cohort. They also accounted for 541,613 response-domain appearances, or 7.93% of the 6,825,821 response-domain appearances. The long tail is enormous by source count, but broader recurring domains account for a much larger share of observed citation activity.

Platform fragmentation is similarly severe. In the full-domain presence analysis, 207,554 domains, 67.21%, appeared on only one of the six AI surfaces, while just 6,084 domains, 1.97%, appeared somewhere on all six. In the stricter Top-100 population, 21,698 of 36,273 domains, 59.82%, appeared in an eligible Top 100 on only one platform, while 1,621, 4.47%, appeared in a Top 100 somewhere on all six platforms.

The fragmentation becomes even clearer when the six platforms are examined inside the same market. Among the 260 markets with eligible Top-100 portfolios on all six surfaces, the median market had 310 distinct domains across the union of its six Top-100 lists. The median market had:

• 185 domains appearing on only one platform
• 126 appearing on at least two platforms
• 73 appearing on at least three platforms
• only 14 appearing on all six platforms

Those 14 all-six domains represented a median of only 4.41% of the market's six-platform Top-100 source union.

The practical implication is that global citation rankings and local citation authority answer different questions. A globally broad source such as Forbes, Reddit, YouTube, or NerdWallet can matter across many markets. But a domain that appears in only one market can still be a deep, six-platform authority inside that category. HearingTracker, WATI, Action Behavior Centers, BookMyForex, and several other examples reached eligible Top-100 portfolios on all six platforms while remaining Top-100 authorities in only one market.

For CMOs, publishers, agencies, and investors, the correct unit of analysis is therefore not simply “which websites are cited most by AI?” It is the intersection of market breadth, platform breadth, local depth, persistence, and the actual business question being answered.

This is a descriptive study of observed citation portfolios. Domain breadth does not establish editorial quality, independence, audience size, traffic, recommendation influence, commercial value, or causation.

Key Findings

The Universal AI Citation Core Is Extremely Small

Questions This Section Answers

1. Which domains appeared across all 283 commercial markets in the AI citation benchmark?
2. How many websites have genuinely broad citation authority across commercial AI search markets?
3. Does appearing in every AI search market mean a domain is a leading source in every market?

The broadest possible definition of universality is simple: did the domain appear at least once somewhere in every qualifying market?

Only five domains cleared that bar:

• reddit.com
• youtube.com
• google.com
• facebook.com
• wikipedia.org

Five domains out of 308,807 is effectively a microscopic universal core.

But even that definition is generous because any presence can include a low-frequency citation that never becomes one of the most prominent sources in the market.

The picture changes when authority is measured by portfolio depth.

At the Top-100 level, only Reddit, YouTube, and Google entered at least one eligible platform Top 100 in all 283 markets.

At the Top-10 level, no domain appeared in a Top 10 in all 283 markets.

That distinction is important. “Appears everywhere” is not the same as “matters deeply everywhere.”

Google.com provides the clearest example. It appeared somewhere in all 283 markets and entered an eligible Top 100 in all 283. But it reached a Top 10 in only 132 markets. Reddit reached a Top 10 in 280, and YouTube in 269.

Wikipedia appeared somewhere in every market, yet reached an eligible Top 100 in 276 markets and a Top 10 in only 91.

The universal core therefore becomes progressively smaller as the authority requirement becomes more demanding.

This is why global citation leaderboards need a depth dimension. A domain can have extraordinary breadth while occupying very different positions across markets.

The Long Tail Is Not a Fringe Phenomenon. It Is Most of the Domain Population

Questions This Section Answers

1. What percentage of AI-cited domains appeared in only one commercial market?
2. How concentrated is AI citation-source breadth across the 283-market benchmark?
3. Does the specialist long tail disappear when analysis is restricted to prominent Top-100 sources?

The full-domain population contains 308,807 distinct registrable domains.

Of those, 192,753 appeared in exactly one market.

That is 62.42% of every distinct domain observed in the strict benchmark.

The opposite tail is tiny.

