Editorial trust
Editorial standards, authorship, and corrections
How Imaginal AI researches, sources, reviews, dates, corrects, and discloses AI assistance across its decision, strategy, and leadership library.
Our publication contract
Readers should be able to inspect how a guide earned its place.
Imaginal AI publishes practical material about decision-making, strategy, business, philosophy, psychology, leadership, advisory councils, and the evidence boundaries of AI mentors. This page explains who is responsible for that material, how AI-assisted editorial work is disclosed, what qualifies a page for indexing, and how readers can challenge or correct a claim.
The standard is designed to make the library inspectable. It does not promise perfect coverage or turn an educational guide into professional advice. It gives readers concrete signals to examine: a named organisational author, publication and modification dates, visible sources, a defined search purpose, a correction route, and an explicit boundary between documented evidence and contemporary interpretation.
Four commitments
The standard applies before a page enters the sitemap.
People before page count
A page must solve a distinct reader problem with substantive original material. Query permutations and unsupported mentor-topic variants remain outside the sitemap.
Evidence before authority
Guides expose the sources and product records used to support their claims. Fluency, age, or a famous name never substitutes for provenance.
Disclosure before mystique
AI-assisted research and editorial tooling is disclosed. Generated prose is not presented as independent human reporting or as a historical mentor speaking.
Corrections before freshness
Material changes update the modification date. Dates are not advanced merely to make unchanged content appear recent.
Editorial standard
How AI assistance is used and disclosed
The library uses AI-assisted research, outlining, comparison, drafting, and editorial tooling. AI may help locate questions, organise a structure, propose counterarguments, detect repetition, or improve clarity. The same technology can also produce plausible but unsupported language, flatten disagreement, or reproduce errors from incomplete context. For that reason, AI output is treated as working material rather than as evidence. A sentence earns its place through the page’s stated purpose, its relationship to visible sources, and the editorial and automated checks applied to the release.
Every guide carries a visible AI-assistance disclosure in its “Sources and method” section. Imaginal AI does not automatically publish a model response, manufacture a byline, or create hundreds of keyword variants from one article. The build rejects duplicated paragraphs, highly similar scaled pages, thin sections, missing disclosures, weak source notes, orphaned guides, and content changes that are not accompanied by an honest modification date. These controls do not replace judgment; they make the minimum publication contract reproducible.
Editorial standard
How sources are selected and represented
Guides prefer primary materials and authoritative institutions when those sources are available and relevant. A primary source may be a named historical edition, an official framework, a standards body, a regulator, an original research paper, or first-party documentation. Secondary explanation may be used to establish context, compare interpretations, or make a technical concept accessible, but it should not be allowed to silently overrule the underlying record. Each reference includes a note explaining why it appears rather than presenting a decorative list of links.
Internal links can support product-specific claims, such as how Imaginal AI isolates mentor corpora or handles source rights, but they are not passed off as independent validation. When a guide applies an older philosophical, strategic, or psychological idea to a modern situation, it should distinguish the documented source from the present-day interpretation. A citation demonstrates where a claim came from; it does not prove that the application is the only reasonable reading. Important decisions still require the user to examine context and competing evidence.
Editorial standard
Edition rights and corpus provenance
Historical age alone does not clear a text for production use. The exact edition, transcription, translation, introduction, annotations, provider terms, territory, and intended use can carry different rights. Imaginal AI therefore maintains edition-level source records rather than granting permission at the title or author level. A source can be marked cleared, preflight-passed, or blocked. Preflight evidence is enough for controlled evaluation, not for production release; production requires an explicit owner decision for the exact checksum-bound edition and launch territories.
Rights-blocked mentors may retain a transparency profile, but they remain outside the production retrieval release and outside the sitemap until the evidence changes. Private custom mentors follow a separate path: users must attest that they have permission to upload the material, and the resulting index stays scoped to that account. This page describes an operational standard, not a legal opinion. Questions about copyright, licensing, privacy, or permitted use should be taken to qualified counsel for the relevant jurisdiction.
Editorial standard
What qualifies a page for search indexing
Search indexing follows evidence and reader value, not the number of routes the application can generate. A resource guide must address a distinct search intent, contain a developed answer rather than a teaser, expose useful sources, answer focused questions, link to relevant next reading, and fit a coherent topic cluster. Topic hubs must do more than enumerate links: each one frames the problem, identifies decision questions, explains how to use the collection, and connects the discipline to other perspectives in the library.
The launch build enforces substantive word floors, unique titles and descriptions, unique visible social images, exact canonicals, static HTML, structured data, crawlable links, and a minimum internal-link graph. Private application routes, thin mentor-topic permutations, unresolved source profiles, legacy deep links, and unreviewed community pages are marked noindex and excluded from the sitemap. Passing a technical gate does not guarantee a ranking. It prevents the publishing system from manufacturing search inventory that the reader would experience as repetitive or incomplete.
Editorial standard
Publication dates, reviews, and material updates
A publication date records when the guide first entered the public editorial set. A modification date changes when the substantive content, evidence, correction, or responsible presentation changes. Spelling fixes and formatting adjustments do not justify a freshness claim. Content hashes bind the body and editorial metadata to the recorded dates, and the release check rejects a changed page whose modification date has not moved to the actual editorial date. It also rejects an unchanged page whose dates were advanced.
A scheduled audit checks every guide’s age, static output, internal relationships, and external references. The current policy requires an editorial review within 365 days, with earlier review when a source, product capability, law, standard, or material factual claim is likely to have changed. A link blocked by automated access is distinguished from a link proven broken. Review may leave a date unchanged when the page remains materially accurate; the goal is dependable maintenance, not the appearance of constant novelty.
Editorial standard
Corrections, challenges, and professional boundaries
Readers can report a suspected factual error, misleading interpretation, broken source, attribution problem, rights concern, or undisclosed conflict through the Imaginal AI support page. A useful report includes the guide URL, the sentence or claim at issue, the reason it appears wrong, and a source that supports the challenge. Material corrections should be reflected in the page, its modification date, and where appropriate its source note. A disagreement about judgment may be documented without being treated as a factual correction.
Imaginal AI’s resources and mentor responses are educational decision-support material. They do not replace legal, medical, mental-health, financial, tax, employment, safety, or other qualified professional advice. The library can help structure questions, alternatives, evidence, assumptions, and follow-up; it cannot inspect every fact, accept a professional duty, or make the decision for the reader. High-stakes or regulated matters should be escalated to a qualified person who can examine the full situation and carry the relevant responsibility.