Can One GEO Strategy Work Across ChatGPT, Gemini, Perplexity, Copilot, Google AI Overviews, and Google AI Mode?

Answer Capsule

A company can run one coordinated GEO program across ChatGPT, Gemini, Perplexity, Microsoft Copilot, Google AI Overviews, and Google AI Mode, but the data does not support using one undifferentiated publisher list, one source map, or one platform-agnostic visibility score.

The reason is structural.

Across the 260 markets in this study with complete eligible Top-100 portfolios on all six AI surfaces, the median market contained 310 distinct domains across the six platform lists.

Most of that source universe was not broadly reusable across platforms:

• median 185 domains appeared in only one of the six Top-100 portfolios;
• median 126 appeared on at least two platforms;
• median 73 appeared on at least three;
• median 46 appeared on at least four;
• median 28 appeared on at least five;
• median only 14 appeared on all six.

Expressed as a share of the market-level source universe, the median market was composed of roughly:

• 59.67% single-platform sources;
• 25.41% sources shared by only two or three platforms;
• 14.90% sources appearing on four or more platforms;
• and only 4.41% sources appearing on all six.

The fragmentation remained even when the analysis focused on the most prominent sources.

At the Top-25 level, the median complete market contained 68 distinct domains, and 34, exactly half, appeared on only one platform. At the Top-10 level, the median market contained 28 distinct domains, and 14 again appeared on only one platform.

The practical operating model is therefore not six separate GEO programs and not one universal program.

It is:

one shared factual and editorial foundation, one market definition, one measurement framework, and platform-specific source extensions where the observed evidence diverges.

The shared layer is real. Google AI Overviews and Google AI Mode were the closest source pair in every one of the 260 complete markets. Some markets such as Mortgage, eSignature Software, Business Checking Accounts, and Credit Monitoring had materially larger four-platform and six-platform shared cores than the benchmark median.

But the platform-specific layer is also real. In the median market, Microsoft Copilot alone contributed 45 Top-100 domains that did not appear in any of the other five Top-100 lists. ChatGPT contributed 36 unique domains, Gemini 36, Perplexity 32, Google AI Overviews 19, and Google AI Mode 18. Copilot was the sole platform with the largest unique Top-100 source contribution in 188 of 260 complete markets, with another 17 markets tied for the lead.

That means a CMO can centralize facts, messaging, content standards, measurement definitions, and a shared set of cross-platform source opportunities. But source acquisition, publisher targeting, competitive source analysis, and recommendation measurement should retain platform-level drill-downs.

This article does not establish that any specific content, PR, schema, sponsorship, or publisher placement causes AI citations. It describes the observed citation-source structure and uses that evidence to define a more defensible way to organize GEO work.

Key Findings

The Wrong Choice Is “One Strategy” Versus “Six Strategies”

Questions This Section Answers

1. Should a CMO run one GEO strategy for every AI platform or create six separate programs?
2. Which parts of AI search optimization can be centralized across ChatGPT, Gemini, Perplexity, Copilot, Google AI Overviews, and Google AI Mode?
3. Which parts of a GEO program need platform-specific measurement and source targeting?

The six-platform data does not support either extreme.

A completely universal program assumes that the same publishers, source domains, cited pages, competitors, and recommendation environment matter everywhere.

A completely separate program assumes there is little meaningful overlap and requires six independent teams, six content systems, six source strategies, and six reporting stacks.

Neither description matches the observed citation structure.

The data instead supports a layered operating model.

Shared layer

A company can centralize what it directly controls:

• company facts;
• product and service descriptions;
• pricing and eligibility information;
• geographic availability;
• evidence standards;
• expert credentials;
• owned content maintenance;
• technical crawlability and indexability work;
• structured internal measurement definitions;
• brand naming conventions;
• competitor definitions;
• and a common research taxonomy.

Those are operational foundations. This benchmark does not test them as ranking factors.

A company can also centralize source opportunities that repeatedly appear across several relevant platforms.

Platform-specific layer

The company should preserve platform-level analysis for:

• source-domain rankings;
• cited URLs;
• publisher shortlists;
• competitive source gaps;
• company recommendation coverage;
• sentiment and positioning;
• platform-specific answer patterns;
• and measurement changes over time.

The evidence for that separation is not theoretical. In the median complete market, nearly 60% of the domains in the combined six-platform Top-100 source universe appeared on only one platform.