The structure is therefore not a smooth continuum in which thousands of domains have similar national breadth.

It looks much more like a very steep pyramid:

• a handful of extremely broad domains;
• a relatively small group with meaningful cross-market breadth;
• and a vast population observed in narrow market contexts.

A common objection is that the long tail may simply consist of weak, incidental domains far below the sources that actually matter.

The Top-100 analysis shows that this explanation is incomplete.

Across 1,655 eligible market-platform Top-100 portfolios, 36,273 distinct domains entered at least one Top 100.

Of those:

• 24,381, 67.22%, appeared in a Top 100 in only one market;
• only 1,081, 2.98%, appeared in Top-100 portfolios across at least 10 markets;
• only 329, 0.91%, reached at least 25 markets;
• only 116, 0.32%, reached at least 50 markets;
• only 36, 0.10%, reached at least 100 markets;
• only 13, 0.04%, reached at least 200 markets;
• and only three domains reached a Top 100 in all 283 markets.

The same pattern exists at the Top-10 level.

Among 2,743 distinct domains that entered an eligible Top 10, 1,873, 68.28%, were Top-10 sources in only one market.

This is one of the strongest findings in the breadth analysis.

The specialist long tail is not simply a collection of marginal citations outside the meaningful source set. Market specificity remains the dominant pattern even among domains that become prominent enough to enter Top-100 and Top-10 portfolios.

A Huge Long Tail by Domain Count Does Not Mean the Long Tail Produces Most Citations

Questions This Section Answers

1. Do single-market domains account for most AI citation activity even though they make up most distinct sources?
2. What is the difference between counting AI citation domains and counting citation occurrences?
3. Can a small set of broad domains still account for a large share of AI citation activity?

The domain-count result needs an important qualification.

The 192,753 single-market domains produced 600,437 resolved citation occurrences.

That is only 6.89% of the 8,718,629 resolved citation occurrences in the strict cohort.

They produced 541,613 response-domain appearances, or 7.93% of 6,825,821 response-domain appearances.

So two facts are simultaneously true:

1. most distinct domains are narrow by market breadth;
2. most citation activity comes from domains that appear in more than one market.

That is exactly why a single metric such as “number of markets cited” or “total citation occurrences” is insufficient.

A globally broad domain can accumulate enormous citation volume because it participates in many categories. A category specialist can have a relatively small global footprint while being one of the most important sources in the specific market that matters to a company.

For a CMO choosing where to spend attention, the relevant question is not whether a domain contributes a large share of all AI citations across the economy.

It is whether that domain has meaningful authority in the company's actual category, on the platforms that matter, and across the buying questions the company is trying to influence.

The Broadest Domains Are Not Equally Deep

Questions This Section Answers

1. Which domains have the broadest Top-100 AI citation authority across commercial markets?
2. Which globally broad AI citation sources also reach the Top 10 most often?
3. Why can a website appear in many AI citation markets without being a dominant source in those markets?

Breadth and depth can diverge dramatically.

The following table compares the number of markets where selected domains entered at least one eligible platform Top 100 with the number where they entered an eligible Top 10.

Reddit and YouTube combine extreme breadth with extraordinary depth. Their broad authority is not merely incidental presence.

Forbes also combines breadth and depth, appearing in an eligible Top 100 in 273 markets and a Top 10 in 184.

NerdWallet is narrower globally, with Top-100 presence in 132 markets, but it reached a Top 10 in 97 of them. Its authority is therefore less universal than Forbes but deeper in a large share of the markets where it appears.

LinkedIn and Quora illustrate the opposite pattern. LinkedIn reached Top-100 portfolios in 248 markets but Top-10 portfolios in only 28. Quora reached Top-100 portfolios in 229 markets but the Top 10 in only 14.

The difference becomes even sharper with several broad but shallow domains:

• cbinsights.com entered a Top 100 in 109 markets but did not enter a Top 10 in any market in the strict portfolio analysis;
• bing.com entered a Top 100 in 96 markets but no Top 10;
• businesswire.com entered a Top 100 in 91 markets but no Top 10;
• medium.com entered a Top 100 in 145 markets but a Top 10 in only five;
• apple.com entered a Top 100 in 149 markets but a Top 10 in only seven.

