Your ServiceNow knowledge base isn't stale. It might be wrong.
Age and inaccuracy aren't the same problem. An article from three years ago can still be correct - and one published last month can already be wrong.
Why KB accuracy actually breaks
None of these are about neglect. They're about a platform and its processes changing faster than anyone circles back to check what was written about them.
Platform upgrades change the steps, not the article
A UI refresh or feature change moves a button or a menu path - the article describing the old one is never updated to match.
A workaround becomes permanent
An article documents a temporary fix for a known issue - the real fix ships, the article documenting the workaround never gets removed.
Screenshots and menu paths drift with every UI refresh
A visual guide is only accurate until the interface it shows changes - which, on an actively maintained platform, is often.
An article gets copied for a related process
Lightly edited from the original for a similar-but-different scenario - inheriting whatever was already wrong with it.
How to check
The useful question isn't "how old is this article" - it's "does this still match what the platform actually does today."
Compare against today's platform, not against when it was written
An article's accuracy is relative to the current state of the system it describes, not to its own publish date. That comparison has to happen against the live platform, not just a read-through of the text.
Age is a hint, not a verdict
A recently updated timestamp doesn't guarantee correctness, and an old one doesn't guarantee it's wrong - a genuinely stable process can be documented correctly for years. Treat age as a reason to check, not as the check itself.
Patterns worth reviewing first
- Articles referencing UI elements or steps that moved in a recent upgradeThe most direct signal of drift - the platform changed underneath the instructions.
- Articles with no update since the process they describe last changedA real gap between "the process changed" and "the article changed."
- Articles copied from another one and never independently verifiedWhatever was wrong in the source likely carried over.
- High-view articlesAccuracy matters more where more people are actually relying on it.
The manual version doesn't scale
Checking this properly means a person reading each article against the platform's current real behavior - not just glancing at a last-updated date. On a knowledge base with hundreds of articles and a platform that keeps changing underneath them, nobody re-verifies that continuously by hand.
Running this review continuously, automatically
Our own platform, neo.ai, reads knowledge content against what the platform actually does today - judging inaccuracy directly, not just flagging articles by age. It's one of seven categories it watches across a ServiceNow instance.
SwissNow