Alexa+ arrives in Italia today. How the new generative voice Ai works

These LLMs do not just generate responses, but can plan and execute actions by connecting to services and devices via APIs, as in the case of TheFork. But the scope is broader. In a series of demos during the presentation event, we saw the possibility of sending documents to Alexa+ via email, and then querying it verbally while keeping the memory (How much are the condominium expenses? Can I get a reminder when the next instalment is due?). Or what time a dinner is and how we agreed in the division of labour on what to bring by email. And then get the recipe, or place a purchase order by dictating to Amazon.

Alexa+ allows for broader interaction in conversational language to turn on lights, start the little robot that mops the floors, make coffee. Obviously with compatible devices. The same goes for music requests to Amazon Music, or starting a film with Prime Video with pointed questions: who wrote the soundtrack?

Alexa+ enables conversations in a more natural language than traditional Alexa. And it changes the role of the activation word 'Alexa' to work. It only needs to be said on the first activation, then the dialogue continues naturally without having to repeat 'Alexa' for every question or activation. Unless there is a long pause, or the action is fully completed. Or another person intervenes, but again it depends on the context.

Profile management becomes important, which the user must do either from the app or from the monitor in Echo models that have one. This is because Alexa+ then keeps track of individual member preferences (songs, allergies, things like that). The timbre of the voice has changed and is more adaptable to the tone of the conversation.

Amazon's idea is to decouple Alexa+ from a single product. Rather it follows the user through their day. At the device level it is integrated in Echo speakers, or the Fire TV but also the Alexa app on the smartphone. Michele Butti, vp Alexa International, states that: "In 76 per cent of cases interaction with Alexa + generates actions that others cannot do, I'm thinking of home automation, making playlists and playing them, reserving a table. And in a very few weeks Alexa+ will also be available via browser'.

This, then, will add text input to voice input, in an experience that Amazon wants to be as seamless as possible. According to Reuters, Amazon is also thinking of a smartphone dedicated to Alexa+, which would go beyond the app paradigm and aim for a total Ai experience, even as an interface. But to date this is only a rumour.