Research Methodology

AI Marketing Consensus Index studies how leading artificial intelligence platforms recommend agencies, software platforms, research providers, and other companies for specific AI marketing needs. Our methodology is designed to answer a relatively simple question:

When multiple leading AI systems independently receive the same marketing problem, which companies do they recommend most consistently?

The answer is not determined by:

• Our preferred vendors
• Affiliate relationships
• Advertising
• Sponsorship
• The companies that provide research infrastructure
• The companies that provide subject-matter support

It comes from the recorded AI responses. Our methodology separates: AI recommendation data from: factual verification from: editorial interpretation from: commercial relationships. That separation is especially important because AI Marketing Consensus Index has disclosed relationships with LLM Authority Index and CiteWorks Studio, both of which operate in markets that may be researched by this publication.

Methodology at a Glance

Each standard consensus study follows this process:

1. Define a specific AI marketing situation
2. Create a neutral standardized research prompt
3. Submit substantially the same prompt independently across multiple AI platforms
4. Preserve the raw responses
5. Identify meaningful company recommendations
6. Normalize company and product names
7. Calculate recommendation coverage
8. Measure recommendation position
9. Analyze reasons, limitations, mentions, and citations
10. Verify material company, service, and platform facts
11. Produce situation-specific rankings
12. Add editorial and subject-matter context
13. Preserve the research as a dated snapshot

1. We Start With a Marketing Problem

Every study begins with a specific use case. We do not begin by selecting the companies we want to rank. For example:

A mid-market B2B company wants to improve how frequently it is recommended across ChatGPT, Gemini, Perplexity, and other AI systems. It wants an agency that can measure current visibility, analyze competitors, identify citation and authority gaps, and implement a strategy to improve recommendation frequency. Which companies would you recommend, and why?

That question is materially different from:

What are the best AI marketing companies?

Specific scenarios allow recommendation patterns to reflect actual buyer needs.

2. We Use Standardized Research Prompts

The same core prompt should be submitted independently to each AI platform included in a study. Minor technical differences may be necessary because AI systems have different:

• Interfaces
• Input requirements
• Search capabilities
• Context limits
• Tooling
• Response formats

But the substantive research question should remain consistent. We do not intentionally rewrite the prompt for one platform in a way that favors a particular company.

3. We Do Not Seed Preferred Companies Into Open Recommendation Studies

For ordinary consensus studies, the prompt should not include a list of companies unless the research question specifically requires a named-company comparison. For example, we prefer:

Which AI visibility platforms would you recommend for an enterprise marketing team that needs prompt tracking, citation analysis, competitor benchmarking, and historical reporting?

rather than:

Which is better: LLM Authority Index, Profound, Semrush, Scrunch, or Otterly?

The first allows the AI system to reveal which companies it naturally surfaces. The second constrains the candidate set. Both can be legitimate research designs, but they answer different questions.

4. Related Companies Are Not Automatically Included

AI Marketing Consensus Index has disclosed relationships with:

• LLM Authority Index
• CiteWorks Studio

These companies should not be inserted into ordinary open-ended research prompts merely because they are related to the project. If an AI platform recommends one of them naturally, that recommendation can be recorded. If it does not, no recommendation is added.

5. Our Standard AI Platform Universe

Our standard research framework is designed around multiple leading AI platforms, including:

• ChatGPT
• Gemini
• Claude
• Perplexity
• Grok
• DeepSeek
• Kimi

The actual set used in a particular study should be disclosed with that study. Platforms may be added, removed, replaced, or temporarily unavailable over time.

6. One Platform = One Vote

For standard recommendation coverage: one AI platform receives one vote. We do not weight a platform more heavily because it:

• Has more users
• Has greater market share
• Is owned by a larger company
• Generates longer answers
• Produces more citations
• Is perceived by our team as more sophisticated

This keeps the primary consensus measure understandable.

7. We Do Not Give Multiple Votes to One Provider by Default

AI companies may offer:

• Multiple models
• Search modes
• Reasoning modes
• Enterprise products
• Consumer interfaces
• API access

For standard public consensus research, different models from the same provider do not automatically receive separate platform votes. Otherwise, one AI company could have disproportionate influence simply because it offers more models.

