The most useful source for an organisation's official policy is the organisation's own approved documentation. It is the one source written by the body that owns the policy, and the only one that can carry a version, a date and a named owner. A comment, an advert or a forum post can only describe the policy at second hand.
The answer, in one line
The organisation's own official policy or approved documentation. It is the only candidate written by the body that owns the policy, with a version, a date and an accountable owner.
Definition#
Official policy: the rule an organisation has formally approved and published as its own, in a document it controls, with a version, a date and an accountable owner. Anything else about the policy is a description of it, and is only as good as its match to that document.
Authorship settles which source wins#
In the multiple-choice version of this question the right answer is always the document, and the wrong answers share a flaw: none of them is written by the owner of the policy. An anonymous comment has no author to hold to account. An unrelated advert is selling something and was never meant to state a rule. An unsourced forum post may once have been accurate and cannot be dated. What separates the options is authorship and accountability.
Three checks turn a document into a source#
A document you have merely found becomes a source once it passes three checks. First, it sits on the organisation's own site or in a document system the organisation controls, which a copy on a third-party page does not. Second, it is the current version: policies are revised, and an archived PDF from two years ago still looks official. Third, it names an owner or an approval date. A policy with neither is a draft until someone confirms it, and the owner is the person to ask when the text is unclear.
Third-party claims are checked by leaving the page#
Sometimes the document is not to hand and the claim arrives from somewhere else. The study that gave the method its name compared professional fact checkers, historians and first-year undergraduates evaluating online information, and identified the strategies the fact checkers used to find trustworthy information efficiently and effectively (Wineburg and McGrew, graded entry). Mike Caulfield later reduced that behaviour to a short sequence, SIFT (graded entry). The part that transfers here is the move: a claim about a policy is only as good as the document it came from, so go to the owner of the policy and read it there. The study predates generative AI, and I have not seen the full text, so the participant numbers are not repeated on this page.
An AI summary of a policy points you to the document#
This section is interpretation, not finding. Asking an assistant what a policy says has become a common first step, and for many people it is also the last one. The answer reads smoothly whether or not it matches the clause, and that smoothness is what makes over-reliance hard to see from the inside. Treat the answer as a pointer: it tells you which document to open. Then apply the three checks above and read the sentence that matters. This costs a minute, which is a smaller price than acting on a rule the organisation never wrote. The same logic is set out as a general habit in the source rule.
When there is no document, say so#
If no official text can be found, the accurate report is that none was found. Filling the gap with the most plausible source turns a missing document into a confident claim. Label what you have as a claim about the policy, name where it came from, and ask the owner.
Key sources
- Wineburg, S. and McGrew, S. (2019). Lateral Reading and the Nature of Expertise: Reading Less and Learning More When Evaluating Digital Information. Teachers College Record, 121(11). Abstract read at the Stanford Digital Repository, 9 November 2026. Graded entry.
- Caulfield, M. (2019). SIFT (The Four Moves). Hapgood, 19 June 2019. Graded entry.
Related SuperSkills research#
See the source rule, over-reliance on AI, the verification bottleneck and how to write an AI use policy that works.
About this research#
Written by Rahim Hirji, author of SuperSkills (Kogan Page, 2026), keynote speaker on AI and human capability, and founder of The SuperSkills Intelligence Company. Source verification is an established field and nothing here is a SuperSkills coinage.
Evidence review · SS-2026-473 · Graded against the published rubric · How I use AI here
Hirji, R. (2026). How do you verify an organisation's official policy?. The SuperSkills evidence base, SS-2026-473. https://thesuperskills.com/research/how-do-you-verify-an-organisations-official-policy. Last reviewed 9 November 2026.
An evidence review by Rahim Hirji, not peer-reviewed research. For a material claim, cite the underlying study as well; every study here carries its own permanent link.
How citations and IDs work