That is too much platform-specific source inventory to erase inside one global score.

A Typical Market Has a Small Shared Core and a Large Platform-Specific Perimeter

Questions This Section Answers

1. How much of a typical AI citation-source market is reusable across multiple AI platforms?
2. What percentage of source domains in a six-platform GEO program are platform-specific?
3. How large is the shared citation-source core across ChatGPT, Gemini, Perplexity, Copilot, Google AI Overviews, and Google AI Mode?

Across the 260 complete markets, the median six-platform Top-100 union contained 310 distinct domains.

Those domains distributed across platform coverage as follows.

The exact bucket medians should not be mechanically added because each median is calculated across markets independently. The percentages nevertheless show the structure clearly.

The typical market has three operational layers.

1. Cross-platform core

Domains appearing on four, five, or six platforms represent the most reusable source layer.

The median market had 46 domains appearing on at least four of six platforms, representing 14.90% of the six-platform union.

That is the best starting point for shared source diligence because those domains already demonstrate observed relevance across much of the platform set.

2. Partial-overlap layer

Domains appearing on only two or three platforms made up a median 25.41% of the union.

These sources are neither universal nor purely platform-specific. They may support paired or clustered strategies.

For example, a source could matter to ChatGPT and Perplexity but not Google AI. Another could recur across Gemini and both Google search surfaces. These relationships should be measured, not assumed from brand ownership or product similarity.

3. Platform-specific perimeter

The largest layer consisted of domains appearing in only one platform's Top-100 portfolio.

The median count was 185 domains, or 59.67% of the six-platform source universe.

That is the strongest quantitative reason not to hand one global publisher list to the entire GEO team.

A universal list can capture the shared core. It cannot describe most of the observed source universe.

The Most Prominent Sources Were Still Highly Platform-Specific

Questions This Section Answers

1. Does platform-specific citation behavior disappear when a GEO program focuses only on the Top 25 or Top 10 sources?
2. Are leading AI citation sources more reusable across platforms than the broader Top-100 source universe?
3. How much platform-specific publisher research remains necessary after filtering to high-ranking sources?

A reasonable objection is that the Top-100 lists contain many marginal sources.

Perhaps the platform-specific perimeter disappears when attention narrows to the most prominent domains.

It does not.

Top 25

Across complete markets, the median union of six Top-25 platform portfolios contained 68 distinct domains.

The median number appearing on only one platform was 34.

That means the median single-platform share of the Top-25 union was 50.76%.

Half of the combined Top-25 source universe remained platform-specific.

Top 10

The median six-platform Top-10 union contained 28 distinct domains.

The median number appearing on only one platform was 14.

The median single-platform share was 48.15%.

Even in the highest-visibility source layer, roughly half of the distinct domains in the combined six-platform universe appeared in only one platform's Top 10.

This matters for resource allocation.

A CMO does not need to investigate hundreds of domains at full depth. But even after narrowing to leading sources, the evidence does not support assuming that the same handful of publishers dominate every platform.

The companion analysis, Top 25 vs Top 100 AI Citation Sources: Where Durable Authority Actually Lives, shows that prominent source portfolios are more persistent over time. This article adds the cross-platform management implication: prominence does not eliminate platform specificity.

Microsoft Copilot Creates the Largest Platform-Specific Source Extension in Most Markets

Questions This Section Answers

1. Which AI platform contributes the most unique Top-100 source domains to a six-platform GEO program?
2. Does Microsoft Copilot require more platform-specific publisher research than ChatGPT, Gemini, Perplexity, or Google AI?
3. How many citation sources are typically unique to each AI platform inside the same market?

For each complete market, the analysis counted domains appearing in one platform's Top 100 and in none of the other five Top-100 portfolios.

The median unique contribution was:

Copilot was not merely highest at the median.

It was the sole platform with the largest unique Top-100 contribution in 188 of 260 complete markets. Another 17 markets had ties for the largest unique contribution.

At the Top-25 level, Copilot's median unique contribution was 10 domains, compared with 7 for ChatGPT, 6 for Gemini, 5 for Perplexity, and 2 each for Google AI Overviews and Google AI Mode.

At the Top-10 level, Copilot still had the largest median unique contribution at 4 sources, versus 3 for ChatGPT, 2 for Gemini, 2 for Perplexity, and 1 each for the Google search surfaces.