8. Platform Is the Stable Public Unit

Exact AI model versions can change rapidly. For that reason, the public methodology generally treats: ChatGPT or: Gemini as the primary platform unit. Where practical, internal research metadata may also record:

• Model
• Mode
• Interface
• Search state
• Tool state
• Date
• Other relevant environment details

This allows greater reproducibility without making rapidly changing model names the primary public ranking unit.

9. We Use Controlled Sessions Where Practical

AI outputs can be affected by prior context and personalization. Where practical, research should be conducted using:

• Fresh sessions
• Minimal unrelated conversation history
• Consistent settings
• No intentional personalization designed to influence the answer

The objective is to reduce unnecessary variation.

10. AI Research Is Not Perfectly Reproducible

Generative AI systems are probabilistic. The same prompt submitted twice may produce:

• Different wording
• Different recommendation order
• Different companies
• Different citations

That does not automatically make either response invalid. It is part of the behavior we are measuring. For this reason, each published study represents a: dated research snapshot.

11. We Preserve Raw Responses

Where practical, the raw AI response should be stored alongside the structured recommendation data. The raw response is important because it allows us to inspect:

• Exactly what was recommended
• Recommendation order
• Conditional language
• Reasons given
• Limitations
• Citations
• Non-recommendation mentions

The published article is an interpretation of that source data. It should not replace the source data.

12. What Counts as a Recommendation?

A company counts as recommended when the AI platform meaningfully presents it as a suitable option for the specified situation. Examples: “My top recommendations are Company A, Company B, and Company C.” “Company A is especially strong for enterprise AI visibility monitoring.” “I would consider Company B if historical prompt tracking is a priority.” These may count as meaningful recommendations.

13. A Mention Is Not Automatically a Recommendation

Suppose an AI answer states: Company A is more enterprise-focused than Company B. That may be a comparison. It does not necessarily mean both companies were recommended. We attempt to distinguish:

• Positive recommendation
• Conditional recommendation
• Incidental mention
• Comparison reference
• Negative reference

14. Conditional Recommendations Can Count

AI marketing products often serve different customer profiles. For example: Company A is a strong choice for enterprise teams, while Company B may be better for small agencies. Both can be recommendations. But the conditions should be preserved. We should not convert: “good for enterprise” into: “best for everyone.”

15. Citations Are Different From Recommendations

This distinction is fundamental to AI Marketing Consensus Index. An AI response can: recommend Company A while: citing Company B's website. Those are different events.

16. Recommendation Data

Recommendation data measures questions such as:

• Which companies were recommended?
• How often?
• At what position?
• For which situations?
• Across which platforms?

17. Citation Data

Citation data measures questions such as:

• Which domains were cited?
• Which URLs were cited?
• Which publishers appear as sources?
• Which sources support AI recommendations?
• Which companies are used as evidence even when they are not recommended?

Recommendation visibility and citation visibility should not be treated as interchangeable.

18. Mentions Are Also Separate

A third measurement is: brand mention visibility. An AI system may mention a company without recommending it and without citing it. Where measured, mentions should remain distinct from:

• Recommendations
• Citations

19. Recommendation Coverage Is the Primary Ranking Signal

The main study-level ranking signal is:

Cross-Platform Recommendation Coverage

Formula: Platforms recommending company ÷ usable platform responses Example: Company A is recommended by 6 of 7 usable AI platforms. Recommendation coverage: 85.7%

20. Usable Responses Determine the Denominator

If one platform fails to produce a usable answer, that technical failure should not be treated as a vote against every company. Example: Five platforms recommend Company A. Only six platforms produced usable responses. Recommendation coverage: 5 ÷ 6 = 83.3% not: 5 ÷ 7 = 71.4%

21. A Response May Be Classified as Unusable

Examples may include:

• Technical failure
• Empty response
• Refusal
• Severe truncation
• Response unrelated to the prompt
• Output that does not meaningfully answer the recommendation question

Where practical, we preserve the reason.

