Programmatically update slot values with dynamic catalogs
With dynamic catalogs (coming soon), you can schedule automatic updates of your catalogs and slots for either live skills or skills in development. Use dynamic updates to specify the frequency of updates, link catalog updates with either shared slots or a skill, and instantly update your live skills.
Receive multiple slot values and respond to more complex utterances
Multi-value slots (beta) let your custom skills understand more complex utterances. With it, you can receive multiple slot values for one request. MVS is suitable for any skill that requires more than one input from the customer to provide an answer.
Use ASR Evaluation to measure your skill’s speech recognition with audio files
You can now test how Alexa interprets spoken word in the context of your interaction model and address any issues by adding sample utterances or slots to your skill’s intents. Using ASR evaluation as a part NLU evaluation, your skill’s interaction model helps Alexa convert your customer’s spoken request to text which is then mapped to an intent and/or slot.
Deep neural networks (DNNs) can help improve Alexa's natural language understanding
We are beginning to roll out Deep Neural Networks to make Alexa NLU more accurate for custom skills through: (1) improved intent recognition and (2) improved slot value recognition.
Retrieve more information on slot values for a better customer experience
With Alexa Entities (preview), you can resolve common entities from a built-in catalogue, traverse our knowledge graph to get interesting facts about those entities, and obtain high-quality, licensed images that can be used on all screen-enabled Alexa devices.
App to app account linking can now be initiated from the Alexa App without the need to re-enter credentials
Customer can now account link from the Alexa mobile app in a few clicks without entering their third party account credentials, when both the Alexa app and third party app are installed on their mobile device.