Authority Constitution
The Authority Constitution is a reference for understanding how organizations are described, compared, cited and recommended in AI answers. It connects definitions, evidence standards and practical decisions. Use it to distinguish an observable pattern from an explanation of that pattern, judge the limits of a study and identify what an organization needs to investigate next. The guidance applies to defined observations; it offers no universal formula for obtaining AI recommendations.
Start with the question you actually face
Different concerns require different evidence. A company missing from an answer raises a discovery and identity question. A website missing from the source links raises a citation question. A company mentioned but passed over for a buyer’s need raises a suitability and assigned-role question. Start by retaining the answer and naming which object is missing or inaccurate.
The presence, citation and recommendation reference explains why these outcomes should be read separately. Use the decision guides to investigate mentions, website citations, recommendations or absence. Google AI Overviews and Gemini have their own investigation pathways because an observation on one interface cannot stand in for another.
A disagreement can be useful even without a headline score. It may reveal a mistaken identity, an unsupported attribute, a difference between local and international contexts or a source that describes the market in a different way. Establish the object and conditions before deciding that the organization has an authority problem.
For a missing name, use the mention pathway; for a missing owned source, use website citations. Recommendation concerns suitability for a choice. If the missing object is not yet clear, start with absence; if a competitor is preferred, examine the comparison.
The Google AI Overviews and Gemini guides apply those questions to their named experiences, keeping the observed interface attached to the conclusion.
Understand the evidence before interpreting the result
The glossary gives the terms used throughout this reference one stable meaning. The reference library explains how those meanings apply to real reading tasks: identifying an organization, interpreting its role, following sources, examining website or social evidence and comparing answers over time.
The governance pages explain what a claim requires, how original research qualifies for publication and how evidence access or a research challenge should be handled. Read an empirical finding with its question, observation period, interface, eligible population and method. A percentage without its denominator can answer a different question from the one the reader assumes.
A public explanation is useful without being a result about your organization. A conceptual example demonstrates a distinction; a dated finding needs retained observations and validated calculations. Neither citations nor a model’s account of its own reasoning reveal every influence on an answer.
Move from understanding to a defensible decision
Use cases show how research can support a market-entry choice, a representation concern, an agency brief or a newsroom investigation. They are decision patterns, not reported client successes. A diagnosis may identify a strength to protect, a factual discrepancy to clarify, an explanation to test or a material unknown that needs more evidence.
AAI Kernel’s research scope connects organizations, markets, competitors, narratives and source environments. A private diagnosis needs an agreed question, suitable evidence and permissions; the client or appointed agency decides and implements the response. A deeper authority architecture requires its own scope rather than being assumed from a visibility audit.
The Observatory’s editorial readings and AAI’s organizational investigations have distinct purposes. A relevant, dated public reading can supply market context; it does not replace private evidence or grant reuse rights. Original AAI publications belong in the research collection only after their findings and disclosure conditions are accepted.
Common questions
- Where should I start if AI describes my company incorrectly?
- If AI describes your company incorrectly, retain the exact answer, verify that it refers to the right organization and compare the disputed statement with dated business facts. Start with the entity-identity reference, then use the representation-risk use case to assess the investigation needed.
- Is the Constitution a set of guaranteed ranking tactics?
- No. Public guidance explains the pathway; a diagnosis of a named organization requires scoped evidence and permissions.
05 Where this leads
Understanding the evidence is the beginning of the decision. The prioritization guide separates a strength to preserve, a fact to clarify and an explanation still to test. When that process identifies a question your team cannot answer from available evidence, the commissioning guide helps define the investigation and the output that would make it useful.