We recently attended the Alexa Live event intrigued to hear the ambitions for development and how Amazon is supporting developers in the evolution of skills.  A point that came across clearly is the long-term vision for Alexa is to create an ambient assistant that is proactive, personal & predictable. This ambient computing should be Ubiquitous, the computing you need is always there, always available, at home, in your ears, in your car or on your phone without being intrusive.  It should get smarter – be natural, contextual, proactive, adaptive to understand customers better over time.  It’s about simplifying customers lives.

This is no longer just about smart speakers but also cars, wearables, PCs, smart home devices.  The Alexa Connect Kit is due to launch in Europe this year, this would allow devices such as microwaves, fans etc. to make their products compatible with Alexa.  Currently there are 140,000 + Alexa compatible smart home devices but Alexa devices are now Matter certified meaning that smart home devices should sync up easily and securely.

Although voice is part of the foundation for a truly accessible ambient experience there is a big focus on multi modal – allowing consumers to interact in the way that is most intuitive, whether that’s voice, touch or even motion.

New Alexa devices now have touch & motion input & responses as well as voice – Alexa is now truly multi-modal.

Echo Show 10 can use motion, connect with people using auto framing in motion when connecting on calls.  Games Developers like Sony and Volley have added motion in their skills to add personality and effects. This has seen very positive results in terms of consumers, Echo show is the fastest growing device and multi-modal skills see more than 3 x increase of monthly active users

Last year 'Quick Links' for Alexa was launched.  This allows people to add links to Alexa skills from mobile apps and websites or ads on social media.  The best integrations were when the quick link brought consumers to a specific task in the skill.  For example, Dominos saw a tenfold increase in monthly users of its Alexa skill by quick linking to order trackers within the Alexa Skill.

On the flipside 'Send to Phone' allows customers to start an activity on Alexa enabled devices and continue on a mobile device. This helps build ambient experiences that make use of both Alexa and mobile devices.

McDonalds & Vixen Labs In the UK used this technology in the McDonalds Alexa Skill they developed to enhance customer experience.  The skill allowed customers to find their nearest store, opening hours etc but also informed them of the deal of the week.  This was available via the app and users could ask their Alexa to open the deal in their mobile app, they would then be sent a notification and it was ready to use when they ordered via the app.

Amazon appear to have put a lot of work into helping develop commonly used frameworks that developers can customise.  This means skill development can be quicker, more cost effective and based on already tested AI routes.  One example of this which ties in with the idea of Ambient computing is vehicle experiences.  Earlier this year the Alexa Connected Vehicle skills API was launched.  This provides pre-built components for connected vehicle use cases E.G. Nissan customers use voice to start & stop engine & lock vehicle.

Another Vertical Solutions for skill development is due to launch later this year, a shared activities API for gaming. Customers interacted with Alexa Games hundreds of millions of times last year, the Jeoopardy! skill has 15m users.  This API will allow people to play other households and Sony are hoping to incorporate it into their Jeopardy! Skill in time for Christmas.

Another example of assisting developers is Alexa Skill Components, Reusable & composable experiences that you can import into your skill & customise as needed, shortening the time it takes to develop a skill.  Initially they are launching the following components:

• Catalogue search & refinement
• Song Requests
• Order tracking, scheduling
• Buying items from Amazon

An example of this which has just launched is with iHeartRadio in the US where the skill can be used to request songs and dedications. Billions of hours of radio, podcast and music is accessed via smart speakers but users want to have deeper connections with creators and this component gives the opportunity to turn Passive listening into active engagement.  Future planned developments include AQ&A with show host, voice driven polls & contests.  They have also launched the Spotlight feature on Amazon Music – which allows artists to directly connect and share content with consumers. 

The other key area of development is to make Alexa more proactive, although clearly this needs to be balanced with privacy concerns so these elements operate on an  ‘opt-in’ basis.  Currently you can set up a Voice profiles so in a household with multiple users Alexa recognises your voice and personalises the experience to suit your preferences. This is being further developed so voice consent can be used to share this  information with skills reducing the friction with skill set up.  Another development to watch out for later this year are event Based triggers – this would notify a skill when specific events such as beginning the morning commute start.  These triggers are used to launch a proactive suggestion, for example if someone begins a morning run Alexa might suggest a playlist.  Already 1 in 4 smart home interactions are initiated by Alexa using predictive & proactive features such as Alexa Guard, hunches and routines.

Another area of innovation is around discovery of skills and how to acquire and engage with consumers.   With screens now becoming more common Featured Skill Cards in the home screen rotation will become available to promote skills. The NFI (Name Free Interaction) toolkit can help your skill to be served using AI without having to repeat verbatim the invocation.  It can also recommend skills that have been interacted with previously.

The final area addressed was the potential to monetise skills.  These ranged from becoming an Amazon Associate with affiliate links, allowing in-skill purchase if your products are available via Amazon and the launch of Paid skills which would require a one off payment for those skills which do not lend themselves to the currently available  in-app purchase or subscription models.