Amazon unveils Alexa+, powered by generative AI
Amazon's Alexa is undergoing its biggest overhaul since debuting more than a decade ago.
On Wednesday, Amazon said it was giving Alexa a new brain powered by generative artificial intelligence. The update, called Alexa+, is set to make the virtual assistant more conversational and helpful in booking concert tickets, coordinating calendars and suggesting food to be delivered. Alexa+ will cost $19.99 a month or be included for customers who pay for Amazon's Prime membership program, which costs $14.99 a month. It will begin rolling out next month.
'Until right this moment, right this moment, we have been limited by the technology,' Panos Panay, the head of Amazon's devices, said at a media event. 'Alexa+ is that trusted assistant that can help you conduct your life and your home.'
With the changes, Amazon is aiming to catch up in generative AI for everyday users. While the Seattle company has in recent months made up for lost time in AI products and services that it sells to businesses and other organizations, its grip on consumer AI products has been narrower. Alexa's upgrades, which were first teased in 2023, are Amazon's biggest bet on becoming a force in consumer AI.
The moves are also an opportunity to reboot Alexa, which has been perceived as having fallen behind other virtual assistants. In recent years, Alexa's growth in the United States has generally stagnated, according to the research firm Consumer Intelligence Research Partners, with people turning to the assistant for only a few main tasks, such as setting timers and alarms, playing music and asking questions about the weather and sports scores.
At Wednesday's event, Panay and other Amazon executives demonstrated how Alexa+ could do those things in a more personalized manner. Alexa+ could identify who was speaking and know the person's preferences, such as favorite sports teams, musicians and foods, they said. They also showed how a device powered by Alexa+ could suggest a restaurant, book a reservation on OpenTable, order an Uber and send a calendar invitation.
Alexa, which was a brainchild of Jeff Bezos, Amazon's founder, debuted in 2014, wowing people with its ability to take verbal requests and translate them into actions. It became a symbol of Amazon's innovation. Over the years, the company has highlighted some Alexa-connected devices, including Echo speakers, a connected microwave, a wall clock and a twerking teddy bear.
But wild experimentation has been out since Bezos stepped down as Amazon's CEO in 2021 and handed the company over to Andy Jassy, a longtime executive. Jassy reined in Amazon's expenses, killed some projects that appeared to have no obvious prospects and oversaw layoffs. In 2023, he hired Panay, a Microsoft executive, to oversee devices.
Panay's top responsibility was to bring generative AI to Alexa and to unlock the promise of the all-helpful assistant that Amazon had long envisioned. Soon after Panay started, Amazon said it was rebuilding Alexa's brain with the kind of technologies that underpinned OpenAI's ChatGPT chatbot.
'The re-architecture of all of Alexa has happened,' Panay said Wednesday.
As Amazon worked to update Alexa, competitors leapfrogged it. ChatGPT, for example, can hold extended, in-depth conversations, with some people developing emotional — and even sexual — relationships with AI personas.
(The New York Times has sued OpenAI and its partner, Microsoft, claiming copyright infringement of news content related to AI systems. The companies have denied the claims.)
Bringing generative AI to Alexa was not easy because the virtual assistant faces challenges that a chatbot does not. Alexa might serve multiple users in a household, for instance, so it needs to distinguish who is speaking and personalize the responses.
Amazon also wants Alexa to be at the center of people's lives and connected to multiple smart devices and services, which is complicated. It must integrate multiple AI systems, including ones built by Amazon and the startup Anthropic, and interact with devices such as smart lightbulbs and with apps including Ticketmaster. Amazon also gave Alexa+ a personality, even training it with comedians to make it funny.
'In the fall, it was just too slow,' Panay said in an interview.
Generative AI has also been afflicted by 'hallucinations,' or when the AI systems serve up incorrect information. Because Alexa interacts with the real world — playing a song, ordering a product, turning off an alarm — Panay said Alexa had to reliably get things right.
He said he believed Alexa+ was finally both fast and accurate. 'I think people will fall in love with it pretty quickly.'
