Enterprise reference
The shared vocabulary behind knowledge operations -
every term owned, reviewed, and dated.
Term index
Last verified 2026-09-02 07:11 UTC
Knowledge library
Last verified 2026-09-01
Whitepaper
No review record
↗
Answer Surface
Any place where an answer about your company is rendered to someone who never visits your website.
↗
Approved Claim
A customer-facing statement that has been reviewed and can be reused across website, sales, support, and AI-answer surfaces.
↗
Canonical Answer
The one approved phrasing for a question that gets asked repeatedly, held in a single place and reused everywhere it is needed.
↗
Claim Owner
The person or function named as accountable for whether a specific claim is still true, as distinct from whoever originally wrote it.
↗
Claim Registry
A maintained record of every public claim a company makes, holding its approved wording, its evidence, its owner, and the date it was last checked.
↗
Claims Library
A set of pre-approved claim statements that teams reuse instead of rewriting the same assertion from scratch, with each use linked back to the approved entry and its references.
↗
Contradiction Surface
Two or more published places that assert incompatible versions of the same fact.
↗
Corpus Reconciliation
Resolving the contradictions inside a set of documents before that set is treated as a source of truth.
Needs review
↗
Generative Engine Optimization
The practice of changing web content to increase how often and how prominently a source is cited in an answer generated by an AI search system.
↗
Golden Record
In master data management, the single authoritative version of an entity, assembled by matching and merging records from several systems.
↗
Grounding
Connecting a generated answer to sources that can be checked; in the strict sense, constraining the answer to evidence that was actually retrieved so each statement can be traced back to a source.
↗
Hallucination
Output from a language model that reads as confident and well-formed but is not supported — either by the material the model was given, or, where none was given, by verifiable real-world fact.
↗
Knowledge Drift
The gap that appears when published statements, internal documentation, and actual product behavior stop matching over time.
Needs review
↗
llms.txt
A proposed Markdown file that gives agents a curated map of a site's most important content, placed at the site root or at any path it covers.
↗
MLR Review
The medical, legal and regulatory sign-off that life-sciences companies require before any external-facing claim is published.
↗
Orphan Claim
A live public claim that no named person is responsible for reviewing, either because no owner was ever assigned or because the person who wrote it changed roles or left the company.
↗
Prompt Injection
Instructions inside an AI system's input, written to change what the system does rather than to inform it — typed in directly, or hidden in content the system reads.
↗
Provenance
The recorded history of a piece of content: where it came from, what changed it, and who vouches for it.
Needs review
↗
Retrieval Boundary
A rule that controls which sources an AI system can reference when generating an answer for a specific audience or workflow.
↗
Retrieval-Augmented Generation
An architecture that searches a document set for relevant passages and gives them to a language model as the basis for its answer.
↗
Source of Truth
The canonical record a team uses to resolve uncertainty when different pages, decks, docs, or AI answers disagree.
↗
Stale-After Date
A date recorded alongside a claim after which its supporting evidence is treated as expired until someone re-checks it.
↗
Substantiation
The evidence held on file that supports a specific claim, kept in a form someone outside the company could check.
↗
Unsubstantiated Claim
A public statement that asserts something checkable and has no evidence recorded against it, whether or not the statement happens to be true.