Amazon today announced a number of new features for developers who want to write Alexa skills. In total, the team released 31 new features at its Alexa Live event. Unsurprisingly, some of these are relatively minor, but some significantly change the Alexa experience for the more than 700,000 developers who have so far developed platform skills.
"This year, given all our momentum, we really wanted to pay attention to what developers really had to take us to the next level of what's engaging [with Alexa] it really means, "said Nedim Fresko, vice president of Alexa Devices & Developer Technologies.
Maybe it's no surprise then that one of the highlights of this version is the beta launch of Alexa Conversations, which the company demonstrated for the first time at its summit on: Mars last year. The general idea here is, as the name suggests, to make it easier for users to have a natural conversation with their Alexa devices. As noted by Fresko, this is a very difficult technical challenge.
"We are observing that consumers really want to speak naturally with Alexa," said Fresko. "But using traditional techniques, implementing naturalness is very difficult. Being prepared with random turns of words, remembering the context, bringing the context, managing the excessive or insufficient supply of information - is incredibly difficult. And if you put it in a certain way and create a state diagram, you get bogged down and you have to stop. So instead of doing all that, people are content with "Okay, well, I'll run robotic commands instead." The only way to break that cycle is to make a quantum leap and the technology required for it, so experienced developers can really focus on what's important to them. "
For developers, this means they can use the service to create sample sentences, annotate them, and provide access to APIs for Alexa calls. Hence, the service extrapolates the whole path that the conversation can take and makes it work, without the developer having to specify all the possible turns that the conversation with their own skills could take. In many ways, this makes it similar to Google's Dialogflow tool, although Google Cloud's focus is a little more on corporate use cases.
"Alexa Conversations promises to be a turning point for developers and will create extraordinary new customer experiences," said Steven Arkonovich, founder of Philosophical Creations, in today's announcement. "We have updated Big Sky's ability with Alexa Conversations and now users can speak more naturally and change their minds during the conversion. Alexa's AI keeps track of it, all with very few inputs from my skill code. "
For a subset of skills - around 400 per hour, according to Fresko - the team will also allow for a new deep neural network to improve understanding of Alexa's natural language. The company says this will lead to a 15% improvement in the accuracy of the skills that will have access to this.
"The idea is to allow developers to gain an advantage in terms of precision without any action on their part by simply changing the basic technology and making our models more sophisticated, we are able to provide an increase in precision for all skills, "explained Fresko.
Another new feature that is likely to attract a lot of attention from developers is Alexa for Apps. The idea here is to allow mobile developers to bring their users from their Alexa skills to their mobile apps. For Twitter, this could mean saying something like "Alexa, ask Twitter to search for #BLM", for example, and the Twitter ability could then open the mobile app. For some searches, after all, seeing the results on a screen and in a mobile app makes much more sense than listening to them read aloud. This feature is now previewed.
Another new feature is Skill Resumption, now available as a preview for American English, which basically allows developers to have their skills in the background and thus provide the necessary updates. It is useful for a ridesharing app, for example, which can then provide users with updates on when their car will arrive. This kind of proactive notifications is something that all help platforms are starting to experiment with, although most users have probably only seen a few of them in their daily use so far.
The team is also launching two new features that should help developers discover their skills as potential users. This remains a serious problem with all voice platforms and is probably one of the reasons why most people use only a part of the skills currently available.
The first of these launches is the Quick Links beta for Alexa, now in beta for American English and American Spanish, which allows developers to create links from their mobile apps, websites or ads to a new UI that allows them to start their skills on a device. "We think this will really help people become more reachable and more recognized," said Fresko.
The second new feature of this bucket is the unnamed interaction toolkit, now previewed. Alexa already had the skills to cast third-party skills whenever the system thought that a particular skill could provide the best answer for a given question. Now, with this new system, developers can specify up to five suggested launch phrases (think "Alexa, when will the next train to Penn Station be?"). Amazon claims that some of the first users of the preview saw interactions with their skills increase by about 15% after adapting this tool, although the company is ready to point out that this will be different for each skill.
Other updates include new features for developers who want to create games and other more interactive experiences. New features here include the beta audio APL, which provides tools for mixing speech, sound effects and music at runtime; the Alexa Web API for games, to help developers use web technologies such as HTML5, WebGL and Web Audio to create games for Alexa devices with screens; and APL 1.4, which now adds editable text boxes, drag-and-drop UI controls and more to the company's markup language for developing visual skills.