This article originally appeared in
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Observer
3 days ago
- Observer
Oracle shares surge to record high on AI cloud demand
NEW YORK: Oracle shares surged 14 per cent to breach the $200-mark for the first time on Thursday, after the company raised its annual revenue forecast, driven by strong demand for its AI-related cloud services. Confidence in the software sector remained strong despite geopolitical tensions, even as analysts warn that US President Donald Trump's tariffs could undermine Big Tech's AI investments. Earlier this year, Oracle, whose cloud offerings help companies build their AI infrastructure, announced a joint venture called Stargate to deliver large-scale computing capabilities to OpenAI. "Oracle's once-stodgy image levels up to 'cloud-native mage,' and the competitive map now looks less like a classic three-player real time strategy and more like a battle-royale with everyone dropping in, looking for compute loot", said Michael Ashley Schulman, partner at Running Point Capital Advisors. Oracle expects total revenue to be at least $67 billion for fiscal 2026, CEO Safra Catz said on a post-earnings call. The Texas-based company's cloud services quarterly revenue rose 14 per cent to $11.70 billion. Its overall revenue of $15.90 billion beat estimates of $15.59 billion. At least nine brokerages have raised their price target post-earnings. Oracle trades at a forward price-to-earnings ratio of 25.86, compared to rivals Microsoft at 31.34 and Amazon at 31.80, according to data compiled by LSEG. Microsoft's stock has gained 12.16 per cent, while Amazon's has decreased by 2.8 per cent so far this year. "ORCL has entered an entirely new wave of enterprise popularity that it has not seen since the Internet era in the late 90s," analysts at Piper Sandler added. Shares of the company were last trading at $201.38. - Reuters
Observer
03-06-2025
- Observer
AI agents will soon assist you in real life
Eid holidays are just a few days away. Some people and families are planning to travel, while others just plan to search and book local hotels and for some activities that they and their children could engage in around the city as a retreat. I know for a fact that I personally need a quick getaway plan immediately after Eid to refresh, re-energise, and prepare for the remaining engagements set for the next balance months (the second half of the year). The good news is that all this (search and booking) is now virtually possible today on the information superhighway (the internet). The bad news is that it would require one to research so as to spend a considerable amount of time in order to have an ideal plan created. For the executives or the likes, this may be an easy affair via their personal assistant. For the rest, it's a tedious manual activity that one cannot avoid, unfortunately. Wouldn't it be just awesome if there were a digital assistant or helper who could fulfil the service for you? Welcome to Artificial Intelligence (AI) agents. What are they and how do they work? is what I will focus on in my article today. First things first, let me clarify that AI agents are not chatbots or voice assistants (like Amazon's Alexa, Apple's Siri, Microsoft's Cortana, and Samsung's Bixby). AI agents are digital assistants/helpers that go beyond just responding to user inputs/prompts where they actually act proactively and not reactively to user/your requests per se. Simply, they understand what you want (as in your goals) and find out ways (as programmed via logic) on what to do to get you the ideal results (for your goals). The typical AI consumer apps that many use today (e.g., ChatGPT, Perplexity, Gemini, Claudia, etc.) are based on LLMs (Large Language Models), yet AI agents are powered by LLMs. They are proactive (don't wait for you to ask things) but instead get things/tasks done for you. Examples include reading your emails, reviewing your calendar, scheduling meetings in advance, analysing and summarising reports, and/or making decisions on your behalf using a logic (programmed or taught to them). The sky is the limit on what these AI agents can do for you, and time will prove their worthiness (as you are witnessing today with ChatGPT and similar apps). How do these AI agents actually work? Quite simply, you set a goal, and it goes out to get it or them done for you. For example, referring to my wish at the beginning of the article today on a getaway travel plan, I would set a plan via the AI agent as 'Look for the best hotel deals in Indonesia under a specific budget X and closer to specific attraction X.' Obviously I will share the budget in mind and a few attractions I have heard of or been recommended. The AI agent would, in return, break down the query/input into smaller actions, surf the web, summarise if it found anything via the links generated, probably draft an email, communicate and negotiate with the hotel(s) and attraction(s) in question, reflect on the outcome received, and return with a recommendation (and sometimes even decide and take action for you, as you had programmed it/them to do). What more? These AI agents can also talk with other AI agents and support one another to get you the final result. Pretty cool? Yes, and trust me, it's not part of a science fiction movie but rather a reality in the making. AI agents are quite new and in the early stages of their full potential. Human intervention and/or supervision may still be required at times to ensure it works and is trained correctly. Nevertheless, like at the beginning of the consumer AI apps of the likes of ChatGPT and the rest, I expect AI agents to mature and become extremely useful in the coming year or so, where almost everyone (including myself) will have a digital helper at the office and/or at home, supporting 24/7 on every and any task, including Eid holiday getaways that I mentioned at the beginning of the article this week. Until we catch up again, I wish you a blessed and happy Eid Mubarak.
