Voice recognition technology still has a long way to go before it works. I don't like automated answering systems, but I much prefer working with Touch-Tone signals than voice commands which seem to be often misunderstood and less precise.
The English language is a horrible thing to automate. Remember English developed from TWO distinct sources and is a blend of syntax, sounds, etc. How do you deal with "their there they're"?
Years ago the hospital I worked at had automated voice dictation recording equipment via the telephone system. One dialed into the system, then dialed various codes to start/stop/playback the tape. The tapes were sent out for transcription.
Note this was controlled by rotary phones. I believe dictation systems was one of the few things the Bell System allowed to be interconnected with their network. The PBX seized the incoming extension line in order to get the dial pulses, and passed them onto to the dictation system. I presume supervision was passed too. (Music for the PA system was also provided by independent playback though the PA system itself was Bell and part of the PBX.)