Copilot was the sole platform with the most unique Top-25 sources in 166 complete markets, and the sole leader at the Top-10 level in 146 markets.

That does not mean Copilot deserves the largest marketing budget.

Citation distinctiveness and commercial audience value are separate questions.

It does mean that a source map built without a Copilot-specific extension will systematically miss a larger set of Copilot-only source relationships than the equivalent omission on most other platforms.

The dedicated Copilot research cluster examines that pattern in greater depth, beginning with Microsoft Copilot Lives on a Different Web.

Google AI Overviews and AI Mode Offer the Most Reusable Pair, but They Still Need Separate Measurement

Questions This Section Answers

1. Can a CMO combine Google AI Overviews and Google AI Mode into one GEO source strategy?
2. How much source research can reasonably be shared between Google's two AI search surfaces?
3. Why should Google AI Overviews and AI Mode still be reported separately if they have the highest citation overlap?

The strongest source-reuse opportunity in the entire study is Google AI Overviews plus Google AI Mode.

Their median Top-100 Jaccard overlap was 53.85%, equivalent to 70 shared domains in two equal Top-100 lists.

More importantly, they were the unique highest-overlap pair in every one of the 260 complete markets.

That supports a practical management decision: Google AI Overviews and AI Mode should share substantial source-research infrastructure.

A team can:

• investigate their shared publishers together;
• compare common source URLs;
• coordinate outreach where the same third-party domains matter to both;
• and maintain one combined research queue for the shared portion of the ecosystem.

But the pair should not be collapsed into one metric.

At the median, 30 domains from each Top-100 list were still not shared with the other surface.

The complete-market sibling advantage also varied substantially. In some markets the two Google surfaces were dramatically more similar to each other than any other pair. In others, another pair came close.

The dedicated Google AI Overviews and AI Mode Are Siblings, Not Twins study focuses on that relationship.

The operating rule is simple:

Share research where the evidence overlaps, preserve reporting where the evidence differs.

Some Markets Support Much More Reuse Than Others

Questions This Section Answers

1. Which commercial markets have unusually reusable AI citation-source portfolios across platforms?
2. Which markets require much more platform-specific GEO work because their citation sources barely overlap?
3. Should a company use the same cross-platform GEO operating model in mortgage, software, legal, health, and other industries?

The amount of reusable source infrastructure varies sharply by market.

Consider the share of the six-platform Top-100 union appearing on at least four platforms.

The benchmark median was 14.90%.

Some markets were much more reusable.

These markets still contained many platform-specific sources, but a larger fraction of the observed source universe could support shared cross-platform research.

Other markets were far more fragmented.

The contrast is operationally significant.

A Mortgage or eSignature program can reasonably invest more heavily in a shared publisher research layer because a larger proportion of important sources recur across several platforms.

A Dental Implants or DUI Lawyers program needs more platform-specific discovery because the combined source universe is much larger and the broadly reusable core is much smaller.

This is why a universal agency playbook can be misleading even if the high-level workflow remains the same.

The workflow can be standardized. The source allocation cannot.

Credit Repair Shows What a Larger Cross-Platform Core Looks Like

Questions This Section Answers

1. What does a relatively reusable six-platform citation-source market look like in practice?
2. How large can the all-six source core become inside a commercial market?
3. Does a large shared source core eliminate the need for platform-specific GEO work?

Credit Repair had one of the largest all-six Top-100 intersections in the complete-market cohort.

Its six-platform Top-100 union contained 281 distinct domains.

Of those:

• 162 appeared on only one platform;
• 50 appeared on at least four platforms;
• 29 appeared on all six.

Twenty-nine shared domains is more than double the benchmark median of 14.

That creates a meaningful shared research core.

A CMO can investigate those 29 domains as a high-priority cross-platform source layer and then expand to the 50 domains appearing on at least four systems.

But the market still had 162 single-platform domains, representing 57.65% of the full six-platform union.

The shared core is larger than normal, but it is not the entire market.

This is a good example of what “one coordinated strategy” means in practice. The team can centralize a larger part of its publisher diligence and still maintain platform-specific extensions.

Dental Implants Shows Why Some Markets Need Stronger Platform-Specific Work

Questions This Section Answers

1. What does an extremely fragmented six-platform citation-source market look like?
2. How many sources can be unique to individual AI platforms in a low-portability market?
3. When should a CMO avoid relying heavily on one shared GEO publisher list?

