question

Matt Kruse avatar image
Matt Kruse asked

My "capture everything" utterance fails now (EchoTTT) - What's the fix?

In my EchoTTT Skill (which is still awaiting certification) I use a single sample utterance like this: requestIntent {do something|request} The goal is to capture everything, and always map to a single intent. This was working before, but a user reported that it no longer works. I tested tonight, and it does indeed fail for any complex sentence spoken to the Echo. If I use a simple phrase like "Alexa, ask Waldo to dance" it works fine. But anything more complex, like "Alexa, ask Waldo to turn on the TV" and it is simple not recognized and mapped to any intent. What's the "new" way for a default capture-everything intent?
alexa skills kitvoice-user interface
10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Steve A avatar image
Steve A answered
Have you tried adding more sample utterances, from one word up to seven or eight (or more)? requestIntent {fish|request} requestIntent {hot water|request} requestIntent {pass the salt|request} requestIntent {where is the bank|request} and so on. I have some skills that need to capture anything, and that seems to work pretty well. Steve
10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

MisterDan avatar image
MisterDan answered
We found that when capturing "all utterances", we had to remove ANYTHING in the form of: myintent xxxx Such as: requestIntent {do something|request} // this is fine finishIntent goodbye // screws up the "all utterances" from working above We had to remove the finishIntent and instead do: requestIntent {goodbye|request} and then handle the "goodbye" in the requestIntent code. Dan E.
10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

jjaquinta avatar image
jjaquinta answered
They are tweaking their language recognition all the time. When they added in custom slots that probably perturbed things a lot. A generic "capture everything" intent is against the goals of what Alexa is trying to accomplish. I've expressed surprise before that it works to the level you need it at all. But, because it isn't a scenario they are trying to optimize, it is one that is subject to perturbation as they optimize other scenarios that are part of their core design. So, you may have had to remove all that other stuff to get this to work before. Try adding it back in now. It's a new system you're trying to game. It's kind of back to square one.
10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Branson Blair avatar image
Branson Blair answered
> requestIntent {do something|request} Matt, I had this exact problem and feel like I'm having a deja vu moment. My word was Yoda instead of Waldo but I ended up having to setup the utterance to be TVIntent turn my TV {on|off|TV} where TV is my slot name. I found it weird to have to parse this out in the utterance. Sounds like it has been changing, in how it handles these requests. I have problems getting Alexa to do a request where the utterance comes at the same time as the ask Skill such as Alexa, ask Yoda to turn my tv on. Of course I had to code 10 different ways to say that but it does work good today but if I say Alexa load or start skill it works better every time.
10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Matt Farley avatar image
Matt Farley answered
Hi Matt, I had the same problem and found my answer here: https://forums.developer.amazon.com/forums/thread.jspa?threadID=9251
10 |5000

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.