The simplest way to manage the answers provided by the Guide AI is to define the content sources on which it’s trained.
Navu builds an index of all the content on your site. The index includes all publicly accessible content on any of the subdomains listed in the Domains page in the portal (typically any subdomain of the main site domain). This includes web pages, PDFs and video transcripts. Navu discovers this content in a variety of ways (wp-json, sitemaps, tracking, and optionally crawling). Pages tagged “noindex” and pages not included in sitemaps that are publicly accessible are still included in the Navu index.
Note that tagging content as “noindex” will NOT exclude it from the Navu index but it may inhibit Navu’s ability to find it. Therefore, Navu recommends explicitly including “noindex” tagged content in the index. To accomplish this go to Settings -> Indexing -> Content Discovery -> Additional URLs and add the URLs for the pages tagged “noindex”.
The GuideAI is trained on all the English content in the index.
If you feel that the Guide AI is not providing an acceptable answer to a specific question, the best solution is to add additional content to your website covering that question.
Pro tip: if you want to quickly train the Guide AI on new content but aren’t yet ready to publish that content to the public, publish a page with the new content, tag it “noindex” and add its URL to the additional URL’s per the instructions above. This can be a simple and effective way of adding content that you’d like the Guide AI to use but that you’re not ready to publish. Note, the Guide AI will never suggest a link or add a citation to a page marked “noindex”.
Conversely, if you don’t want the Guide AI to use certain content, you can list the URLs it should ignore by going to Settings -> Indexing -> Content Discovery -> Exclude URLs.
The Navu index typically picks up new content within a few hours of it being published and the Guide AI will be up to date shortly after that.
