What 126 Million AI Prompts Tell Us About Institutional Truth
The Semrush AI Visibility Index 2026 Through the FORMEX Lens
What happens when AI can find an organization—but cannot reliably determine what is true about it?
Semrush's 2026 AI Visibility Index examines how brands are mentioned, cited and represented across ChatGPT, Gemini, Google AI Mode and Google AI Overviews.
FORMEX asks a deeper institutional question: Does what AI understands and represents sufficiently correspond with what is actually true?
FORMEX™ Evidence Brief No. 04 examines the Semrush AI Visibility Index 2026 through the lens of institutional truth, authority and correspondence. Semrush analyzed more than 126 million real U.S. AI search prompts across 22 industries and four AI platforms. The research shows that brands may be mentioned without being cited, that different AI platforms rely on different sources, and that third-party content increasingly shapes how organizations are represented. FORMEX interprets this as an institutional underwriting issue: visibility asks whether AI sees you, authority asks whether AI relies on evidence about you, and correspondence asks whether the resulting representation sufficiently reflects what is actually true.
1. AI Has Become Part of the Decision Environment
AI systems increasingly stand between institutions and the people trying to understand them.
They search, retrieve, compare, summarize, compress and increasingly recommend.
The study covers ChatGPT, Gemini, Google AI Mode and Google AI Overviews across 22 industries.
An organization is no longer the sole author of its own representation.
2. Visibility Is Not Authority
One of Semrush's most important findings is the distinction between a brand mention and a citation.
VISIBILITY
Does AI know enough about you to mention you?
AUTHORITY
Does AI rely on your evidence—or evidence about you—to support its answer?
Being seen is not the same as being relied upon.
3. FORMEX Adds a Third Condition: Correspondence
Semrush measures visibility and citations.
FORMEX introduces a third question:
Does what AI understands, cites and represents correspond with what is actually true?
Can AI find and mention you?
Can AI find evidence it regards as credible?
Does the resulting representation sufficiently reflect what is actually true?
4. Authority Is Increasingly Distributed
Semrush found that AI systems draw from combinations of owned websites, third-party publishers, customer reviews, community discussions, retailers and reference platforms.
FORMEX describes this condition as Distributed Authority™.
What body of evidence is teaching humans and machines who we are?
5. There Is No Single AI Representation
Semrush reports results separately across the four AI platforms because their behavior differs enough that aggregation would hide important differences.
One AI system may understand an organization well.
Another may know little about it.
Another may depend primarily on third-party descriptions.
Institutional identity must increasingly survive translation across multiple machine environments.
6. AI Compresses Institutional Identity
AI frequently reduces complex institutions into short answers, summaries and recommendations.
That creates a new institutional test:
Can the institution withstand compression without losing what is essential or becoming something it is not?
The greater the ambiguity in the underlying evidence, the more an AI system must infer.
How much does AI have to guess rather than verify?
7. The Correspondence Gap™
FORMEX describes the distance between what is true and what is understood to be true as the Correspondence Gap™.
In the AI era, that gap can emerge across an increasingly long chain:
↓
What Can Be Evidenced
↓
What Can Be Found
↓
What AI Can Access
↓
What AI Interprets
↓
What AI Represents
↓
What the Stakeholder Understands
↓
What the Stakeholder Decides
Every transition creates an opportunity for correspondence—or distance.
8. The Economics of Correspondence
Representation matters because representation precedes decisions.
A Correspondence Gap™ may appear economically as:
What is the economic consequence of the distance between what is true and what AI-mediated stakeholders understand to be true?
9. Visibility Can Amplify the Gap
The natural response to weak AI visibility is often to publish more, optimize more, increase PR or improve AI-search performance.
Those actions may be appropriate.
FORMEX asks a question that logically comes first:
What exactly are we preparing to amplify?
If identity, claims, evidence and authority are coherent, amplification may increase value.
If they are not, increased visibility may simply make the gap more visible.
10. Amplification Risk™
FORMEX describes the risk of increasing exposure before the underlying institutional condition is sufficiently clear as Amplification Risk™.
Before increasing visibility, underwrite what greater visibility will amplify.
11. Measurement and Underwriting Are Different
Semrush can increasingly measure where an organization appears, how often it appears, which sources AI cites and how performance differs between platforms.
Those measurements are valuable.
FORMEX asks what sits beneath them.
MEASUREMENT
What signal is appearing?
UNDERWRITING
What institutional condition is producing the signal?
Measurement identifies the signal. Underwriting examines the condition producing it.
12. The FORMEX AI-Representation Test
Can AI find the institution?
Can AI determine what the institution actually is?
Can material claims be independently substantiated?
Which sources does AI rely upon?
Do different platforms describe substantially the same institution?
What external sources are shaping the institutional narrative?
How much must AI guess rather than verify?
Does the resulting representation sufficiently reflect what is true?
What would greater visibility amplify?
What decisions may be affected by any gap?
13. What the Evidence Does—and Does Not—Establish
The Semrush research does not validate FORMEX™.
Nor does it establish that every difference in AI representation produces measurable economic harm.
What it does establish is that AI platforms:
Those conditions make institutional correspondence increasingly consequential.
AI did not create institutional ambiguity.
It did not create unsupported claims, fragmented authority or the Correspondence Gap™.
AI makes these conditions increasingly visible, measurable and consequential.
What was once difficult to see can increasingly be observed.
What can be observed can increasingly be assessed.
What can be assessed can begin to be quantified.
That is where correspondence becomes an institutional and economic asset.
If the answer is uncertain, there may be a Correspondence Gap™ worth examining before greater exposure is pursued.
Understanding AI Visibility and Institutional Truth
What is AI visibility?
AI visibility describes how often and how prominently an organization or brand appears within AI-generated answers and discovery environments.
What is the difference between visibility and authority?
Visibility asks whether AI mentions an organization. Authority asks whether AI relies upon credible sources and evidence connected to that organization when forming its answer.
What is the Correspondence Gap™?
The Correspondence Gap™ is the FORMEX term for the distance between what is true about an institution and what stakeholders—including AI-mediated stakeholders—understand to be true.
What is Amplification Risk™?
Amplification Risk™ is the risk that increasing visibility, marketing, content, PR or AI-search exposure will amplify an unresolved institutional gap rather than increase value.
Independent evidence informing this Brief: Semrush, AI Visibility Index 2026, published June 2026. The study analyzed more than 126 million U.S. AI search prompts across 22 industries and four major AI platforms.
Study: Semrush — AI Visibility Index 2026 Study
Methodology: Semrush — AI Visibility Index Methodology
Announcement: Semrush — Expanded 2026 AI Visibility Index
FORMEX™ Evidence Brief No. 04 | August 2026