With Alexa+, Amazon Nails the Value Proposition for genAI
Amazon’s long gestating genAI version of Alexa, Alexa+, is coming with all the capabilities you’d expect and more integration than you might think. The key is that it will be part of Prime (for now).
Context
Amazon is a big believer in AI, "not because it is cool,” says CEO Andy Jassy, but “to solve real, practical problems." Jassy assures us that Amazon has a great relationship with NVIDIA, but also designs its own Trainium 2 silicon for better price/performance, and works with its own models and everyone else's. Amazon is using AI internally for customer service, Marketplace product intake forms, better sizing information, shopping help with Rufus, and inventory management. Amazon is all about AI. However, while Amazon was early with Alexa, it has languished in the age of genAI.
Alexa may have stagnated, but that doesn’t mean it isn’t successful. There are over 600 million Alexa-capable devices, and usage is growing: up 20% in 2024 over 2023. However, Amazon traditionally did not monetize its hardware (like Apple), preferring to sell Echo speakers and displays, Fire tablets, sticks, and televisions, and other hardware like wearables either at cost or even at a loss.* The business model drove engagement, engagement drove Prime subscriptions, and Prime subscriptions drove ecommerce sales. Internally, this was called, “downstream impact.” And while I have written that there were other strategic reasons to invest in voice AI, there were two clear problems with downstream impact accounting:
1. The added ecommerce profit lift did not cover the cost of the hardware division in the past
2. genAI is wildly expensive to provide
Amazon needs to find a way to charge for its AI services in the home, while still keeping the strategic option of a potential computing platform that it controls and can plug its other services into.
New Capabilities
Alexa has always been able to set timers, play music, keep your child occupied, answer basic search queries, and set more timers. By combining its own AI with Claud models from Anthropic, Alexa+ is more conversational and personalized; the more you interact with it, the better the experience is supposed to get. It is also able to take actions on your behalf. The idea, according to product manager extraordinaire Panos Panay is that, "Alexa+ can plan a date, book the restaurant, and text the babysitter." Trip planning was another proposed showcase for Alexa+ genAI, because everyone pitching genAI seems to think that consumers are constantly planning trips to places and trust AI to do a better job than humans.
Alexa was already integrated into tens of thousands of products, and Amazon is leaning on those APIs to enable Alexa+ to do more with them. Amazon showed off a handful of vendors at its launch event, and the press release name-checked GrubHub, OpenTable, Ticketmaster, Yelp, Thumbtack, Vagaro, Fodor’s, TripAdvisor, Amazon, Whole Foods Market, Uber, Spotify, Apple Music, Pandora, Netflix, Disney+, Hulu, and Max along with smart home devices from Philips Hue, Roborock, and Lutron. For web services that don’t have API integration, Alexa+ Agents can do things for you on the web. The example given was scheduling an appliance repair, but I’m skeptical that this will consistently work well – there are too many corner cases, poorly implemented web services, and login/account complications never mind the security and privacy questions.
Alexa+ understands context from past interactions, devices and locations in the home, and anything you’ve asked it to remember. Alexa+ can suggest multi-device routines and make changes based on indirect natural language. This can be simple and powerful; Alexa could already play music or control lighting in different rooms, but now that Alexa+ better understands natural language and concepts like floorplans, you can ask to turn off the lights downstairs, “play music everywhere but don’t wake the baby,” or turn up the thermostat if you complain to Alexa that you’re cold. The most impressive live demo we saw was moving music from a movie from an Echo Show to a Fire TV and then asking to "jump to the scene in the movie." I have confirmed that jumping to movie clips is not supported in the version of Alexa+ being launched in March, but it is expected to be added eventually.
Generative AI also allows Alexa+ to have deeper conversations around sports and media. That conversation can be multimodal; Amazon showed off a demo where you’ll be able to ask Alexa+ to jump to a specific movie or TV clip, but this is a feature that will not be implemented at launch. Amazon knows that kids use Alexa as an indefatigable companion, so there are enhanced integration with Amazon’s Kids+ subscription. My 12-year-old already uses the Echo Show to skirt his screentime restrictions on other devices; with Alexa+ he’s never going to leave the thing alone.
