For years, talking to a virtual assistant felt a lot like talking to a robot. You had to use specific phrases, learn certain commands and resign yourself to listening to responses that were as mechanical as they were predictable. “Play music”, “turn on the lights”, “what will the weather be like tomorrow?” or “Close the garden gates.” The exchange It worked, but had little conversation. Alexa+, the new generation of Amazon’s assistant, wants to change precisely that.
Powered by generative artificial intelligence models (similar to those used by ChatGPT), the new platform marks the biggest redesign of Alexa since its launch in 2014. The ambition is evident: go from being a tool that executes orders to becoming a kind of conversational assistant able to understand context, remember details and perform more complex tasks. The goal? Getting as close as possible to giving Alexa a “physical body.”
But the question is: what has really changed? Perhaps the most important improvement is also the least spectacular. Until now, voice assistants worked best when the user learned to speak their language. Alexa+ tries to reverse the relationship and learn the user’s language.
The change is noticeable: the voice does not change, the details are not modified, but the result is different and a logical exchange begins to occur. One of the most common stories with Alexa was to ask her for something (weather information, a song and a recipe) and in many cases she would (pardon the redundancy) ignore it or make a capricious interpretation. Now the system is capable of following long conversations, interpreting questions formulated in different ways and remember the context of previous exchanges without having to constantly repeat information. Alexa+ has context.
An example: we ask you about a restaurant with a grill near us, you give us the options and then, if we ask you which of them has a vegan menu, terrace or park for minors, you understand that we are referring to the previous question. It seems simple, but it represents a huge leap compared to the “Pre Plus” options.
Another notable novelty is the ability to remember personal preferences. Alexa+ can learn routines, tastes and habits to offer more personalized responses. Amazon hopes that the assistant acts less and less like a voice search engine and more like a digital secretary able to anticipate certain needs. In theory, this allows you to ask for more complex recommendations, organize activities, or manage everyday tasks without having to start from scratch each time.
The influence of ChatGPT and other conversational models is evident. On devices with a screen, such as the Echo Show, Alexa+ transforms part of the interaction into a visual conversation similar to that of current chatbots. The user can follow the exchange both by voice and by text, something unthinkable in the first generations of home assistants.
The difference is that Alexa + also maintains its connection with the smart home, calendars, music, security cameras and the rest of the Amazon ecosystem. Another welcome development is that ability to summarize information and answer open-ended questions more accurately.
And, the last one (in the absence of testing it in a more in-depth way) is personalization, both in preferences and in the tone of interaction. Amazon has even incorporated different personality styles to adapt the way the assistant responds.To access it you must request, through Amazon websiteearly access to the Plus configuration. Once we have it, it is configured very easily, simply with a voice command, asking to activate the option.
Beyond the specific functions, Alexa+ reveals something more important. For years, Voice assistants seemed to have reached a ceiling. They were useful for simple tasks, but incapable of carrying on a complex conversation.. The arrival of large language models has changed the rules of the game.
Amazon, Google, Apple and OpenAI are now competing to build a new generation of assistants capable of understanding human language in a much more flexible way. Alexa+ is Amazon’s bet in that race.
Perhaps the best way to summarize the change is this: the original Alexa answered questions. Alexa+ tries to participate in a conversation. And although he still makes mistakes, he is probably the first time that a domestic assistant begins to look more like an interlocutor than a play written in an instruction manual.
Of course, as a postscript and personal request. This innovation is very interesting and a way to establish devices with a screen (like the Echo Show with which we use Alexa +) in areas such as the kitchen. Therefore, dear Jeff (Bezos): design some cover, protection or barrier that reduces the possibilities that the vapor from this environment (which carries water, but also fats) does not affect the speakers or the screen. Because they are going to suffer it.