Mixed Speech to Text and DTMF inputs

The Speech to Text (STT) module allows both types of inputs: Speech and, if enabled by the user, also DTMF (number entry via the dial pad).

So what happens if a customer starts saying something and then starts typing numbers on the number pad? Or if they customer starts talking while inputting numbers?

There is an easy rule:

DTMF inputs always overwrite speech input

The reason:

• Speech analysis lags behind DTMF inputs
• If we would allow speech inputs to be mixed with DTMF inputs we would not be guaranteeing the right order of the input 
• This would cause incorrect inputs

Use cases

These use cases all require that DTMF is enabled for the STT module. If no DTMF is enabled, only speech inputs will be recognized.