One Power Virtual Agent versus many
Typically, in an organization, you will start with one agent in a specific area of your website, serving a specific purpose. You will want to gauge the success, make improvements, and expand on that. Once you see the value, you will decide how to implement similar functionality across other areas of your site.
It is very seldom that a single bot will serve an entire site. Usually, you will structure the functionality based on the various functional areas of your public website. For example, in the Support area of your site, you will have a bot serving topics concerning user support. In the Sales area, you will have another bot serving more sales-focused topics, and so on.
Segmenting and separating the bot's functionality is beneficial from both a maintenance and support perspective, but also from the point of view of managing complexity, testability, and overall UX. After all, you do not want a bot to ask you whether your issue is sales...