Dental Implants sits near the opposite end of the spectrum.

Its six-platform Top-100 union contained 409 distinct domains.

Only 18 domains appeared on at least four platforms, and only 2 appeared on all six.

Meanwhile, 286 domains appeared on only one platform, representing 69.93% of the source universe.

The platform-specific contributions were large:

• ChatGPT: 47 unique Top-100 domains;
• Gemini: 51;
• Perplexity: 62;
• Microsoft Copilot: 67;
• Google AI Overviews: 32;
• Google AI Mode: 27.

A single shared publisher list would therefore omit a large part of the observed source environment on every platform.

The Article 11 pillar found the same fragmentation in the pairwise overlap matrix. Google AI Overviews and AI Mode were still the closest pair in Dental Implants, but their overlap was only 45.99%, while Google AI Overviews and Copilot fell to 3.63%.

This is a category where platform-specific source mapping is not a minor optimization. It is a large part of the observed citation environment.

One GEO Program Should Have Three Source Workstreams

Questions This Section Answers

1. How should a marketing team organize GEO source research across six different AI platforms?
2. What belongs in the shared source workstream versus platform-specific GEO workstreams?
3. How can a CMO avoid building six redundant AI search teams while still respecting platform differences?

The evidence supports three source workstreams inside one coordinated program.

Workstream 1: Cross-platform core

Start with sources appearing across four or more priority platforms in the market.

These domains offer the best observed source reuse.

For a complete market, the median count is 46.

Inside this layer, flag the domains appearing on all six platforms. The median market has 14.

This is where shared publisher research, relationship diligence, source quality review, and owned-information consistency can be concentrated first.

Workstream 2: Partial-overlap clusters

Next identify domains that recur on two or three platforms.

The median market has a substantial population in this layer, representing about 25.41% of the total Top-100 source universe.

These sources are candidates for grouped platform strategies rather than universal ones.

Examples could include a ChatGPT-Perplexity source cluster or a Gemini-Google cluster, but the exact grouping should be measured inside the market rather than assumed globally.

Workstream 3: Platform-specific extensions

Finally investigate sources that appear on only one platform.

This is not a small cleanup list. It is the majority of the market-level source universe in the median market.

Prioritize the platform-specific layer by:

• source depth;
• response presence;
• persistence over time;
• commercial relevance;
• competitor visibility;
• page relevance;
• publisher quality;
• and the importance of that AI platform to the business.

A platform-specific source at rank 4 may deserve more attention than a six-platform source that sits near rank 90 everywhere.

That is why breadth and depth should remain separate, as developed in Citation Breadth vs Citation Depth: Which AI Authority Metric Should CMOs Care About?.

What a CMO Can Centralize Without Overclaiming

Questions This Section Answers

1. Which GEO activities can be centralized even when AI platforms cite different sources?
2. What should a company standardize across ChatGPT, Gemini, Perplexity, Copilot, and Google AI measurement?
3. Which shared GEO practices are operationally sensible without claiming they are proven AI ranking factors?

The evidence in this study concerns source portfolios, not causal optimization factors.

That distinction matters.

A coordinated program can still centralize the following for operational reasons:

Company truth layer

Maintain one authoritative internal record for:

• company name and ownership;
• products and services;
• pricing ranges;
• service areas;
• eligibility rules;
• customer segments;
• expert credentials;
• product specifications;
• guarantees and exclusions;
• current leadership;
• and other facts AI systems may encounter across the web.

The purpose is factual consistency, not a claim that a database entry causes AI citations.

Research taxonomy

Define markets, competitors, prompts, recommendation outcomes, citation metrics, and source categories consistently across platforms.

Without a shared taxonomy, cross-platform measurement becomes difficult to interpret.

Editorial standards

Use one standard for evidence quality, source citation, disclosure, review dates, and factual maintenance across owned content.

Again, this is a governance recommendation, not a ranking-factor claim.

Measurement protocol

Standardize:

• collection windows;
• platform labels;
• response counts;
• domain normalization;
• prompt documentation;
• recommendation definitions;
• and change logs.

This creates comparable platform reports even when the results differ.

Shared-source diligence

Investigate the cross-platform source core once, then reuse the research across the platforms where the source actually appears.

