The upgraded digital assistant powered by Amazon’s generative AI Alexa+ has deployed to over 100,000 users, CEO Andy Jassy said Thursday in the company’s revenue call.
That’s far from the 600 million Alexa devices, but the company has made some progress with its Alexa+ rollout, which was first announced in February. At the time, Amazon said Alexa+ would be caught up in the waves over the coming months.
Amazon’s new digital assistant aims to have the agent ability to speak it in a more natural style and ultimately use third-party apps on behalf of the user. Alexa+ should be able to generate the original response on the fly, similar to the audio modes of Openai’s ChatGpt and Google’s Gemini, rather than the given responses of older Alexa and Siri systems.
However, as reported by the Washington Post at its launch, the Alexa+, which is rolling out today, lacks some of the key features the demo demonstrated in February. The report notes at launch that Alexa+ could not use third-party apps like Grubhub, generate bedtime stories for kids, or brainstorm gift ideas. It is unclear when these features will become ALEXA+.
“There are a lot of features we plan to add in the next few months,” Jassy said over the phone.
In the opening comment, Jassy claimed that Alexa+ was one of the first action-oriented AI agents for consumers. However, he pointed out that the technology is still “primitive” and “inaccurate.” Currently, most multi-step AI agents have a low accuracy of 30% to 60%, says Amazon CEO. Jassy has set the target for Alexa+, the company’s web browsing agent to enhance its Nova Act, to achieve 90% accuracy in this domain.
Amazon’s Alexa+ deployment appears to be moving faster than Apple’s new LLM-driven Siri deployment. When asked about the new Siri delay in Apple’s Thursday revenue call that coincided with Amazon, CEO Tim Cook said the company needs “more time to complete the work.”
TechCrunch Events
Berkeley, California
|
June 5th
Book now
Along the way to the SuperCharge Legacy Digital Assistant with Generator AI, both Apple and Amazon reportedly encountered failures and delays. Some of the biggest hiccups are using tools using LLM and integrating with other systems. In doing so, Alexa and Siri can complete practical tasks such as setting timers and reading text, but it has proven to be more difficult than expected to implement.