None of those observations implies that the domains are weak publishers, weak brands, or poor marketing opportunities. The data supports a narrower claim: market breadth and local citation depth are distinct dimensions of AI authority.

The companion analysis Citation Breadth vs Citation Depth: Which AI Authority Metric Should CMOs Care About? develops that distinction in more detail.

Fewer Than 1% of Top-100 Domains Reached 25 Markets, but They Generated More Than One-Fifth of Market Appearances

Questions This Section Answers

1. How much of Top-100 AI citation activity is generated by the small group of domains with broad market coverage?
2. Do a few globally broad domains account for a disproportionate share of AI Top-100 market appearances?
3. How concentrated is broad AI citation authority among domains that enter Top-100 portfolios?

The long tail dominates the number of distinct domains. Broad domains still contribute disproportionate portfolio coverage.

There were 88,197 domain-market Top-100 appearances across the 36,273 unique Top-100 domains. This counts each domain once per market in which it reached at least one eligible platform Top 100.

The implication is a two-layer source economy.

The broad core is tiny by domain count, but it contributes a large amount of repeated cross-market presence.

The long tail is enormous by source count, but each individual tail domain usually contributes to a small number of markets.

For strategic measurement, both layers matter.

A global authority score will naturally favor broad recurring domains. A local category strategy can miss important specialist sources if it relies only on that global score.

Platform Breadth Is Just as Fragmented as Market Breadth

Questions This Section Answers

1. How many AI-cited domains appeared on only one of ChatGPT, Gemini, Perplexity, Copilot, Google AI Overviews, or Google AI Mode?
2. What percentage of AI citation domains appeared somewhere on all six platforms?
3. Does appearing on all six AI platforms mean a domain is broadly authoritative within individual markets?

Market breadth is only one axis. Platform breadth creates a second layer of fragmentation.

In the full-domain presence analysis:

67.21% of all observed domains appeared on only one AI surface.

Only 1.97% appeared somewhere on all six.

The strict Top-100 portfolios show the same basic shape.

Among the 36,273 domains that reached at least one eligible Top 100:

• 21,698, 59.82%, appeared in Top-100 portfolios on only one platform;
• 6,445 appeared on two;
• 3,202 appeared on three;
• 1,964 appeared on four;
• 1,343 appeared on five;
• 1,621, 4.47%, appeared in a Top 100 somewhere on all six platforms.

At Top 10, the population was smaller but still fragmented:

• 1,434 of 2,743 domains, 52.28%, entered a Top 10 on only one platform;
• 142, 5.18%, entered a Top 10 somewhere on all six.

These are global platform-breadth counts. They do not mean that a domain appearing on all six surfaces did so in the same market.

A domain might be cited by ChatGPT in finance, Gemini in software, Perplexity in health, and Google AI in consumer products. That creates six-platform global breadth without creating six-platform authority inside any one category.

This is why Six AI Engines, Six Different Webs measures source overlap within markets instead of treating global platform presence as portability.

Inside a Single Market, the Six AI Platforms Still Draw From Very Different Source Pools

Questions This Section Answers

1. How many distinct Top-100 domains does a typical commercial market accumulate across six AI platforms?
2. What share of a market's AI citation sources are unique to one platform versus shared by all six?
3. How large is the true six-platform source core inside a commercial market?

The strongest demonstration of fragmentation comes from the 260 markets with eligible Top-100 portfolios on all six AI surfaces.

For each of those markets, the six Top-100 lists were combined into a union and each domain was counted by the number of platforms on which it appeared.

The median market contained 310 distinct domains across those six Top-100 portfolios.

The median composition was:

The middle half of complete markets had between 298 and 331 distinct domains in the six-platform union.

The middle half had only 10 to 17 domains shared by all six platforms.

That means a typical market can have hundreds of important cited domains across the six systems while the genuinely common all-six core is only a small fraction of the source universe.

