We are in the process of trying to move our Web Chat integration over to Web Messaging. We have run into an issue when trying to set up our Dialogflow Bot on the new flows.
The issue is that in the Web Chat Architect flows, the dialogflow bot is configured with the field "agent input text" left blank. This allows us to send a response to the customer in architect and then the dialogflow bot takes the input from the next message from the customer and uses that to identify the intent. It works perfectly.
On Web Messaging however, this "agent input text" field cannot be left blank. It defaults to message.message.body. which pulls in what ever the initial opening message was from the customer when they started the chat. This means I can't grab the intent from the customer at the stage I want (we call the dialogflow bot multiple times in the flow) and instead I am defaulted to always looking at the opening salvo.
I can switch the field from literal, to expression, then back to literal which allows me to publish with the field empty. However after a few minutes the system defaults the field and re-enters message.message.body in the literal field passing over the string "message.message.body" to dialogflow.
Is this difference in set up built by design? Does anyone know of a way to disable this requirement for web message to have the configuration the same as it is with chat? Or failing that, is there a work around which I am missing which will give us the same experience we have on Chat?
