Methods and apparatus to control a voice eXtensible Markup Language (VXML) session are disclosed. An example application server comprises an interface to send a first protocol message to a media server to initiate a communication session to provide first VXML content to the media server, and a session controller to detect an event at the application server and to direct the interface to send a second protocol message to the media server to interrupt the first VXML content at the media server and to direct the media server to obtain second VXML content via the communication session.