This is the market-level version of the same phenomenon visible globally: a small shared core surrounded by a much larger specialist or platform-specific periphery.

Some Markets Have Ten Times More Six-Platform Source Agreement Than Others

Questions This Section Answers

1. Which commercial markets had the largest shared citation-source core across all six AI platforms?
2. Which markets had almost no Top-100 domains shared by all six AI platforms?
3. How much can six-platform citation-source agreement vary by industry?

The size of the all-six core varies substantially by market.

Markets with larger all-six Top-100 cores

Even these relatively aligned markets remain far from universal agreement. Credit Repair had the largest all-six source core in this analysis at 29 domains, meaning 252 other domains appeared somewhere in its six-platform Top-100 union without appearing across all six.

Markets with tiny all-six Top-100 cores

Dental Implants accumulated 409 distinct Top-100 domains across the six systems, yet only two appeared on all six.

Car Accident Lawyers accumulated 410, with only three shared by all six.

Those are not small differences around one global benchmark. They are structurally different citation environments.

A company in Credit Repair and a company in Dental Implants should not use the same assumptions about how transferable publisher authority will be across platforms.

The industry implications are developed further in There Is No Universal GEO Strategy: How AI Citation Persistence and Portability Differ by Industry Across 283 Markets.

A Domain Can Be a One-Market Specialist and Still Be a Six-Platform Authority

Questions This Section Answers

1. Can a niche website be a Top-100 source on all six AI platforms while having little global market breadth?
2. Which specialist domains show deep AI citation authority inside one commercial category?
3. Why can global citation rankings understate the strategic importance of niche publishers and category sources?

A global breadth metric can make category specialists look insignificant.

The portfolio data shows why that can be misleading.

The following domains entered eligible Top-100 portfolios in only one market in the strict benchmark, yet appeared in that market's Top 100 on all six AI platforms.

These are exactly the kinds of sources a global citation ranking can bury.

HearingTracker is a particularly clean example. In the Hearing Aids market it ranked:

• #1 on ChatGPT
• #4 on Gemini
• #1 on Google AI Mode
• #2 on Google AI Overviews
• #1 on Microsoft Copilot
• #1 on Perplexity

WATI showed a similarly deep market-specific pattern in AI Chatbots:

• #1 on ChatGPT
• #2 on Gemini
• #7 on Google AI Mode
• #7 on Google AI Overviews
• #1 on Microsoft Copilot
• #1 on Perplexity

Lifeway Mobility illustrates a slightly broader local authority profile in Stairlifts. It ranked:

• #20 on ChatGPT
• #1 on Gemini
• #1 on Google AI Mode
• #1 on Google AI Overviews
• #5 on Microsoft Copilot
• #16 on Perplexity

The point is not that any of these domains is objectively the “best” publisher or business in its category.

The point is that global market breadth can dramatically understate local AI citation authority.

The supporting article Why Niche Publishers Matter More in AI Search Than Global Citation Rankings Suggest focuses on this category-specialist effect.

Broad Publishers Still Matter, but Their Value Comes From a Different Kind of Authority

Questions This Section Answers

1. When should a CMO favor a broad AI-cited publisher over a category specialist?
2. How do broad domains such as Forbes, NerdWallet, Money.com, G2, and Healthline differ from single-category sources?
3. Should publisher selection optimize for global market breadth or local citation depth?

The specialist-long-tail finding should not be interpreted as an argument against broad publishers.

Broad domains offer something category specialists often cannot: repeated authority across many commercial contexts.

For example:

• Forbes entered Top-100 portfolios in 273 markets and Top-10 portfolios in 184.
• NerdWallet reached 132 Top-100 markets and 97 Top-10 markets.
• Money.com reached 114 Top-100 markets and 61 Top-10 markets.
• G2 reached 86 Top-100 markets and 49 Top-10 markets.
• Healthline reached 96 Top-100 markets and 40 Top-10 markets.
• Bankrate reached 89 Top-100 markets and 41 Top-10 markets.

