Techniques for voice instant messaging (VIM) on a programmable telephone set include receiving configuration data that includes VIM ID data that indicates a particular group of telephone sets. A start signal is received that indicates a start of a new message directed to the particular telephone set over a particular communication channel associated with the VIM ID. In response, VIM audio data received over the communication channel is recorded. It is determined whether the particular telephone set is currently in use. If not, then the VIM audio data is played over a speaker on the particular telephone set. These techniques allow instant, persistent audio messages to multiple recipients without a computer.