The efficiency comes from measured overlap rather than assumed overlap.

What Must Stay Platform-Specific

Questions This Section Answers

1. Which parts of GEO measurement should never be collapsed into one global AI visibility score?
2. Why should publisher outreach and competitor source analysis remain platform-specific?
3. What can be hidden when ChatGPT, Gemini, Perplexity, Copilot, and Google AI results are combined?

Several parts of the program should retain platform-level visibility.

Source rankings

A domain that is Top 5 on ChatGPT may be absent from Gemini.

A global average can hide that distinction.

Cited URLs

Even when the same domain appears on two platforms, the cited pages can differ.

Domain-level overlap does not establish page-level overlap.

Recommendation behavior

Citation authority and recommendation authority are different measurements.

A platform can cite a publisher while recommending a competitor. A platform can recommend a company without citing the company's domain.

Competitive positioning

Competitors can have very different recommendation coverage and source support by platform.

Sentiment and factual inconsistencies

A company may be described accurately on one AI surface and incorrectly on another.

Source opportunity gaps

The median market contains hundreds of platform-specific source relationships. Aggregating them into one list can erase the exact publisher or third-party opportunities that distinguish the platforms.

How to Allocate Effort Across Platforms

Questions This Section Answers

1. How should a CMO allocate GEO resources when AI citation-source ecosystems differ by platform?
2. Should platforms with more unique citation sources automatically receive more budget?
3. What evidence should determine whether a company invests more heavily in ChatGPT, Google AI, Perplexity, Gemini, or Copilot?

Citation distinctiveness should determine measurement granularity, not budget by itself.

Copilot contributed the largest unique source extension in most markets, but that does not prove Copilot deserves the largest investment.

Budget decisions require external business inputs such as:

• customer usage;
• referral traffic;
• assisted conversions;
• lead quality;
• search volume or query exposure where measurable;
• sales attribution;
• audience demographics;
• and strategic market importance.

The citation data answers a narrower question:

If we care about this platform, how much of its source ecosystem can be covered by shared work versus platform-specific work?

A reasonable resource process is:

1. rank platforms by business importance;
2. quantify the shared source core in the relevant market;
3. investigate that core once;
4. identify high-depth platform-specific source gaps;
5. prioritize those gaps based on competitive and commercial relevance;
6. measure recommendations separately;
7. review changes over time.

That approach minimizes redundant work without pretending that one platform's source map represents the others.

The Executive Dashboard Should Be Unified at the Top and Separated Underneath

Questions This Section Answers

1. How should a CMO report GEO performance across six AI platforms without hiding platform differences?
2. Can an executive AI visibility dashboard use one overall score while preserving platform-level evidence?
3. Which cross-platform GEO metrics should be summarized centrally and which should remain drill-downs?

Executives need a concise summary.

Analysts need the underlying platform differences.

Both can exist in one reporting system.

A practical dashboard can have three levels.

Executive level

Show:

• total recommendation coverage across priority platforms;
• priority-market visibility;
• major factual inconsistencies;
• major competitive gains or losses;
• changes in durable source authority;
• and high-priority platform gaps.

Platform level

For each platform, show:

• company recommendation rate;
• competitor recommendation rate;
• top cited domains;
• company-owned citations;
• third-party citations;
• source depth;
• source persistence;
• and high-value missing publishers.

Cross-platform source level

Show:

• domains appearing across four or more platforms;
• all-six domains;
• platform-specific Top-10 and Top-25 sources;
• sources gaining or losing persistence;
• and source relationships where the platforms disagree most.

A single score can be retained for executive convenience if its weights are transparent and the platform drill-downs remain available.

An opaque score should not replace the evidence.

The dedicated Which AI Citation Metrics Belong on a CMO Dashboard? article develops that reporting architecture further.

Methodology

This analysis uses the July through September 2026 publication-grade cohort from the LLM Authority Index persistence-portability research program.

The strict benchmark contains:

• 283 qualifying tracked markets;
• 876,605 retained AI responses;
• 8,718,629 citation occurrences resolved to registrable domains;
• six consumer AI surfaces: ChatGPT, Gemini, Perplexity, Microsoft Copilot, Google AI Overviews, and Google AI Mode.

For platform portfolio construction, registrable domains were ranked within each market-platform cell by the number of distinct retained responses containing the domain. Repeated citations to the same domain inside one response counted once for ranking.