That kind of breadth can be valuable to an advertiser operating across multiple adjacent products or to a publisher trying to demonstrate portfolio-level authority.

But broad authority is not a substitute for market fit.

A CMO deciding between a broad publisher and a category specialist should ask at least four separate questions:

1. Breadth: In how many relevant markets does the domain appear?
2. Depth: How highly does it rank within this specific market's source portfolio?
3. Platform breadth: Which AI surfaces cite it in this market?
4. Persistence: Does the source remain prominent across multiple observation periods?

The 2026 AI Commercial Citation Authority Index is designed around those separate dimensions rather than one arbitrary composite score.

Why a Global “Top AI-Cited Websites” List Can Mislead a Marketing Team

Questions This Section Answers

1. Why can a global list of the most-cited websites in AI search mislead a CMO choosing publishers?
2. What does a global AI citation ranking miss about market-specific source authority?
3. How should marketing teams combine global citation breadth with category-level publisher research?

A global ranking naturally rewards domains that participate in many markets.

That is useful if the question is:

Which domains have the broadest observed citation footprint across commercial AI search?

It is incomplete if the question is:

Which publishers or sources matter most for my product category?

The evidence shows why.

A domain such as LinkedIn can enter Top-100 portfolios across 248 markets while reaching the Top 10 in only 28. A niche source can appear in a Top 100 in only one market and still rank near the top of all six AI platforms inside that category.

Those two domains are not competing for the same kind of authority.

A useful source-selection process therefore needs two lanes:

Broad-authority lane

Identify domains with meaningful breadth across relevant adjacent markets and platforms.

These can be useful for:

• multi-category brands;
• corporate reputation programs;
• broad editorial distribution;
• investor or enterprise visibility;
• cross-category research;
• and large media partnerships.

Category-authority lane

Identify domains that are unusually deep, persistent, or portable inside the exact market.

These can be useful for:

• product-level comparisons;
• high-intent buying questions;
• vertical PR;
• category sponsorships;
• specialist editorial partnerships;
• and highly targeted authority-building.

The two lanes should then be reconciled with actual audience quality, editorial standards, ownership, commercial access, pricing, and brand fit.

Citation data tells a marketing team where AI systems used sources. It does not automatically determine where marketing dollars should go.

The Long Tail Also Changes How Publishers Should Think About Expansion

Questions This Section Answers

1. Can a specialist publisher build valuable AI citation authority without becoming a general-interest media company?
2. How should a niche publisher decide which adjacent categories to expand into?
3. Does AI citation breadth reward publishers for being generalists, or can category depth be commercially valuable on its own?

The breadth data suggests that niche authority is not an edge case. It is a normal feature of the observed citation ecosystem.

That creates an important strategic option for publishers.

A specialist publisher does not necessarily need to become a generalist site to build an AI-relevant commercial asset.

Instead, it can deepen authority in one market and expand selectively into adjacent categories where its editorial expertise and source credibility are transferable.

The evidence supports a progression such as:

1. establish depth in one category;
2. demonstrate persistence across time;
3. document which AI platforms cite the domain;
4. identify adjacent markets with similar buying questions or source ecosystems;
5. expand where there is genuine editorial fit;
6. measure whether authority transfers rather than assuming it will.

The article How Publishers Can Build Category Authority Without Becoming a Generalist Media Site focuses on that expansion path.

The evidence does not say that every narrow domain is valuable. The full long tail includes company sites, retailers, institutions, communities, niche publishers, incidental sources, and domains with very little citation activity.

The useful subset is the long-tail domain that combines local depth with recurrence, cross-platform presence, or clear editorial relevance.

A Practical Market-First Source Selection Framework

Questions This Section Answers

1. How should a CMO identify the publishers AI systems actually rely on in a specific industry?
2. What evidence should a marketing team collect before sponsoring content on an AI-cited domain?
3. How can a company separate broad AI citation authority from market-specific publisher opportunity?

A practical workflow starts with the market rather than the publisher.

Step 1: Define the commercial market and buying questions

The category needs to be narrow enough that the source environment is strategically meaningful.

