Amazon’s Offensive For the AI Platform of the Future

At Amazon’s Developer Event in Las Vegas, the company came out swinging, looking to be the sole platform for machine learning and artificial intelligence of the future. Their announcements are important in a couple of key contexts.

Thinking about the Echo and Alexa platform, we are at the stage in the game where the early entrant may be the winner or a majority winner. We are in a race for the voice-first platform and the players are Apple, Google, Microsoft, and Amazon. While I’m not going to discount the potential of all four, I believe I can create the strongest arguments that Apple and Amazon are the best positioned to be the dominant players in markets where they compete. Apple, for reasons I outlined here, where they can likely be the default, and Amazon because of the reasons I will lay out. First, I need to make a grounding point.

With voice, we are talking about the opportunity to develop a relationship we are comfortable with. Having used all the main voice first UIs, I still think Amazon is the best because you can address it by name. I know this sounds like something small but it is a big point in how we interact. Saying, “Hey, Siri”, “Hey, Cortana” or “OK Google” are natural ways to speak but saying, “Alexa” is a more organic way to kick off a “conversation”. Yes, the other three can fix this but, right now, we are at a stage where the consumer experience with a voice-first UI matters. The key for all of these companies is to get regular consumers using their voice-first UI and to use it for more than just saying, “play a song” or “set the alarm.” Apple, Google, Microsoft, and Amazon need people to train their AI agents. That will only come as they use them often and for more advanced things necessary for making these agents smarter.

The reason the Echo is interesting to me is it forces you to a voice-first paradigm. This is on point with what Carolina outlined yesterday, where adding a screen could hinder the Echo because it allows the user to go back to an old habit or UI metaphor by using touch instead of voice. This is a similar counterpoint I offer to Apple’s positioning that the best place for Siri is in your computers. The problem with this viewpoint is it allows us to use our old behaviors of touching or typing vs. being “forced” to speak to Siri. Consumers will almost always stick with default behaviors when given the option. The beauty of the Echo is we are not allowed old behaviors but forced to create new ones. It just so happens to be the ones that are helping Amazon take the lead.

It is with this grounding I think about how certain companies are positioned for the future. If Amazon continues on this path, they are likely to have more consumers using their AI interface in ways that are truly helping them build an AI platform than other competitors in the market. This is also a part of their grand strategy in announcing that anyone can now build on top of their AWS platform — including natural language processing, machine and visual processing, and even their artificial agents using a set of tools Amazon now offers.

I think about who else is laying the groundwork for third parties to develop such comprehensive solutions in AI and machine learning and come up empty. Not to discount what Google has done but their strength is in search, not AI. Microsoft has some compelling assets around Cortana but they have yet to prove the masses will embrace these new tools. Apple is focused on a more introverted approach to AI. Amazon is looking to democratize it faster and with a holistic toolset more than the others.

While it is still early, the foundation is being laid. A big part of the analysis is to look at who is laying the right foundation and looking to plug as many holes as they can. Right now, I feel that is Amazon. They are putting the pieces together to offer a suite of tools which will enable the next generation platform of AI and machine learning that can tie together hardware, software, and services from a standard AI and ML platform. The kicker is, anyone can build these tools, not just Amazon.

Interestingly, Amazon has made it possible for all sorts of companies to create competing products to the Echo. Which begs and interesting question: Is the Echo simply a showcase for Amazon’s cloud services which power it? Is Amazon’s end goal to not necessarily make the hardware but provide the platform to enable a new generation of hardware built around their AI platform?

All of this to say, I’m getting pretty bullish on Amazon.

For the interview with Amazon’s VP of Alexa, check out this article by Steven Levy.

Published by

Ben Bajarin

Ben Bajarin is a Principal Analyst and the head of primary research at Creative Strategies, Inc - An industry analysis, market intelligence and research firm located in Silicon Valley. His primary focus is consumer technology and market trend research and he is responsible for studying over 30 countries. Full Bio

8 thoughts on “Amazon’s Offensive For the AI Platform of the Future”

  1. “The kicker is, anyone can build these tools, not just Amazon.”

    You mean anyone can build the tools that tie into Amazon’s platform or anyone can build these tools to compete with Amazon?

    1. anyone can build the tools. They don’t need to integrate with Amazon commerce more specifically, but someone can build an Echo competitor that goes after something else than shopping for example.

      1. In response to Ben’s article in June 2015 (17 months ago), I commented

        “Amazon, by offering Echo voice to other companies, is essentially making it the new AWS. ”

        It seems that after more than a year, Amazon is now seriously pushing this.

        There are several implications;

        1. Amazon’s goal is not to drive profits to Amazon’s other services. Instead, if they are thinking like AWS, their AI business will be a very profitable and independent segment.
        2. Scale will be the differentiator and Amazon will have most of it, as new startups come up with innovating new ideas.
        3. AI, at least in the applications that most people use it for, will be commoditised and accessible to even the smallest single person start-ups. This will be even more so if Google and Microsoft enter the same market and compete to reduce prices.
        4. Like how AWS empowered tech startups to beat incumbents, this new AI device will likewise breed a new brand of AI powered services, that will run over what the incumbents currently offer. Companies that try to differentiate based on their AI prowess and not by their ideas, are the ones likely to lose first.

        All this is derived from a comparison with what AWS has already achieved, so it’s not so much of a prediction but rather a history review.

        1. “This will be even more so if Google and Microsoft enter the same market and compete to reduce prices.”

          You know they will. MS just put together a 5000-person team dedicated to AI & machine learning

          I’m just curios where Apple stands because they don’t have anywhere near the infrastructure or manpower dedicated to ambient computing / AI

          1. Yes I know they will.

            Regarding ambient computing, I think it’s important to note that Amazon’s efforts are in no way limited to ambient computing. Their AWS-ish strategy for AI has much broader reach and implications for all AI applications except maybe for the AI-on-your-device approach that Apple seems to be taking.

            Regarding Apple, it is easy to imagine that they are hard considering the privacy implications, which are huge for ambient devices, especially those that connect with third party services. Putting these always-on devices that can potentially send back anything they hear over the Internet (and have the AI to figure out what you’re doing from ambient noise alone) is a huge concern for me. There has already been allegations that the Facebook app is always using the smartphone microphone to overhear your conversations and is using that to target ads to you.

          2. Regarding Microsoft’s 5000 people team. The idea behind an AWS-ish service is that this infrastructure can become obsolete very quickly. AWS made it unnecessary for startups to invest in their own server farms and maintenance teams, unless they were using some arcane setup. This is what could very quickly happen with AI.

Leave a Reply

Your email address will not be published. Required fields are marked *