A strict Top-100 platform portfolio required at least 100 retained responses and at least 100 resolved domains.

This article's cross-platform source-structure calculations use the 260 markets with complete eligible Top-100 portfolios on all six platforms.

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

That produces:

• the six-platform union;
• exact one-platform sources;
• sources appearing on at least two, three, four, five, or six platforms;
• and platform-specific unique-source counts.

The Top-25 and Top-10 analyses use the first 25 and first 10 domains from the same eligible ranked platform portfolios. No lower-support platform cells were added.

All figures are descriptive summaries of observed source portfolios.

Limitations

This article converts citation-source structure into an operating framework. The evidence supports the structure, but not every possible marketing conclusion.

The study does not test GEO tactics causally

It does not establish that schema, content publication, PR outreach, sponsorship, backlinks, page format, or publisher placement caused a platform to cite a source.

Shared domains are not necessarily shared pages

Two platforms can cite the same registrable domain while citing different URLs, authors, product pages, or topics.

Unique sources are not automatically valuable targets

A domain can be unique to one platform for reasons that have little commercial value. Publisher ownership, editorial quality, relevance, audience, ad policy, and cited-page context require separate diligence.

Platform overlap does not reveal technical architecture

High overlap does not prove shared retrieval infrastructure. Low overlap does not prove technical independence.

The historical collection is not a certified identical-prompt panel

The stored data does not establish identical prompts, model versions, geographies, retrieval settings, and collection conditions across all platform-month cells.

The strategy framework does not prescribe budget weights

The citation data describes source reuse and fragmentation. It does not measure platform audience size, conversion value, or channel ROI.

Citation authority is not recommendation authority

Being used as a source is different from being recommended as a company, product, or service. Those outcomes should remain separate in both measurement and executive reporting.

Conclusion

One GEO strategy can work across ChatGPT, Gemini, Perplexity, Copilot, Google AI Overviews, and Google AI Mode if “one strategy” means one coordinated operating system.

It should not mean one identical publisher list.

The six-platform evidence is too fragmented for that.

In the median complete market, the six Top-100 portfolios produced 310 distinct source domains. About 59.67% appeared on only one platform. Only 14.90% appeared on four or more, and only 4.41% appeared on all six.

The fragmentation persisted at the top of the rankings. Half of the median Top-25 union and roughly half of the Top-10 union consisted of single-platform sources.

At the same time, there is enough shared structure to justify coordination.

The median market had 46 sources appearing on at least four platforms. Google AI Overviews and AI Mode formed a consistently reusable pair. Markets such as Mortgage, eSignature Software, Business Checking Accounts, and Credit Monitoring had larger-than-average cross-platform cores.

The correct model is therefore not one strategy or six.

It is:

one coordinated GEO program with a shared factual foundation, a measured cross-platform source core, and platform-specific extensions driven by the observed market evidence.

That model preserves efficiency without hiding the differences that matter.

Related LLM Authority Index Research

• Six AI Engines, Six Different Webs provides the full 15-pair citation-overlap matrix.
• Which AI Platforms Share the Most Citation Sources, and Which Diverge the Most? focuses on the pair rankings and distribution tails.
• Google AI Overviews and AI Mode Are Siblings, Not Twins examines the Google search-surface relationship.
• How Publishers Should Think About Platform-Specific AI Citation Demand applies the overlap evidence to publisher strategy.
• Microsoft Copilot Lives on a Different Web focuses on Copilot source divergence.
• Persistence vs Portability: Why a Source That Wins on ChatGPT May Not Transfer to Gemini or Google AI compares temporal durability with platform transfer.
• How to Find the Market-Specific Publishers AI Systems Actually Cite provides the market-level publisher discovery workflow.
• Citation Breadth vs Citation Depth: Which AI Authority Metric Should CMOs Care About? explains why cross-platform reach and local prominence must remain separate.
• Which AI Citation Metrics Belong on a CMO Dashboard? provides the executive measurement framework.
• AI Citation Authority vs AI Recommendation Authority separates source authority from company recommendation behavior.

About LLM Authority Index

LLM Authority Index measures how AI systems cite sources, surface publishers, and structure authority across commercial markets. Its research separates persistence, portability, market breadth, platform breadth, source depth, citation authority, and recommendation authority so that cross-platform measurement can remain coordinated without hiding meaningful platform differences.