Skill quality is the biggest factor driving success for developers building with Alexa. High-quality skills result in increased customer engagement, which in turn drives opportunities for monetization and business growth.
“Alexa sets a very high bar for all facets of the customer experience,” says Abhishek Cariappa, senior operations manager for the Alexa Skills Insights team, which evaluates existing skills and works with you – the developer – to identify areas of improvement.
From improving the accuracy of natural language understanding to building compelling multimodal experiences, the Alexa Skills Insights team provides five tips to boost skill quality, and build high-quality customer experiences.
Five Tips to improve Alexa skill quality
1. Improve natural language understanding accuracy
With the Alexa Skills Kit, you can create skills with a custom interaction model. You implement the logic for the skill, and also define the voice interface through which users interact with the skill. To define the voice interface, you map users' spoken input to the intents your cloud-based service can handle.
“The robustness of the interaction model is what ultimately determines whether the skill can cater to all possible customer requests,” says Pranima Das, Senior Manager for Alexa Skills Insights program. “Without a robust interaction model, the skill can become frustrating to use.”
The Alexa Skills Kit, which provides self-service APIs, skill components, and other tools for building on Alexa, provides critical features to help you build robust interaction models. The intent confidence dashboard also helps map utterances to various intents. For more information, check out these tips on improving NLU accuracy.
2. Develop a robust voice user interface
The voice user interface (VUI)
is speech-recognition technology that facilitates human-computer interactions. When building Alexa skills, it is important to keep key VUI best practices in mind to avoid confusion and complexity.
People intuitively follow certain guiding principles to navigate their conversations. Grice’s maxims of conversation are a good resource for understanding these rules. When mapping out the VUI in your Alexa skill, here are relevant principles in mind:
Maxims of quantity: Be informative, but do not provide unnecessary information to the customer.
Maxims of quality: Provide truthful information with evidence to support it.
Maxim of relation: Present relevant information to the customer.
Maxims of manner: Be as clear, brief, and orderly as possible. Avoid using obscure words or ambiguous language that the customer may not understand.
Learn more about designing high-quality voice experiences in the Alexa Design Guide.
3. Focus on developing quality content and offer a unique value proposition
To build a high-quality Alexa skill, you need accurate and original content that is relevant and up to date. For example, news trivia skills need to provide up-to-date answers to questions. At all times, you should look to design content that makes your skill stand out from among the competition.
“If your Alexa skill delivers content that cannot be replicated by other skills, it can become quite attractive to customers,” says Cariappa.
Implementing localized content for different regions is another way you can boost relevancy. A trivia skill for US audiences may not translate well to audiences in other countries. Incorporating cultural nuances when creating content can go a long way in delivering a powerful customer experience.
Finally, categorizing your skill accurately can help customers discover your skill easily is essential to driving skill discovery. For more information, check out this index of Alexa skill types.
4. Deliver personalized experiences
Personalization is an effective way for you to forge deeper connections with customers.
“The overall experience can be different and personal for each customer,” says Cariappa. “Leveraging the customer location is a great way of making this happen. For example, weather skills can use location data to deliver reminders to customers. To give another example, shopping skills can recommend nearby restaurants, shops, and activities in a customer’s area.”
5. Build multimodal experiences with the Alexa Presentation Language
Alexa Presentation Language (APL) is a responsive layout language that lets you build visuals to render on multimodal devices that support Alexa. You can use APL to mix speech, sound effects, graphics, images, animations, video, and more, providing another layer of customer interaction.
In recent years, customers have responded positively to the multimodal experiences enabled by Alexa. For multi-modal devices made by Alexa’s partners and device makers - devices like TVs, PCs, smart displays, and tablets - Alexa’s active customer base has grown by almost 3X since 2020. In addition, the number of partner devices with screens has grown by nearly 7X since 2019.
“APL opens up a new multimodal dimension for a skill,” says Cariappa. “You can support more experiences, such as displaying a chess board while using a voice chess skill.”
The Alexa team provides tools, dashboards, and other resources to build powerful voice experiences. Log into the Alexa developer console today to begin improving the quality of your skills.
You can look for Alexa skills insights recommendations provided on the developer console, or reach out to the team at firstname.lastname@example.org with your skill name and skill ID information for quality improvement recommendations. We will get back to all relevant inquiries.