I have a speak block followed by a choice block containing two paths: continue and exit. I also enabled the “else” to reprompt. It works fine in Voiceflow.
However, in the Google Action test console, the Action speaks the speak block AND the reprompt. This means it doesn’t give the user the opportunity to answer continue or exit. Any tips on why this is?
seems it is not supported yet because it’s a new feature. try else “path” as current workaround.
@kun432 I tried an else “path” that speaks and then points back to the choice block but I get the same result. The Actions says, “Would you like to continue? Sorry I didn’t get that. Try again.” instead of “Would you like to continue?”
For some reason it always selects “else” in if condition
I tried my simple test and reprompt for else works now. can you try again?
BTW, I found else does not work after stream block. In Alexa, playing longer audio with stream block might be different from simple audio in Speak block. I’m not sure about this is same as Google.
@kun432 Yes, this plays after a stream block. However, no matter what I do, the path goes to “else.” For example, if I do “reprompt,” it doesn’t give the user an opportunity to answer. If I have a “path” that doesn’t go to anything, the session immediately ends. If the “path” speaks the reprompt and goes back to the block (the same as a “reprompt” but more explicit), then it does the same thing as the first case.
What are your suggestions?
@kun432 Does this mean I can’t have any speak blocks after a stream block in Google?
BTW this prototype works perfectly fine in the Voiceflow test. I’m testing in the Google Action simulator and it doesn’t work. It was working last week but now it is not. Is there a Voiceflow bug going on?
@kun432 Hi, do you know if this bug is going to get fixed soon?
Hi, sorry for check late. I checked google’s document, tested my sample with stream block, examined GCP’s log in deep and found something like a clue why choice doesn’t work after stream block. but I don’t know much in detail about Google’s mediaplayer behaviours, and also am not sure if this is a bug or Google’s right behaviour, so I reported this in detail to Voiceflow.
also, I don’t have much experiences with Google, but
Voiceflow’s prototype tesing is neat but it is not same as Alexa or Google does. I recommend you should ALWAYS test on Alexa Developer Console for Alexa or Actions Simulator for Google.
The problem is choice block after stream block, not speak block after stream block because speak block after stream block works in my test.
Comparing to Alexa, Voiceflow’s Google support seems not enough currently. If you find any bugs, you should “report a bug” via Intercom, not here. You may not get answers or fixes soon, but it will help improving Voiceflow’s google support.
As I mentioned before, Stream block, which use Audio Player in Alexa and Media Player in Google, is very different from using audio in speak block. IMO, those are for playing audio “mainly”, not for conversation, and there’s a trade-off between using longer-audio/less conversational with stream block or using short-audio/more conversational with speak block.
Anyway, I will post here if there are any further info although I cannot promise to solve this soon, sorry…