Alexa+ and the new intelligence
Amazon launched Alexa+, an advanced AI system that uses generative models to manage complex tasks and coordinate services, becoming increasingly autonomous.
Amazon unveiled a new and improved version of Alexa called Alexa+ on Wednesday. This new system is described as 'model agnostic', meaning it always uses the best artificial intelligence model for each specific task. During a press event in New York, Amazon VP Daniel Rausch explained that Alexa+ leverages Bedrock, Amazon's cloud platform designed to allow organizations to experiment with generative AI models. Alexa+ utilizes in-house models like Amazon Nova, as well as models from Anthropic, a strategic partner of Amazon.
One of the innovations brought by Alexa+ is the introduction of a system called 'experts', which allows it to manage and perform specific tasks. This approach enables Alexa+ to execute Amazon services and third-party services more efficiently. Rausch stated that from day one, Alexa+ will be compatible with 'tens of thousands' of devices and services. Among the offered services are news, thanks to collaborations with publications like Time, Reuters, and the Associated Press, which provide updated information on current events and market movements.
Alexa+ can perform complex tasks, such as navigating websites to complete activities on behalf of the user. For instance, it can search for a professional to repair a broken oven, autonomously contacting the repair shop. Rausch explained that Alexa takes a couple of minutes to complete navigation and then sends a notification to the user to inform them that the task has been completed. Additionally, with the 'experts' system, Alexa+ can coordinate multiple services simultaneously, such as making a dinner reservation via OpenTable, booking an Uber to the restaurant, and texting a contact with the details. Developers will be able to access this functionality through a new SDK, the Alexa AI Multi-Agent SDK, which will be available soon in preview.
