Speech recognition in IVRS has become more of a necessity after mobile phones have arrived. POTS ( Plain Old Telephone ) System have telephone instrument with proper keypad along with head-set. Procedure of calling is suitable for pressing DTMF based interaction with IVR System!
You normally lift the headset with left hand, dial pressing keys on telephone instrument using your right hand. So, when IVR is connected and you are listening to the voices by IVR System as well as pressing keys! While in a mobile phone, the tiny instrument itself is head-set, so when you are listening to IVR instructions, it is almost impossible to press keys without removing from your ear and looking at keys! The one solution is to use ear-phone or blue-tooth powered wireless headset!
Most probably, this has given prompted IVR developers as well as CTI vendors to come out with speech recognition for IVR Input. Though, speech recognition has evolved a lot, still speech recognition can not claim to perform 100%. This happens mainly because of pronunciation, accents of language from one region to another.
There are many speech recognition engines available. Nuance, Loquendo, Microsoft, Spinvox etc. are few companies which provide speech recognition engines on various technologies.
IVR with speech recognition capabilities are supposed to more advanced and suitable for mobile user, but it is actually very difficult to predict if the IVR application and speech engine will be able to recognise speech command properly. But IVR with Speech Recognition has been on the rise and every IVR Company now a days provide speech recognition. CTI vendors too support speech recognition in their hardware.
But still many people like me, will prefer to use DTMF for giving command to IVR as it sounds more technical. 🙂 Most probably some people will take time get used to the idea of commanding some invisible machines by speaking!