22. Recommendation Position Is a Secondary Signal

If two companies have equal recommendation coverage, placement can help distinguish them. Example:

Company A

Recommended by 5 of 7Average position: 1.8

Company B

Recommended by 5 of 7Average position: 3.7 Company A generally demonstrated stronger placement.

23. Coverage Generally Comes Before Position

Suppose:

Company A

6 of 7 recommendationsAverage position: 3.0

Company B

3 of 7 recommendationsAverage position: 1.2 Company A will generally rank higher because this publication is primarily measuring: cross-platform agreement.

24. Top Recommendation Frequency May Add Context

We may also measure:

• #1 recommendation frequency
• Top-three frequency
• Best position

These can help explain recommendation strength. They do not normally override substantially stronger cross-platform coverage.

25. Qualification for Full Analysis

A company should generally receive recommendations from at least: 2 independent AI platforms before receiving the same depth of fit analysis as companies with meaningful cross-platform support. Single-platform recommendations may still remain visible in:

• Tables
• Raw data
• Historical records

26. Genuine Ties Are Allowed

If two companies perform essentially identically, we may report a tie. We do not need to manufacture false precision merely to create: #1 and: #2.

27. Company Name Normalization

AI systems may refer to the same company using:

• Full legal name
• Brand name
• Product name
• Abbreviation
• Older name

We normalize equivalent entities where appropriate. For example: A clearly identical brand spelling variation should not create two companies in the ranking.

28. We Do Not Collapse Materially Different Products

A company may offer several different products. A recommendation for: Company A's AI monitoring platform does not automatically mean the AI system recommended: Company A's agency services. Product or service context should be preserved where material.

29. Agency, Software, Research Provider, and Consultancy Are Different Entity Types

The AI marketing market includes:

• Agencies
• SaaS platforms
• Research organizations
• Consultancies
• Hybrid providers
• Content optimization software
• Monitoring platforms
• Analytics tools

We should not automatically treat them as interchangeable merely because they compete for some of the same budgets.

30. Broad “Solution” Studies Can Include Multiple Entity Types

Some research questions are intentionally broader. Example: A large company wants to measure AI recommendation visibility, benchmark competitors, identify influential citation sources, and then improve performance. It is open to software, research services, an agency, or a combined solution. In that case, multiple provider types may legitimately compete in the same study.

31. Category-Level Rankings

Individual studies may be aggregated into category rankings. Our primary research categories are:

• AI Search & GEO Agencies
• AI Visibility & LLM Monitoring Platforms
• AI Search Audits & Market Intelligence
• AI Citation & Authority Building
• AI SEO & Content Optimization Tools

32. Category Rankings Reward Breadth

A company that appears strongly across many studies demonstrates broader category-level recommendation strength. For example:

Company A

Appears in 14 of 18 AI Visibility studies.

Company B

Appears in 5 of 18. Even if Company B occasionally ranks first, Company A has demonstrated broader recommendation coverage across the category.

33. Category Metrics May Include

• Consensus Score
• Studies Appeared In
• Total Recommendations
• Recommendation Rate
• Average Position
• Best Position
• Top Recommendation Frequency

The underlying metrics should remain visible where practical.

34. Consensus Score

A Consensus Score may summarize category-level performance. It may incorporate factors such as:

• Breadth across studies
• Recommendation coverage
• Recommendation position

The exact formula should remain consistent within a methodology version. The score should not become an unexplained mystery rating.

35. Site-Wide Rankings

AI Marketing Consensus Index may also publish broad rankings across multiple categories. A site-wide ranking measures: recurring recommendation presence across our research universe. It does not necessarily mean: best AI marketing company for every use case.

36. Scenario-Specific Rankings Usually Matter More

A company may be: #1 overall while another company is substantially stronger for:

• Enterprise monitoring
• Agency use
• AI citations
• B2B GEO
• Ecommerce
• Competitive intelligence

For a buyer with a specific need, the scenario-level study is generally more useful than the site-wide leaderboard.

