First generation IVR systems used DTMF (touch-tone) input from telephone keypads and provided pre-recorded voice output. This was before speech recognition technology became practical. Later generations added speech input (2nd gen) and multi-modal capabilities including visual/video output (3rd gen).