AIUNIFY Chat supports two user-facing ways to bring multiple website pages into an Agent:
These should not be confused with one another.
On the Agent's Sources page:
The interface warns users to choose Sitemap only when pasting a real sitemap.xml.
The deployed validation expects a sitemap-like URL, normally one ending in .xml or containing “sitemap” in its path.
For one ordinary webpage, select Single URL instead.
A Sitemap Source can represent multiple indexed pages. The Sources list can display the number of pages indexed compared with the total known page count.
Use Refresh while background crawling is still in progress.
During initial onboarding, AIUNIFY Chat can automatically discover candidate pages from a website.
The discovery process can use sitemap information and common website paths to build a list of candidate URLs.
The onboarding interface displays Found [count] pages and tells the user:
Pick the ones you want the agent to learn from. You can add more later.
Each discovered URL has a checkbox so you can control which pages become Knowledge Sources.
Each selected URL becomes its own URL Source and is queued for crawling.
In the deployed customer interface inspected for this manual, the discovered-page checkbox workflow is exposed during onboarding rather than as a general bulk-discovery button on the normal Sources page.
After onboarding, add more content through the normal Sources page using Single URL, Sitemap, pasted text, or supported file/integration workflows.
The onboarding interface can report that no pages were found from the sitemap or common paths. In that situation, continue and add URLs manually from the Agent's Sources page.
Do not automatically select every discovered page. Prioritize pages containing accurate, current, visitor-relevant information so the Agent's Knowledge remains focused.