Today via a blog post, Amazon have announced the preview of the new Alexa Skills Kit. Although the preview is currently only available for US skills the new kit demonstrates that we are working on a maturing system. Much of the functionality within the new skills kit could be created by the skills developer, however, the advantage of using built-in intents can be seen when trying to cater for a large number of sample utterances to cover each potential method an end-user may use to request information. It is quite easy to miss the obvious or to be overwhelmed by the myriad of possibilities. Built in intents takes this out of the developers' hands and displaces the burden.
The new skills (although limited in number) look to be quite powerful. With great power, however, comes an increased level of complexity. In this case defining the intent now looks a tad more cryptic. 1 such example of an intent in the Amazon documentation is as follows:
This intent is declaring that our end users may request details for a member of a particular band. For example, an utterance could be 'Who is the lead singer of Bolt Thrower'.
If we were to create this functionality ourselves, not only would we need to create all variants of the utterances but we would also need to create slots. This is not the case with these new intents, slots are automatically created on our behalf. In the above example the following slots would be made available:
resulting in a request such as:
"value": "bolt thrower"
I believe handling such requests may initially cause a greater overhead but the time and effort saved with no need for creating utterances and slots should be worth it.
At present the categories available are limited (the list can be seen here) but this is bound to grow over time.
Once I have had a chance to play around with these I plan on writing up a tutorial and give you guys a proper hands on with them.In the meantime let me know what you think. Will you be making use of the new skills kit?