Amazon is poised to release a significant upgrade to its Alexa voice assistant, incorporating generative artificial intelligence. According to sources familiar with the matter, the company has been working on the project and recently scheduled a press event for late February to showcase the new features.
An illustration of the Amazon Alexa device.
The release of this revamped Alexa marks the most substantial update to the product since its introduction over a decade ago, which helped popularize digital assistants. The new, generative AI-powered Alexa presents both a massive opportunity and a considerable risk for Amazon, which has over half a billion Alexa-enabled devices in the market. The goal of the revamp is to enhance the product so that it can engage in more natural conversations to convert some of its extensive user base into paying customers, and to generate a financial return for the currently unprofitable business.
The new AI service will be able to handle multiple sequential prompts, and company executives have stated that it could act as an “agent” on behalf of users by taking actions on their behalf. This represents a key difference from the current version, which generally handles just one request at a time. The company is scheduled to make a final decision on the “street readiness” of the new Alexa service at a meeting on February 14.
This AI-powered Alexa faces challenges familiar to generative AI systems from companies such as OpenAI and Alphabet, including the potential for generating fabricated responses (hallucinations). Amazon could also be expanding the reach of Alexa as a daily tool, integrating it into cars, televisions, thermostats, and mobile phones, where it could be used for scheduling and even shopping.
Initially, Amazon plans to roll out the updated Alexa service to a limited number of users without charging a fee, though a monthly fee of $5 to $10 is under consideration. The company will also continue to offer “Classic Alexa,” the existing free version. Reports indicate that Amazon has stopped adding new features to the classic version.
Jeff Bezos, Amazon’s founder, first envisioned Alexa, hoping that it would become a service that resembled the voice-activated computers seen on “Star Trek.” Nearly a decade ago, then-chief of devices at Amazon, Dave Limp, said of the service, “Someday in the future – that might be years or decades away – it could answer everything that you would ever ask it.”
Due to concerns regarding the quality and response speed of the AI, the move to upgrade Alexa has seen delays, according to those familiar with the matter. Internally, Amazon has called the new service “Banyan” and also “Remarkable Alexa,” although there are still questions if either will become its product name. The company executive Rohit Prasad acknowledged some of the challenges in developing the completely new service which included eliminating hallucinations, during a January Financial Times interview.
Analysts at Bank of America estimate that Amazon could generate $600 million annually if 10% of active users — estimated at about 100 million devices — paid $5 per month for the service.
The new Alexa is designed to retain customer preferences to assist with music or restaurant recommendations and allow multiple prompts in a single session, so that, for example, a user could further modify a hamburger order before it’s delivered. The product will be compatible with existing Alexa devices.
At least in part, Amazon will rely on Anthropic, a startup in which it has already invested $4 billion, for the artificial intelligence software to underpin the new AI in Alexa. In September 2023, Amazon demonstrated an early version of the updated Alexa. However, because of quality concerns, Amazon repeatedly postponed the launch date, and that rendition was never released to the public.
