ChatGPT for Amazon is ready to go! AI-infused Alexa is expected to enter the “subscription era”

The Zhitong Finance App learned that according to people familiar with Amazon's plans directly, the company is undergoing major transformation of its Alexa voice assistant service, which has been at a loss for ten years, into generative AI to fully incorporate a conversational generative artificial intelligence service with two levels of service — equivalent to the voice version of ChatGPT. Furthermore, Amazon management is considering charging about $5 per month, and there is even a possibility of charging a fee of $10. Currently, it has not been determined. After payment, users can get a higher-level version of Alexa's smart voice service. This move means that Amazon's long-term loss-making free Alexa voice assistant business is expected to enter an era of subscription payments.

The project is internally known as “Banyan”, which refers to a growing banyan tree. The project will be the first major overhaul of Amazon's classic voice assistant Alexa since it was launched with the Echo series speakers in 2014. According to people familiar with the matter, Amazon named this new voice assistant “Extraordinary Alexa.”

Citing information revealed by people familiar with the matter, the media reported that Amazon has asked employees to prepare the latest version of Alexa combined with generative AI dialogs by the August deadline, and pointed out that Amazon CEO Andy Jassy (Andy Jassy) is personally very interested in seeing Alexa revitalize. In an April letter to shareholders, Jasi promised to launch a “smarter, more capable Alexa,” but did not provide further details.

However, people familiar with the matter stressed that the company's plans for Alexa, including pricing and release dates, may be changed or cancelled depending on the progress of the “Banyan Project.”

An Amazon spokesperson said in a statement: “We have integrated generative artificial intelligence technology into different components of Alexa and are working to implement and deploy it on a large scale — there are already over 500 million Alexa-enabled environmental devices in homes around the world to provide our customers with more proactive, personalized, and trustworthy voice assistant assistance.”

This service provides voice answers to users' inquiries, such as local weather, and can also be used as a voice call center to control household appliances. This voice assistant service is one of Amazon founder Jeff Bezos (Jeff Bezos)'s most popular projects. The technology he envisioned could mimic a fictional voice computer in the “Star Trek” series.

In the field of AI chatbots, Amazon is unwilling to lag behind Google and Microsoft

For Amazon, keeping up with competitors in the field of generative artificial intelligence is critical because so-called AI chatbots launched by Google (GOOGL.US), Microsoft (MSFT.US), and OpenAI, such as ChatGPT developed by OpenAI, and Google Gemini have gained more massive user attention. These AI chatbots can respond completely to complex prompts or queries almost instantly. Microsoft, on the other hand, relied on OpenAI's majority shareholder to embed the GPT-4 AI model that OpenAI is proud of into various flagship applications such as the Office series and the Microsoft Azure cloud platform, and quickly became a global leader in AI applications, and its performance and stock price have continued to grow since 2023.

Compared to ChatGPT launched by OpenAI and Google Gemini, Amazon has yet to make significant achievements in this general-purpose AI chatbot field. Amazon previously launched Amazon Q, an AI chat assistant focused on the AWS platform for cloud services, but Amazon Q focuses on quickly calling the various functional modules of the AWS cloud service rather than a general-purpose AI chatbot similar to ChatGPT. Amazon Q is a fully integrated productivity assistant launched by Amazon AWS, designed specifically for developers of the AWS platform. Amazon Q provides access and scheduling management capabilities to the AWS platform and its connected systems through a chatbot interface. Amazon Q helps enterprise-level customers quickly get relevant answers to pressing questions, resolve issues, generate content, and quickly use data and expertise from enterprise information repositories, related code, and enterprise systems to respond to business customer questions.

ChatGPT was released at the end of 2022, and triggered a frenzy of global capital investment in artificial intelligence companies, which in turn drove the total market value of chip giant Nvidia (NVDA.US) to surpass Amazon and Google, and once this week it became the listed company with the highest market capitalization in the world. AI GPUs such as the Nvidia H100/H200/GB200 are the core hardware that drives major artificial intelligence applications such as ChatGPT and Sora. Since 2024, the computing power demand for various AI applications such as ChatGPT, Claude, and Sora, and the computing power demand for AI large-scale model iterative training terminals has continued to explode, spurring a sharp increase in demand for server AI chips such as Nvidia's AI GPUs. Amazon AWS is also one of Nvidia's most core customers.