“Financial services” is too broad for many decisions. Mortgage, Credit Repair, Business Checking Accounts, Life Insurance Companies, and Money Transfer Fintech can have very different source ecosystems.

Step 2: Build platform-specific source portfolios

Do not assume ChatGPT's sources represent Gemini, Copilot, Perplexity, or Google AI.

The Persistence vs Portability study shows why.

Step 3: Separate broad domains from category specialists

Use market breadth to identify globally recurring sources and local depth to identify category authorities.

Step 4: Add persistence

A one-time Top-10 appearance is different from a domain that remains prominent month after month.

The month-to-month persistence study provides the durability benchmark.

Step 5: Inspect the cited URLs

Domain-level authority is not enough for a media decision.

Review:

• which URLs were cited;
• the topics they answer;
• editorial ownership;
• publication dates;
• whether the page is commercial or editorial;
• whether sponsored content is disclosed;
• and whether the source is relevant to the company's actual customer questions.

Step 6: Measure recommendation authority separately

A source can be cited without a company being recommended.

A company can be recommended without its own domain being cited.

The distinction is developed in AI Citation Authority vs AI Recommendation Authority.

Step 7: Keep global breadth as context, not as the decision rule

Global breadth can help identify influential publishers and platforms. Local depth determines whether they matter enough in the specific market to deserve further diligence.

Methodology

The core benchmark covers completed observations from July 1 through September 30, 2026 across six consumer AI surfaces:

• ChatGPT
• Gemini
• Perplexity
• Microsoft Copilot
• Google AI Overviews
• Google AI Mode

The strict qualifying cohort contains:

• 283 qualifying tracked markets
• 876,605 retained AI responses
• 8,718,629 citation occurrences resolved to registrable domains
• 6,825,821 response-domain appearances
• 308,807 distinct registrable domains

Full-domain market breadth

Market breadth in the full-domain analysis means that the registrable domain appeared in at least one retained response-domain observation assigned to the qualifying market under the frozen analysis rules.

This is an any-presence measure. It does not require a domain to reach a Top 100 or Top 10.

Full-domain platform breadth

Platform breadth means the domain had at least one observed presence on the AI surface somewhere in the qualifying cohort.

This any-presence measure can include cells that are too thin for strict portfolio comparisons.

Six-platform global breadth therefore does not mean that a domain appeared on all six platforms inside every market.

Top-100 and Top-10 market breadth

For the stricter depth analysis, each eligible market-platform portfolio was ranked by distinct response presence.

A platform cell required at least:

• 100 retained responses;
• 100 resolved registrable domains;
• and an actual ranked list of exactly 100 domains.

There were 1,655 eligible market-platform Top-100 portfolios across the 283 markets and six surfaces.

A domain's Top-100 market breadth is the number of markets where it entered at least one eligible platform Top 100.

Top-10 market breadth uses the first 10 domains from those same eligible portfolios.

Within-market six-platform analysis

There were 260 markets with eligible Top-100 portfolios on all six platforms. This population is distinct from the stricter 246-market subset that also requires all three monthly portfolios.

For the six-platform breadth analysis, the six Top-100 sets were unioned within each market, and each domain was counted by the number of platform portfolios in which it appeared.

Important population distinction

The full-domain presence population and the strict Top-100 portfolio population answer different questions.

The full-domain population measures observed breadth anywhere in the retained data.

The Top-100 population measures prominence inside cells that meet the strict support threshold.

The two should not be substituted for each other.

Limitations

This analysis supports descriptive claims about domain breadth and source-portfolio depth. It does not establish several stronger claims.

Domains are not synonymous with publishers

A registrable domain can represent:

• a publisher;
• a company;
• a retailer;
• a government entity;
• a community;
• a marketplace;
• a platform;
• an academic institution;
• or another source type.

The term “specialist long tail” describes narrow observed market breadth. It is not a verified classification of every domain as a specialist publisher.

Breadth does not establish quality

A domain appearing in many markets is not necessarily more accurate, independent, editorially rigorous, or commercially valuable than a narrow domain.