Observer
25-05-2025
- Observer
In India's congested cities, delivery apps cash in
Mumbai—In India's sprawling financial hub, Mumbai, armies of "dabbawalas" have for decades crisscrossed the city by foot and bicycle, delivering home-cooked food to office workers keen to avoid the searing heat and traffic-snarled streets. Now, across the country, young entrepreneurs are taking that tradition to the next level with the explosion of shopping apps that allow customers to get hold of not only food and drink but anything else from clothes to iPhones -- within minutes. The so-called quick commerce apps are redefining the retail game, disrupting not only e-commerce titans such as Amazon with their speed and efficiency but also long-established "mom-and-pop " stores that are no longer convenient enough. At a warehouse managed by online grocer BigBasket in central Mumbai, employees work with military-like precision to pull off deliveries in just 10 minutes. These warehouses are known within the industry as "dark stores", a reference to being closed off to customers. When a new order is received, a worker leaps into action, darting through aisles filled with everything from fizzy drinks to vegetables, packing a bag of groceries handed to a motorbike rider -- the modern-day "dabbawala", Hindi for "lunchbox man". Local tech companies have poured in billions to set up these nifty logistical networks across big cities, fuelling India's rapid shopping industry. - 'Unprecedented' - For millions of customers, it's an easy way to avoid shopping in the sweltering heat -- visiting multiple food stalls -- and spending hours navigating the country's notorious traffic jams. Growth has been "very strong", BigBasket co-founder Vipul Parekh told AFP, pointing to forecasts that indicate a compounded annual growth rate of more than 60 percent over the next two to three years. "When you talk of a large industry transforming and growing at this pace, that is unprecedented," he said. Delivery apps such as Getir or Jokr have faltered in Europe and the United States in recent years, as pandemic-induced demand wore off and rising inflation pinched customer wallets. But sales in India have soared from $100 million in 2020 to an estimated $6 billion in 2024, according to projections by market analysis firm Datum Intelligence. This could hit $40 billion by the end of the decade, according to investment bank JM Financial. Companies say India's quick commerce growth is partly down to the sheer scale of people living in tightly packed cities within a roughly two-kilometre (one-mile) radius of a "dark store", said Parekh. "The revenue potential in that catchment is very high," he said. A lack of many traditional supermarket grocery chains in India aids the business model, he said. Rinish Ravindra, a regular user, admits that they make him "lazy", but argues that the convenience is unbeatable. "I just press a bunch of keys and all of it comes delivered to my home," says the 32-year-old, who works in Mumbai's film industry. Local players have made rapid progress, but competition is heating up. Amazon is getting its act together, along with Walmart-owned Flipkart and billionaire Mukesh Ambani's Reliance Industries as they belatedly roll out rapid delivery offerings. "One of the problems with e-commerce players like Amazon is that, until now, they've relied on these big fulfilment centres that sit on the outside or outskirts of cities," said Satish Meena of Datum Intelligence. "These aren't suited for rapid delivery, which is why they now need to invest to build their dark store networks within urban areas." - 'Just order it online' - However, a more crowded industry threatens the sustainability of the sector that has already seen one prominent start-up go bust. "I sense that the market is good enough for two to three players," said Rahul Malhotra of Bernstein, a research firm, adding that the total addressable market may be worth around $50-$60 billion. "Some of the early movers, with hyperlocal capabilities obviously, have an advantage here." The sector could also face challenges from thousands of small, family-run shops. The Confederation of All India Traders, a leading industry group that claims to represent over 90 million small businesses, has called for "a nationwide movement" against newer platforms. Its president likened quick commerce to being a "modern-day East India Company", a reference to the rapacious British power that began in the 17th century to seize swathes of India, preceding colonial rule. For now, customers are voting with their wallets. "When I think of groceries, I think, 'I can just order it online'," said Ravindra.
