Alexa — Inside Out
(Author: Jieke Pan; Editor: Saira Timmerman)
“Alexa, set the alarm to 8-o-clock in the morning”, “Alexa, tell me a joke.”, “Alexa, check the traffic.”
Amazon Alexa, known simply as Alexa, is the voice assistant platform developed by Amazon and launched alongside the Amazon Echos devices in 2014. Today, over 100 million of devices from Amazon and third-party device manufacturers are connected Amazon Alexa Voice Service. Natural voice experiences that offer customers a different way of interacting with the technology they use every day.
Amazon offers the custom skill development via the Alexa Skills Store, where developers can publish their own skill to be enabled for Alexa compatible devices. Alexa Echo devices are just a hardware interface; where all the magic happens behind the scenes is through the cloud platform of Amazon Web Services.
So, if you are an Alexa skill developer, how can you continuously improve your Alexa skill?
Optimizations
Invocation naming
An invocation name is used for starting an Alexa skill. It must be very clear and meaningful, and something that a user can remember easily, as a user needs to call the skill’s name to initiate conversation. Ideally a skill name should consist of one or two words and match with a complete invocation sentence, for example: “Alexa, Open GoodNes”, where “GoodNes” can be the invocation name.
Intent
Intent is equivalent to each function that your skill will perform. Users can call different intents to perform different functions. Define clear intent for each action’s needs.
Utterances
Alexa utterances are synonyms of sentences that can be used for calling different Intents. While building an Alexa skill there is an option to write utterances manually, these utterances are sentences which user can tell to go to particular intent, or to open particular stage inside your skill. You should write as many utterances as possible. Use AMAZON.FallbackIntent and AMAZON.SearchQuery to handle the expected requests from the users and try to get as much content as possible from them.
Summary
Ready to scale, easily to be developed, changed and deployed. These are a few important components that serverless stack of AWS can bring. Typically we are referring to Lambda, APIGateway and DynamoDB. It is also very important to make sure your lambda function’s performance is matching the required response time of any voice invocation from Alexa Voice Service.
Designing for Voice User Interface is very different than designing for screen or a website. Continuous user testing and analytics help to improve your Alexa skill accuracy over time. In order to capture as much content as we can from the original speech of the user.
Author Information
Jieke has a Chinese and Portuguese background and is fluent in several languages. Passionate about Cloud Computing, Agile, Extreme Programming, TDD, Clean Code, Software Security and Team Leadership.
He is a software engineer and architect with more than 15 years of professional experiences in mobile career network provisioning, mobile payment gateway, cloud based enterprise management middleware platform, as well as healthcare digital solutions.
As a VP of Engineering, deriving the optimal quality from a team consisting of people with very different cultural backgrounds is second nature to him.
Originally published at https://www.mobiquityinc.com.
(We hope you enjoyed reading this, and don’t forget to follow us if you’d like to read more interesting stories.
If you are looking for a job opportunity in the IT industry, come to join us at Mobiquity.)