A narrow domain is not necessarily more expert.

Top-100 and Top-10 membership is cutoff-sensitive

A domain ranked #11 is excluded from a Top-10 analysis even if its response presence is nearly identical to #10. A domain ranked #101 is excluded from the Top 100.

Portfolio cutoffs are useful measurement tools, not natural boundaries in source quality.

The study is observational

The historical collection does not establish an identical panel of prompts, model versions, countries, retrieval settings, and collection configurations across every market and platform.

The study therefore describes observed citation portfolios rather than controlled causal platform effects.

Citation authority is not recommendation authority

Being used as evidence does not prove that an AI system recommended a company, product, or service.

Breadth is not future performance

Observed market or platform breadth does not guarantee that the same domain will retain its position in future months.

Persistence must be measured separately.

Global platform breadth is not local portability

A domain can appear somewhere on all six platforms while appearing on only one or two platforms inside a specific market.

Cross-platform portability must be measured within the relevant market.

Conclusion

The 283-market benchmark shows that AI citation authority has two very different layers.

The first is a small broad core.

Reddit, YouTube, Google, Forbes, Wikipedia, Facebook, NerdWallet, Money.com, and a relatively small number of other domains appear across large portions of the commercial AI citation landscape.

The second is a massive long tail.

62.42% of all distinct domains appeared in only one market. Even inside the strict Top-100 portfolio universe, 67.22% of domains were Top-100 sources in only one market. Among Top-10 sources, 68.28% were Top-10 authorities in only one market.

The same fragmentation appears across platforms. 67.21% of the full-domain population appeared on only one surface, and in the stricter Top-100 universe, 59.82% of domains appeared on only one platform.

Inside a typical six-platform-complete market, the fragmentation becomes tangible. The median market accumulated 310 distinct Top-100 domains across the six systems. 185 appeared on only one platform. Only 14 appeared on all six.

That is why a global citation leaderboard cannot substitute for market-specific source intelligence.

The broad core matters because it accounts for substantial repeated citation activity and can offer cross-market reach.

The specialist long tail matters because it contains domains that can be extremely deep within a narrow category. A one-market domain can still rank in the Top 10 across all six AI surfaces.

For CMOs, the strategic unit is therefore market × platform × depth × time.

For publishers, the opportunity is equally clear: category authority can be commercially meaningful even without universal breadth, provided it can be documented with evidence rather than vague claims about “AI visibility.”

And for researchers, the result argues against treating AI search as one unified citation ecosystem. The observed web is far more fragmented, specialized, and market-dependent than a single global ranking suggests.

Related LLM Authority Index Research

• The Persistence-Portability Gap at Scale establishes the 283-market methodological benchmark.
• Why Niche Publishers Matter More in AI Search Than Global Citation Rankings Suggest focuses on market-specific authority hidden by global rankings.
• Citation Breadth vs Citation Depth: Which AI Authority Metric Should CMOs Care About? separates cross-market reach from local prominence.
• How to Find the Market-Specific Publishers AI Systems Actually Cite provides a market-first source-discovery framework.
• How Publishers Can Build Category Authority Without Becoming a Generalist Media Site examines specialist publisher expansion.
• How Publishers Can Turn Persistent AI Citation Authority Into a Sponsorship and Distribution Asset examines commercialization of durable authority.
• Six AI Engines, Six Different Webs measures within-market platform fragmentation.
• The 2026 AI Commercial Citation Authority Index combines separate breadth and depth dimensions without collapsing them into an arbitrary score.
• AI Citation Authority vs AI Recommendation Authority separates evidence-source authority from company recommendation behavior.
• The 2026 AI Citation Authority Study documents the earlier high-stakes consumer research population.

About LLM Authority Index

LLM Authority Index measures how AI systems cite sources, surface publishers, and structure authority across commercial markets. Its research separates market breadth, platform breadth, source depth, persistence, portability, citation authority, and recommendation authority so that a globally visible domain is not automatically mistaken for the most important source in every category.