This article was AI-generated as part of an experimental historical-content project. The date reflects the period being analyzed rather than the date the article was originally written.
Google’s November core update finished rolling out yesterday, after a little over three weeks. That makes this a good moment to look at the whole picture, because the past two months changed more about how companies are found than most years do.
In late October, Google said AI Overviews were expanding to more than 100 countries and more languages, with over a billion monthly users. A few days later OpenAI launched ChatGPT search, giving paying users web answers with links to sources. The Internet Archive spent days offline after a cyberattack before the Wayback Machine came back in read-only mode. And in November, Google tightened the wording of its site reputation abuse policy.
Plenty of teams will do a version of this audit in January, usually after something has already gone wrong. Doing it now is cheaper. Here is what I would include.
Re-baseline the results for your names
Now that the core update is complete, capture the first page of results for the company, its main brands and each senior executive. Compare against your last snapshot from before November 11, if you have one. Look less at individual positions and more at which kinds of sources moved in or out: your own pages, independent news, Wikipedia, reviews, forums.
Check AI Overviews in your markets and languages
If you operate in any of the newly added countries, run your core queries there and in the local language, not only in English from your headquarters. An overview in Hindi or Japanese may cite different sources from the English one, and nobody on your team may have looked.
Ask ChatGPT search the same questions
Use the questions a stranger would ask: who runs the company, is it legitimate, what happened with a past controversy. Note the sources it cites. Its source choices will not always match Google’s, and the differences tell you which pages each system treats as authoritative about you. This is the same discipline as with Google’s answers: corroboration is the hard part.
Read your Wikipedia and Wikidata entries line by line
Leadership changes, acquisitions, renamed products and new headquarters from 2024 may not be reflected yet. Small factual errors matter more than they used to, because they travel into Knowledge Panels and AI answers. Where something is wrong, propose a correction on the talk page, disclose your affiliation, and bring a reliable independent source.
Look for name collisions
Search each executive’s name without the company attached. Another person with the same name, especially one with a lawsuit or a scandal, is a classic source of confusion for search engines and AI answers alike. It is far better to know about a collision before an investor or reporter finds it.
Keep your own record
The Wayback Machine outage was a reminder that “it’s archived somewhere” is not a records policy. Keep dated copies of key statements, leadership pages, and the results and AI answers you capture in this audit. If a dispute comes up next year about what was said and when, you will want your own evidence.
Review third-party content on your domains
Google’s November 19 clarification says using third-party content to exploit a site’s ranking signals violates its policy regardless of first-party involvement or oversight. If your site hosts partner sections, sponsored guides or white-label content, make sure they are there for readers rather than for rankings.
None of this requires new tools. It requires someone to spend a few focused days looking at the company the way an outsider would, across all the places an outsider now looks. That used to mean ten blue links. As of this December, it means search results, AI answers in several languages, a chat interface with citations and a Wikipedia article that all of them read.