Echos aren’t really used for writing, so Alexa+ isn’t being pitched as a way to edit the tone of your document, but it can accept multi-modal input (camera) with contextual awareness so you can feed it documents and ask it to take actions on it. The best example at the launch event was forwarding a soccer schedule email and then having Alexa+ automatically enter all the information on the calendar –complete with reminders the day beforehand every time you are in charge of bringing snacks. Solving family calendars would address a real pain point for families.
Alexa+ can also be taught preferences, which can feed into its actions and recommendations. We'll see how well Alexa+ understands the rules of keeping kosher when I start testing it.
As you would expect, Alexa+ is fully integrated with Amazon’s own products. If you have Ring cameras or doorbells Alexa+ can give you a summary of things around your house – which is incredibly useful if you're using it for remote security -- including package delivery (naturally). In Panos' case, he discovered that his Amazon driver had made friends with his dogs.
Alexa+ works on most Echo devices in some capacity, but it has been designed first for Echo Show models with touchscreens. There is also a new Alexa.com web site, and a new smartphone app. Alexa+ is free with Amazon Prime or $20/month on its own.
Analysis
Alexa is about to get all of the genAI improvements that every other tech company is promising, along with some unique integrations with Ring, Echo speakers, FireTVs, and "thousands" of third-party services and websites. The key is Amazon's pricing model: Amazon is making Alexa+ free with a Prime subscription, or $20/month without one. The standalone pricing is just there to set perceived value, since literally no one will pay for this separately (Prime currently costs $15/month or $139 annually, a full $100 less than buying Alexa+ monthly). This framing is crucial: by including Alexa+ in Prime, Amazon sidesteps the challenges getting people to add another subscription, for a service they're not sure they'll need, and that won't work perfectly/have full integration with services at launch.
We've seen a lot of AI services launch recently, and while the models are amazing at coding and writing bad poetry, most practical integrations have been underwhelming. How well will Alexa+ work in the real world? Who knows. But consumers are likely to give Amazon the benefit of the doubt since they aren't paying more for it up front. Amazon can establish a foothold in genAI alongside everyone else, and it only needs to add enough marginal utility to justify future Amazon Prime price hikes.
Amazon is also hoping to overcome another strategic debt with Alexa+. While Amazon has sold millions of smart speakers, Amazon was not able to get traction on the computing platforms consumers use most: smartphones and the web. The new Alexa+ experiences focus primarily on 15" & 21" Echo Shows with touchscreens, a new smartphone app, and a new website. Amazon is leveraging all the relationships and APIs it has built up over the years for its smart speakers to now make Alexa capable of taking action around your house and online.
It’s far from perfect, but if Amazon can actually get consumers to use Alexa+ across platforms thanks to its installed base of Echo speakers and FireTVs, Amazon at least has a chance to be relevant in the scrum of companies vying for consumer AI subscription dollars. Apple and Google natively live on the primary computing device, smartphones, but Apple appears to be well behind what Amazon is doing with Anthropic, and Apple never established a commanding presence in the home. Google needs to move upmarket – Apple has the majority of premium users globally – but it is working on that with Pixel and Samsung, and Google can extend into the home with smart speakers and televisions of its own. Microsoft is all-in on the PC, but has little leverage outside of it. Meta is so starved for a platform that it can control that it is trying to jump straight to your face. The dedicated AI model-makers like ChatGPT are likely to end up ingredient brands across platforms (at least for consumers); Anthropic is already a part of Alexa+.
To discuss the implications of this report on your business, product, or investment strategies, contact Techsponential at avi@techsponential.com.
*Kindle and Ring are exceptions. Kindle is a razor/blade model where the hardware has always been sold at a profit and it directly drives content sales. A Kindle is basically a vending machine that consumers happily pay for. Ring is a subscription model, and while the subscription fees are modest, they apply per camera (so multi-camera households pay more), and Ring has been able to generate volume sales. The hardware is also sold at a profit, though margins appear to vary widely by product and during sales.