37. Research Data and Factual Verification Are Separate

AI platforms may recommend companies based on claims that are:

• Wrong
• Outdated
• Incomplete
• Misinterpreted

We preserve the recommendation. We then verify material supporting facts separately.

38. Verification Does Not Rewrite the AI Vote

Example: Six AI platforms recommend Company A because they believe it monitors seven AI systems. Current verification shows it now monitors four. The recommendation record remains: 6 of 7 because that is what the AI systems actually recommended. The article should separately note the current verified capability.

39. Verification Can Affect Editorial Interpretation

Although verification does not change the historical vote, it can affect our conclusion. For example: Company A received the strongest AI consensus, but current verification indicates that a capability repeatedly cited by the AI platforms is no longer offered. That context belongs in the article.

40. Facts We May Verify

Depending on the company and study, verification may include:

• Company identity
• Service type
• AI systems monitored
• Recommendation tracking
• Citation tracking
• Mention tracking
• Prompt tracking
• Historical data
• Competitor benchmarking
• Source analysis
• Content optimization
• GEO services
• AEO services
• Technical SEO
• Reporting
• API availability
• Integrations
• Target customer
• Public pricing
• Free trials
• Agency services
• Current product availability

41. We Prefer Primary Sources

For direct company and product facts, preferred sources generally include:

• Official product pages
• Official pricing pages
• Official documentation
• Official feature pages
• Official help centers
• Official company announcements
• Other primary company materials

Third-party sources can provide context but should not automatically override clear current first-party product information.

42. Marketing Claims Are Not Automatically Facts

Companies in this market frequently use phrases such as:

• #1 GEO platform
• Leading AI visibility solution
• Best AI search agency
• Most accurate LLM tracker

These remain marketing claims unless independently supported. We should distinguish: what the company says about itself from: what our research shows.

43. AI Citations Are Not Verification

If an AI system cites a webpage, that citation is part of the research record. It is not automatically factual proof. A cited page can itself contain:

• Promotional claims
• Old information
• Inaccurate information
• Third-party opinion

Verification is a separate step.

44. Cross-Platform Consensus Does Not Mean Independent Evidence

Several AI systems may rely on overlapping sources. Therefore: Seven AI systems agree does not necessarily mean: Seven independent source ecosystems independently proved the claim. Our methodology measures recommendation consensus, not epistemic independence.

45. LLM Authority Index's Role

LLM Authority Index provides AI research data and measurement infrastructure supporting AI Marketing Consensus Index. Its contributions may include:

• AI response collection
• Recommendation datasets
• Citation datasets
• Prompt-level measurements
• Historical tracking
• Competitive analysis
• Research infrastructure

This relationship is disclosed because LLM Authority Index may itself appear in relevant AI Marketing Consensus Index studies.

46. LLM Authority Index Receives No Automatic Ranking Credit

Providing research infrastructure does not create:

• Recommendation votes
• Ranking points
• Qualification
• Competitor exclusions
• Preferred placement

If LLM Authority Index is recommended by four platforms, its recorded result is four. If it is recommended by zero, its result is zero.

47. CiteWorks Studio's Role

CiteWorks Studio provides AI search strategy and subject-matter support to AI Marketing Consensus Index. Its contributions may include:

• GEO/AEO research taxonomy
• AI search strategy
• Citation and authority concepts
• Practical interpretation
• Research-question development
• Market terminology
• AI visibility strategy

CiteWorks Studio may itself appear in studies involving agencies or AI search services.

48. CiteWorks Studio Receives No Automatic Ranking Credit

Its participation does not create:

• Additional AI recommendations
• Preferred ranking
• Higher Consensus Score
• Automatic article inclusion
• Competitor suppression

Its ranking must come from the same underlying research process applied to unrelated agencies.

49. Related Companies May Rank Below Competitors

If the research produces: Competitor A — 7 of 7 CiteWorks Studio — 3 of 7 that result should remain. Likewise: Competitor B — 6 of 7 LLM Authority Index — 2 of 7 should remain visible if that is what the data shows.

