Creating natural voice experiences is a challenge for every voice app developer.
The ability to provide a natural experience goes beyond understanding individual words or sentences. To successfully provide a natural voice experience your skill must know how to respond to a wide range of conversational phrases and unexpected terms while also having a conversational memory, to sustain long back-and-forth sessions.
Amazon Alexa has now made available the beta version of their Alexa Conversations platform, which is an AI-driven approach to dialog management that enables you to create skills that customers can interact with in a natural, unconstrained way – using the phrases they prefer, in the order they prefer – while freeing you to focus on the highest value parts of your experience.
Put more simply, it’s a deep-learning based approach that voice app developers can use to create more natural human-like voice experiences on Alexa.
Alexa Conversations is now available as a beta in the US for English with features like advanced dialogue creation, context carryover and corrections.
By using Alexa Conversations as a part of your voice development stack, you will need less back-end code for dialogue, state management and less model training data.
Unlike traditional dialog managers, with Alexa Conversations you do not have to code or even identify all of the different ways customers might interact.
Instead, you provide a set of example dialogs and how they map to specific services in your skill. Alexa Conversations then applies deep learning to extrapolate the many possible dialog paths that customers might use and uses this learning in real-time to automatically manage customer interactions. This includes keeping track of state, maintaining context, accepting customer-driven corrections, and then calling your services at the right time.
Before Alexa conversations you would need to write code to track all the possible and often complex interactions within your voice app, and this becomes cumbersome and error-prone very quickly.
Rather than go through the trouble, many voice skill builders would sometimes reduce their skills functionality or wait to discover edge cases they had not handled elegantly.
Alexa conversations now does all this for you.
As an example, the team at iRobot wanted to go beyond commands like, “Start Cleaning”. Now, with Alexa Conversations their customers can say things like, “Vacuum the living room and mop the kitchen on Tuesdays and Thursdays” and if they forget to say the date or the room the skill will ask for clarification.
This is the kind of thing that can better help customers with everyday tasks.
I know that these appear to be extreme details for voice user interface designers, but these are the kinds of things that lead the way to the natural interactions and complex tasks that customers are after.
Learn more about how to use Alexa Conversations to create natural voice experiences here: https://developer.amazon.com/en-US/alexa/alexa-skills-kit/dialog-management
Watch the video below from the Alexa Developers team to discover more about creating natural voice experiences using Alexa: