Content freshness: how decay hurts AI visibility

When an engine answers a question, it's making an implicit promise that the answer is true now. For anything time-sensitive - pricing, features, statistics, best practices, anything dated - an old source is a risk. Engines hedge that risk by preferring content that signals it is current and by favoring newer sources when they exist.

Freshness is relative, not absolute. A page doesn't decay in a vacuum; it decays as the world moves on around it. The moment a competitor publishes a more accurate, more recent answer to the same question, your older page becomes the weaker citation even if it hasn't literally changed.

Decay is usually a slow accumulation of small staleness rather than one dramatic event.

Show, accurately, when a page was last meaningfully updated. A visible and structured 'last updated' date helps both readers and engines gauge currency - but only if it's truthful. Bumping the date without changing the content is a hollow signal that erodes trust when the stale facts inside contradict it.

Reflect real maintenance in the content itself: refresh the statistics, update the examples, fix dead links, and revise any guidance that's changed. Currency in the body is what justifies the freshness signal in the metadata.

You can't update everything constantly, so triage. Identify your highest-value pages - the ones that earn citations, traffic, or pipeline - and the ones covering the fastest-moving topics, and review those on a regular cadence. Lower-stakes evergreen content can be reviewed far less often.

Treat a refresh as a real edit, not a date change: re-verify the facts, add anything new the topic now demands, and prune anything that's become wrong. A well-maintained page can hold its citations for years; a neglected one quietly bleeds them.