One of the major advantages of embedded analytics, as the name suggests, is that it's embedded into business applications. This means users don't have to switch between apps to get insights or track performance. Though the underlying analytics engine was likely built using a BI solution, users hardly ever realize it because the analytics features are set up to blend with the UI of the application it's embedded into.
Embedded analytics offers context-specific dashboards depending on the page or portal in which it's embedded. For example, in the case of a CRM application, the page that displays leads would display analytics on the likelihood of converting those leads into paid customers, whereas the invoice section would show the total outstanding payments due for the current month.
The downside of such tight integration is that the analytics offered is restricted to a single application. For example, if sales reps wish to view the post-sale satisfaction rating of the accounts they handle, it wouldn't be possible because customer support is usually handled outside the CRM application.
Another challenge embedded analytics poses is not being able to provide ad-hoc customization of reports because software vendors offer predefined configurations that may not suit everyone.