Amazon Alexa supports longer voice skills and self-learning techniques

Amazon Alexa has released new updates for third-party developer, including a longer voice skill time and self-learning ability, which can bring more use cases for the voice assistant.
 
Amazon has increased the voice skill time limit from 90 seconds to 240 seconds. This can eliminate the need to enter the Alexa AudioPlayer service.
 
When developers have longer sound clips, such as podcasts, to play in the voice skill, they need to use AudioPlayer, which is not part of the Alexa Voice Skills. After listening to the clip, therefore, users would need to re-invoke the voice skill in order to return to the service they are using. This is not a smooth user experience.
 
With the update, users can continue to interact with the voice skill after listening to the audio. This may improve the user experience when accessing skills related to music, storytelling, narrative-style gaming or messaging.
 
Google Assistant, on the other hand, has a 120-second limit for its voice applications.

Self-leaning ability makes Alexa more intelligent

 
Amazon posted on its developer blog a few days ago stating Alexa’s new self-learning technique which allows the assistant to automatically recover from its own errors. For instance, when users mistakenly tells the assistant to play “Good for What,” it can self-correct by directly playing the song “Nice for What” from Drake.
 
According to Amazon, Alexa will learn from several contextual clues such as users’ historical activity, preferences and what Alexa skills they are using, where the Alexa device is located.
 
The self-learning technique also enables Alexa to perform name-free skill interaction. For example, by saying “Alexa, get me a car,” users may get their Uber rides without specifying a service or skill name; or by saying “Alexa, start cleaning,” they can activate Roomba vacuum to work.