Internal documents from Amazon between August and November 2024 obtained by Fortune shed light on the company’s difficulties in developing its AI-enhanced version of Alexa, known internally as Project Banyan.
The new iteration, initially planned for release in September 2024, then pushed forward to October, has again faced delays due to technical setbacks, the latest on November 14. These delays highlight challenges like latency and limited device compatibility, factors that could affect Amazon’s ability to compete with the latest voice-based AI tools from tech giants such as Google and OpenAI.
Persistent Latency and Compatibility Issues
Latency, or the delay between a user’s command and Alexa’s response, has emerged as a significant obstacle. This technical shortcoming impacts the fluidity of user interactions, making them less natural and more frustrating.
Testers reported an average satisfaction score of 4.57 out of 7, missing Amazon’s target of 5.5. These results underscore the work needed to bring Alexa up to par with competitors like Google’s Gemini Live, which Salesforce CEO Marc Benioff recently lauded for its fast, zero-latency responses, calling it the “future of consumer AI”.
Compounding the latency issue is the limited compatibility of the new Alexa with older hardware. According to internal memos, approximately 3.8 million current Echo users—nearly 10% of the active user base—will require hardware upgrades to access the new version. To address this, Amazon is preparing a “recycle and save” campaign aimed at encouraging device upgrades through discounts.
Initial Skill Limitations and Subscription Plans
The new Alexa is expected to launch with only 300 available Skills, a sharp drop from the 100,000 Skills supported by the current version. This limited feature set poses a risk for Amazon’s planned monetization strategy, which aims to includes a premium Alexa Plus subscription at an estimated $10 per month. The premium service is expected to introduce generative AI-powered interactions and advanced features.
This move aims to shift revenue generation from device sales to service-based income. However, internal discussions noted user frustration could arise if they feel they’re paying for a product with fewer capabilities.
To enhance the assistant’s utility, leaked documents also mention potential partnerships with Uber, OpenTable, Instacart, and Ticketmaster. These integrations would allow users to complete tasks such as booking rides or making dinner reservations directly through voice commands. Amazon aims to onboard approximately 200 partners within three years to expand Alexa’s scope.
Lauren Raemhild, an Amazon spokesperson, highlighted to Fortune the challenges inherent in building such a complex AI system. She noted that while generative models are effective for generating conversational responses, ensuring they can perform specific user tasks accurately without errors is another hurdle.
Despite these challenges, Amazon is determined to enhance Alexa’s appeal with advanced features like personalized voice recognition and moderated interactions for younger users. The existing version of Alexa will remain free, with Alexa Plus offering new, exclusive features for paying subscribers.
Financial Background and Strategy Shift
Amazon’s financial struggles in its devices division add another layer of urgency to the successful rollout of Project Banyan. Reports from July 2024 indicate losses of over $25 billion from 2017 to 2021 in the division encompassing Echo, Kindle, and Fire TV products.
The original strategy to sell devices at a loss while banking on increased spending within Amazon’s ecosystem did not deliver as expected. Echo devices were mainly used for simple tasks like setting timers or checking the weather, limiting their profitability.
Amazon´s downstream impact (DSI) metric, which estimated the overall financial benefits of a product by tracking related ecosystem spending, initially justified investments in devices. However, this metric was not as effective for the Echo line as it was for other Amazon products like the Kindle.
Since becoming CEO in 2021, Andy Jassy has pushed for profitability, scaling back unprofitable projects and emphasizing revenue-generating initiatives. During a recent earnings call, he stated that the future of voice assistants lies in their ability to take practical actions, not just deliver information.
Rising Competitive Pressures
Amazon’s development struggles are playing out against a backdrop of fierce competition in the AI voice assistant market. Google’s Gemini Live, unveiled at the Pixel 9 event in August, has quickly gained traction.
Known for its rapid, hands-free interactions and multitasking capabilities, Gemini Live is being praised by Marc Benioff and others for setting a new benchmark in consumer AI performance.
OpenAI’s Advanced Voice Mode for ChatGPT, which expanded to worldwide users in October, offers customizable intonation and near-instantaneous response times of approximately 320 milliseconds. These features have pushed the boundaries of what users expect from AI-driven voice assistants, adding pressure on Amazon to refine its offering to compete effectively.
Meanwhile, Microsoft’s Copilot, has become a fixture in productivity settings, assisting users across Office 365, though it has faced criticism for lacking the dynamic voice capabilities of its rivals. Benioff referred to Copilot as “Clippy 2.0,” pointing out its limitations in keeping up with modern helpful AI engagement.
But it would be unfair do directly compare Microsoft’s Copilot, which not specifically targets the voice assistant market, to Gemini Live, ChatGPT´s Advanced Voice Mode, or a future AI Alexa version. It´s mainly an AI agent, integrated across Microsoft 365 applications like Teams and Word for productivity tasks.
While Project Banyan faces numerous challenges, its outcome could influence Amazon’s standing in the landscape of AI voice technology. Competitors are already setting new standards with features that prioritize speed, personalization, and multitasking. Amazon’s ability to address its current technical and strategic hurdles will be key to its success in this increasingly competitive market.