50. Related Companies May Also Rank First

If the research independently produces: CiteWorks Studio — 6 of 7 or: LLM Authority Index — 7 of 7 that result can be published. The appropriate response is not to suppress valid data. It is to: disclose the relationship prominently.

51. Mark B. Huntley's Role

Mark B. Huntley, J.D. serves as: AI Search & Visibility Research Reviewer where indicated. His review may address:

• GEO/AEO terminology
• Recommendation interpretation
• Citation methodology
• Visibility metrics
• Competitive intelligence
• Research framing
• Unsupported conclusions
• Distinctions between citations, mentions, and recommendations

52. Mark's Business Relationships Are Disclosed

Mark has financial and operational interests associated with:

• CiteWorks Studio
• LLM Authority Index

Those relationships create a potential conflict of interest. They should be disclosed.

53. Mark Cannot Change Recommendation Counts

Editorial review does not allow Mark to change:

• Raw AI responses
• Recommendation count
• Recommendation coverage
• Recommendation position
• Consensus Score

because he prefers a related company. The dataset determines the research result.

54. Human Review Can Challenge the Interpretation

Reviewer input can identify issues such as:

• A platform being categorized incorrectly
• An AI claim being outdated
• Citation visibility being confused with recommendation visibility
• A tool being described as an agency
• A recommendation being overgeneralized
• A metric being interpreted too aggressively

Those issues can be corrected or qualified without rewriting the original AI vote.

55. Reviewer Attribution Must Reflect Actual Review

If a page displays: Reviewed by Mark B. Huntley, J.D. a meaningful review should actually have occurred. Reviewer labels should not be added automatically merely because Mark is affiliated with the project.

56. Commercial Relationships Are Separate From Research

AI Marketing Consensus Index may earn revenue through:

• Affiliate relationships
• Lead generation
• Advertising
• Sponsorship
• Data licensing
• Research projects
• Related businesses

These relationships do not affect the underlying recommendation data.

57. Commercial Value May Influence Topic Selection

We may choose to study: Best GEO Agencies instead of an obscure low-demand topic because GEO agencies have meaningful commercial demand. That is acceptable. Commercial value can influence: which question we research but not: which company wins.

58. Sponsorship Cannot Purchase a Research Outcome

If a company financially supports a study, that support must be disclosed where appropriate. A sponsor may potentially help define:

• Topic
• Audience
• Research scope

It cannot purchase:

• Recommendation counts
• Rank
• Consensus Score
• Competitor removal

59. Research Dates Matter

AI marketing is changing rapidly. Every study should display a meaningful: Research Date so readers can understand when the AI responses were collected.

60. Verification Dates Are Separate

We may also display: Company Details Verified: Not yet supplied because product features may change after the AI research was conducted. The verification date should not replace the original research date.

61. Publication and Modification Dates Are Separate

Recommended date fields include:

researchDate

When AI research occurred.

datePublished

When the study was first published.

companyDetailsVerifiedDate

When material company/product facts were last checked.

dateModified

When meaningful editorial changes occurred.

reviewedDate

When actual human review occurred where applicable.

62. Reruns Create New Research Snapshots

When a study is rerun, we should not overwrite the previous research. Instead: old snapshot remains and: new snapshot is created. This enables longitudinal analysis.

63. Historical Data Is Part of the Product

Over time, historical research may allow us to measure:

• Recommendation-share gains
• Recommendation-share declines
• New market entrants
• Changes in agency visibility
• Changes in software visibility
• Platform-specific differences
• Citation-source changes
• Category leadership changes

64. A Rerun Is Not a Correction

If Company A moves from: 6 of 7 to: 3 of 7 three months later, that does not mean the first study was wrong. The recommendation environment changed.

65. Corrections Are Different

A correction occurs when our own record was wrong. Examples:

• Company counted twice
• Recommendation omitted
• Position recorded incorrectly
• Entity normalization error
• Arithmetic error
• Wrong product fact

Those should be corrected.

66. Methodology Versioning

If we materially change:

• Platform set
• Recommendation classification
• Ranking formula
• Qualification threshold
• Category aggregation

we should document the change. Where practical, studies should retain a methodology version identifier.

