Jul 22, 2017 · 1 min read
Errr… no? This is actually one of the biggest complaints that Alexa developers have right now — the way the intent model works we don’t get the fully rendered text from Amazon, which means we can’t actively refine our models to fix gaps that are causing slot/intent misses.
The only way to do this realistically is to describe your intent model as a single catch-all intent and then do language parsing on your own, at which point you lose much of the value that Alexa provides.
Kevin McLaughlin at TheInformation actually wrote an article about this just recently (paywalled, fwiw): https://www.theinformation.com/facing-new-rivals-amazon-may-open-up-alexa-data-for-developers
