We have updated the recognition system.

Now recognition works according to the duration settings set in the Question block. That is, not only the time during which the Subscriber must start speaking is taken into account, but also pauses between words and the total duration of the session.

You can read more about the parameters in the article "Recognition session duration"