67. Historical Studies Should Not Be Retroactively Rewritten Without Reason

If Methodology 2.0 changes how recommendation position is weighted, old Methodology 1.0 studies should not silently be rewritten as though the new methodology always existed. Versioning helps preserve analytical integrity.

68. Research Automation

Some parts of the workflow may be automated. Potential automation includes:

• Prompt execution
• Response collection
• Company extraction
• Entity normalization suggestions
• Recommendation counting
• Ranking calculation
• Citation extraction
• Article drafting
• Historical comparison

Automation does not remove the need for validation.

69. Automated Extraction Must Be Audited

Before relying on extraction at scale, manually compare test studies against raw responses. Check:

• Was every recommendation captured?
• Were incidental mentions incorrectly counted?
• Were negative references misclassified?
• Was ranking position recorded correctly?
• Were aliases normalized correctly?

70. AI May Assist With Editorial Production

AI tools may assist with:

• Summarization
• Drafting
• Organization
• Comparative analysis
• Formatting
• Identifying potential inconsistencies

But article-generation AI should not invent research data.

71. Structured Data Controls the Metrics

Fields such as:

• Recommendation count
• Recommendation coverage
• Average position
• Consensus Score
• Platform count

should come from structured research data. The article generator should explain these metrics, not calculate them independently from memory.

72. The Generator Cannot Invent Company Capabilities

If a current product feature has not been verified, the article generator should not confidently invent it. Instead it should:

• Omit the claim
• Qualify the claim
• Flag it for verification

73. The Generator Cannot Invent Reviewer Commentary

AI-generated copy must not fabricate statements such as: Mark Huntley says... unless the statement comes from actual approved reviewer input.

74. We Preserve Disagreement

Sometimes the AI systems will disagree substantially. That is useful data. A study may legitimately produce:

• Weak consensus
• Multiple near-ties
• Fragmented recommendations

We do not need to force a decisive winner.

75. Weak Consensus Should Be Identified

If the leading company is recommended by only: 3 of 7 platforms the article should not imply overwhelming market agreement. A better characterization may be: plurality leader or: weak consensus.

76. Absence Is Not Automatically a Negative Judgment

If a company is not recommended, that does not necessarily mean the AI platform considers it poor. It may reflect:

• Lower brand awareness
• Newer market entry
• Less indexed information
• Narrower positioning
• Different source availability
• Prompt interpretation

Non-appearance should be interpreted cautiously.

77. AI Recommendations Can Favor Established Brands

Large companies may have advantages because they have:

• More web mentions
• More reviews
• More publisher coverage
• More documentation
• Stronger brand recognition
• Longer histories

Our research measures recommendation behavior. It does not claim that recommendation behavior is free of these biases.

78. New Companies May Be Underrepresented

A newer platform or agency may be excellent but still appear infrequently because AI systems have limited information about it. That is a limitation of AI recommendation consensus. It should not be hidden.

79. Public Web Presence Can Influence AI Visibility

AI systems may be influenced by:

• Company websites
• Media coverage
• Comparison articles
• Reviews
• Community discussions
• Reference sources
• Search results
• Other public information

Part of what this research measures is the result of that broader information environment.

80. We Do Not Claim to Measure Objective Product Quality

AI Marketing Consensus Index does not claim that cross-platform recommendation frequency is a complete measure of:

• Software quality
• Agency execution
• Client satisfaction
• ROI
• Customer support
• Strategic sophistication

Consensus is one signal.

81. We Do Not Claim Every Company Has Been Hands-On Tested

Unless explicitly stated for a particular study, our standard methodology is not: hands-on product testing of every company. The core methodology is: multi-platform AI recommendation research plus factual verification and editorial analysis.

82. We Do Not Manufacture Firsthand Experience

Articles should not claim:

• We used the platform
• We hired the agency
• We spoke with the company
• We tested the dashboard
• We purchased the product

unless those things actually occurred.

83. Research Questions Can Evolve

As the AI marketing market changes, new study types may emerge. Examples could include:

• AI agent visibility
• Shopping AI visibility
• AI reputation monitoring
• Agentic search optimization
• New answer-engine categories

New research should remain consistent with the core methodology principles.

