Limit Knowledge Source crawling to a subdomain

Currently importing a URL into the Customer Agent knowledge source is pulling data from our other sales sites that are on a different subdomain than our external help desk - which is scewing the data tha the bot has to work with, resulting in confused customers.