Amazon’s upgraded digital assistant powered by generative AI, Alexa+, has rolled out to over 100,000 users, CEO Andy Jassy said on the company’s earnings call Thursday.

While that’s a far cry from the 600 million Alexa devices out there, the company is making some progress on the rollout of Alexa+, which was first unveiled in February. At the time, Amazon said that Alexa+ would roll out in waves over the coming months.

Amazon’s new digital assistant aims to let users talk with it in a more natural style, and eventually have agentic abilities that allow it to use third-party apps on a user’s behalf. Alexa+ should be able to generate original responses on the fly, much like the voice modes in OpenAI’s ChatGPT and Google’s Gemini, rather than the predetermined responses of the old Alexa and Siri systems.

However, as The Washington Post reported at launch, the Alexa+ that’s rolling out today lacks some of the key features the company demoed in February. The report notes that, at launch, Alexa+ did not have the ability to use third party apps like GrubHub, generate a bedtime story for children, or brainstorm a gift idea. It’s unclear when those features will make it into Alexa+.

“We have a lot more functionality that we plan to add in coming months,” Jassy said on the call.

During his opening comments, Jassy claimed that Alexa+ is one of the first action-oriented AI agents for consumers. But he noted that this technology is still rather “primitive” and “inaccurate.” Currently, most multi-step AI agents have a low accuracy rate, between 30% and 60%, the Amazon CEO said. Jassy set a goal for the company’s web-browsing agent that powers Alexa+, Nova Act, to achieve 90% accuracy in this domain.

Amazon’s rollout of Alexa+ seems to be progressing faster than Apple’s rollout of its new, LLM-powered Siri. When asked about the new Siri delays on Apple’s Thursday earnings call, which occurred simultaneously wiith Amazon’s, CEO Tim Cook said the company needed “more time to complete the work.”

Along the road to supercharge legacy digital assistants with generative AI, both Apple and Amazon have reportedly run into snags and delays. Some of the biggest hiccups are around getting LLMs to use tools and integrate with other systems. Doing so allows Alexa and Siri to complete practical tasks such as setting timers and reading texts, but implementing it has proven harder than expected.