84. Transparency Is Part of the Methodology

Because related companies participate in the project, disclosure is not merely a legal footer issue. It is part of how the research should be evaluated. Readers should be able to understand:

• Who contributes data
• Who contributes strategy
• Who reviews research
• Which related companies might be ranked
• Which safeguards prevent those relationships from altering AI votes

85. Our Related-Business Safeguards

For ordinary open recommendation research:

• CiteWorks Studio is not automatically seeded into prompts.
• LLM Authority Index is not automatically seeded into prompts.
• Competitors are not automatically seeded either.
• Raw AI responses are preserved.
• Recommendation counts are generated from the recorded responses.
• Related companies receive no bonus weighting.
• Competitors can outrank related companies.
• Related companies can fail to appear.
• Related companies can receive critical analysis.
• Historical poor performance should not be deleted.
• Mark's review cannot change the underlying AI vote count.
• Related-company appearances receive prominent disclosure.

86. Reproducibility

Because AI systems are nondeterministic, exact reproduction cannot always be guaranteed. However, we improve reproducibility by preserving:

• Prompt
• Platform
• Research date
• Raw response
• Model/mode where practical
• Search state where practical
• Structured recommendation output

This provides a meaningful audit trail.

87. Research Data Provided by LLM Authority Index

Where applicable, the site may display: AI research data and measurement infrastructure provided by LLM Authority Index. This statement describes the infrastructure relationship. It does not imply LLM Authority Index controls the published ranking.

88. Strategy Support Provided by CiteWorks Studio

Where applicable, the site may display: AI search strategy and subject-matter support provided by CiteWorks Studio. This describes CiteWorks Studio's role in:

• Research framing
• Strategy
• Taxonomy
• Interpretation

It does not mean CiteWorks Studio determines research outcomes.

89. Reviewer Disclosure

Where Mark reviews a study, the page may display: Reviewed by Mark B. Huntley, J.D., AI Search & Visibility Research Reviewer along with a link to his reviewer profile and the related-business disclosure.

90. How to Interpret Our Research

Readers should ask:

• What exact marketing situation was studied?
• Which AI platforms were included?
• How many responses were usable?
• Which companies were recommended?
• How frequently?
• At what position?
• Were recommendations conditional?
• Were companies cited or merely recommended?
• Were material capabilities verified?
• Is a related company involved?
• What was the research date?

These questions provide more context than a simple numbered list.

91. Our Methodology Principle

Our research methodology can be summarized in one sentence:

AI Marketing Consensus Index submits standardized marketing questions across multiple leading AI platforms, preserves and structures their recommendations, ranks companies primarily by cross-platform recommendation coverage, separately verifies material facts, adds transparent editorial context, preserves historical results, and prevents related-company or commercial relationships from changing the underlying AI recommendation data.

Related Methodology Pages

How We Rank →

Platforms We Analyze →

How We Verify AI Marketing Companies, Platforms & Services →

Data & Research Limitations →

Research Definitions & Terminology →

Editorial Standards →

Editorial Independence →

Human Review Policy →

Related Business & Conflict of Interest Disclosure →

How LLM Authority Index and CiteWorks Studio Contribute →

Corrections & Updates →

Transparency Disclosure

AI Marketing Consensus Index has related business relationships that are relevant to the subject matter we research. LLM Authority Index provides AI research data and measurement infrastructure supporting AI Marketing Consensus Index. CiteWorks Studio provides AI search strategy and subject-matter support.

Mark B. Huntley, J.D. serves as AI Search & Visibility Research Reviewer where indicated and has financial and operational interests associated with LLM Authority Index and CiteWorks Studio. Because either related company may appear in our research, their relationships are disclosed. Those relationships do not alter:

• Raw AI responses
• Recommendation counts
• Recommendation coverage
• Recommendation position
• Consensus Score calculations
• Competitor inclusion
• Historical results

Read the Full Related Business & Conflict of Interest Disclosure →

Related business disclosure · Research methodology