Re-indexing
Re-indexing is when a search engine or AI platform fetches a page again and replaces its stored copy with the current one.
Re-indexing is when a crawler comes back to a page, reads it again, and replaces the copy it stored with the version that's live now. Until that happens, the search engine or AI platform keeps working from the old copy.
That's why fixing a page is a two-step job. You publish the change, which you control. Then you wait for the crawlers to come back, which you don't. An answer can keep repeating your old price for weeks after you changed it.
Why some fixes show up faster
AI platforms get their information in different ways, and each one updates at its own pace.
A platform that searches the web while it answers will often read a page's current version, as long as its search index has picked up the change. One that answers from what it learned in training only learns about your change when its model is trained again, which can be months away. Most answers use a mix of both.
For example, Fernhill Coffee lowers its subscription price from £14 to £12. Within a few days, answers that search the web start quoting £12. Answers that don't search carry on saying £14, and so do the review sites that still list the old price, until they're updated and crawled again too.
How to speed it up
You can't force a crawl, but you can make one more likely:
- Ask Google to crawl the page again in Search Console, as Google explains here
- Tell Bing and other search engines that use IndexNow that the page changed
- Update the page's date and your sitemap when the content changes
- Link to the changed page from pages that crawlers visit often
Then ask the other sites that repeat the old fact to update their pages too.
In AnswerOrbit
Connect your server logs and the Bot visits tab in Website and AI traffic shows which AI bots read which pages, so you can see when a changed page was read again. In Recommendations, Before and after a date shows your numbers before and after the day you made a change.