Hi All! First of all, I'm Andrew, I'm new on this forum and new in the fantastic world of tasker. I'm still learning the basis, and maybe for this reason sometimes I have strange behaviour. Ok, let's go to the task (it is my version of WhatsApp voice messaging): The task is simple: After recognition of "send whatsApp" A voice tells me to input the contact name Autovoice is used to recognize the name. A %name variable is set to %avword() A voice asks me if the %name is correct AutoVoice is used to recognize yes or no If the answer is yes, -Launches WhatsApp -Autoinput goes to the search -Write command inserts the %name -Autoinput will then "click" on the first field of the search results -The chat opens -Voice asks for message text - Autovoice recognition is started -%message is set to save %avword() -Voice asks "is this the message? %message" -if answer is yes --write is used to input %message --autoinput "click" is used to send message -if answer is no, go back to ask message -endif Else if answer "no" go back to input the name edif This is in brief what I made. This is what I have: me: "send WhatsApp" phone: "please tell me the name" me: "Andrea" phone: "Andrea WhatsApp is that correct?" me: "no" (same for yes) phone (received word "no WhatsApp") and task ending (MODIFIED version: yes, continue, ELSE go back, lead the system to go in loop as yes is detected as yes WhatsApp) If I modify the triggering command from "send WhatsApp" to "send" and I say only one word for each recognition all works and message is sent. As soon as I say two or more words the last one is added to the rest of the commands Please help!