How this would work:

1. User enters in a new host entry wizard and adds any monitoring types they want to scan for and credentials: Agent, SNMP, SSH?, port scan, services, etc.
2. Zabbix does a scan against it, looking for monitoring agent, available listeners, services running, web monitoring, etc.
3. On the next page, zabbix presents the user with a categorizes list with chechboxes of all items found and the user can check items they would like to monitor for changes
4. User sets overall alert level for the host
5. Zabbix starts monitoring all the discovered items for that system.

Why? This would streamline monitoring and make it much easier for new users to add monitoring to systems.