Alexa will soon have a better idea of what you’re asking and won’t need to ask you as many clarifying questions to provide a response. That’s because Amazon is making a concerted push to expand the conversational and response skills available for its Alexa voice assistant, according to Venture Beat.
Alexa originally could manage 15 command categories called “intents,” which are used with voice inquiries. Common intents include stop, cancel, help, yes, and no. Alexa skills programming also includes “slots,” which are similar to data types. Examples of slots include date, number, city, and time. Depending on the type of question or request, Amazon also enriches intents, so if a user asks about the weather, for example, Alexa knows about various weather factors and attributes, as well as the most common ways users ask about the weather.
The Alexa Skills Marketplace, where users can locate, download, and activate skills, originally had about 100 skills. That number has now grown to more than 5,000. Skills are similar to mobile apps, but they are built specifically for voice interchange with Alexa, use Alexa’s command set and vocabulary, and relate to an outcome from or controlled by Alexa. The outcome could be telling a joke, setting an alarm, dimming Alexa-compatible smart lights, adding an appointment to your calendar, giving a local weather report, and more.
During the next few weeks, Amazon plans to add hundreds of new commands, slots, and intents to the Alexa Skills Kit, initially for use in the U.S. only. The Alexa Skills Kit contains tools developers use to create skills. By including new slots and intents, Amazon will simplify skill building and save time for developers.
In the interest of improving how Alexa understands and responds to user requests, Amazon has been tracking how questions are asked in order to improve the voice assistant’s natural language understanding. The goal is to make it easier for users to obtain an appropriate response or action from Alexa without needing to use what feels to them like “unnatural language” when shaping requests.
Making it easier to build Alexa skills while at the same time increasing Alexa’s ability to respond to users’ usual language when they make requests or inquiries will accelerate both the number of available skills and people who want to utilize them.