In addition to cloud computing giants Google, Microsoft, and Amazon, consumer electronics giant Apple (AAPL.US) is also advancing its artificial intelligence strategy, including its Apple Intelligence embedded in the iPhone. Supported by this technology, the Siri voice assistant has greatly increased in intelligence, including more comprehensive and complex conversational answers.

Some Amazon employees involved in the project said that the “Banyan” project represents a “desperate attempt” to revive this voice assistant service. The free service has never been profitable, and in the past 18 months, it has been caught off guard by the rise of competitive generative artificial intelligence products. These employees said senior management told them that this is a critical year for Amazon, and voice assistant Alexa must finally prove that it can bring meaningful sales to Amazon.

Amazon's classic voice assistant, Alexa, has long been accessed mainly through Amazon TV and Echo speaker devices, and is mainly used to set timers, quickly check the weather, play songs, or answer simple questions. Amazon once hoped to increase sales for its e-commerce business through this service, but that hope fell short, mainly because users liked to see the actual products they purchased first so they could compare them.

Amazon cut thousands of jobs in its voice assistant Alexa division at the end of 2023, as part of a major restructuring after the COVID-driven e-commerce boom lost momentum, while investing more of the company's resources into generative AI.

A “must win” battle

People familiar with the matter said that Amazon hopes to embed generative artificial intelligence, and Alexa customers will ask it for shopping suggestions, such as which gloves and hats to buy for hiking trips, similar to the text-based service Amazon launched on its website Rufus earlier this year.

According to people familiar with the matter, senior management emphasized that 2024 is a “must win” year for Alexa, and this Alexa battle must be won. Similar to Amazon Prime membership, Kindle, and Fire devices, Alexa is the brand most closely related to the Amazon platform.

But the artificial intelligence version of the service, which was unveiled in September last year, has yet to be released to the wider public, and rivals have already rolled out multiple updates to their chatbots. In the demo, Alexa lost the robot's tone and answered questions such as when the soccer game started. “You can now have an almost human conversation with Alexa,” Dave Limp (Dave Limp), Amazon's hardware director at the time, promised. But he later left the company.

People familiar with the matter said that Amazon is working to replace the current free version of the “Classic Alexa Voice Assistant” with a streamlined AI-based first-tier version, while the other tier uses more powerful artificial intelligence software to handle more complex queries and prompts, and users need to pay at least $5 per month to use it. They said Amazon would also consider a price of around $10 per month. People familiar with the matter stressed that Amazon is currently not considering linking it to Amazon's $139 annual Prime membership.

People familiar with the matter said that, as conceived, the paid version can perform more complex tasks, such as writing short original emails, sending emails, and easily ordering food from Uber Eats, all of which can be done with a single prompt. They said it could also eliminate the need to repeatedly say “Alexa” when talking to software and provide more personalized AI services.

Amazon is also plagued by problems in developing artificial intelligence and other challenges, such as illusions — that is, false or misleading information generated by software, and low employee morale in the department.

People familiar with the matter said that Amazon also plans to comprehensively strengthen the home automation service provided by Alexa and embed more high-end AI service capabilities in this process. Alexa can now connect wirelessly to so-called smart home devices, so they can be voice-controlled; for example, users can turn on porch lights every night at 8 p.m., with just one short sentence.

People familiar with the matter said that Alexa, which incorporates the latest generative AI technology, can not only successfully complete the above tasks, but can also deeply learn some of users' personal habits, so it can turn on the TV, broadcast the user's favorite weekly program, or turn on the coffee pot after the user's morning alarm sounds. These are all achieved through a heavy reminder function called Alexa's “Reminder” by Amazon. However, the implementation of such services requires customers to buy more smart home devices that support Alexa.

People familiar with the matter said that Amazon has been developing more smart devices since last year and hopes to introduce this smart service into more rooms, such as home energy trackers and carbon monoxide detectors that support Alexa's voice assistant.