How does Amazon LEX change the game?

How does Amazon LEX change the game?

If you haven't heard the news, Amazon just announced LEX, a new conversational interface service added to its AI toolkit. Currently in limited preview, LEX is based on Alexa's core ASR and NLU technology and is Amazon's answer to competing NLU and "Chatbot" frameworks currently available in the market. Although I've yet to explore the service's full set of capabilities, one area that caught my attention is its multi-turn capabilities and slot filling. Technically, this would allow developers to build conversational interfaces that perform back and forth interactions to provide and collect information. If elegantly done, this has the possibility of simplifying the process of developing conversational interactions, moving from a basic Question-Answer dialogue system, to an interactive dialogue interface. This isn't new however, and Amazon is yet to demonstrate how their service compares to others. Although I wouldn't qualify this as a game changer, I do believe that the service's introduction into the AWS ecosystem will significantly push the adoption of multi-channel conversational interfaces in the market. With its massive reach into the developer community, you can count on seeing a good number of LEX-powered products go live next year.

But where does that leave startups and smaller innovators competing in the NLU and conversational interface space? Probably not in a good space for many, but potentially in an interesting position for others. This will force small players to adapt, specialise and further innovate. Indeed, when we built our core NLU engine for our BBX platform, multi-turn conversational design and slot filling was developed as a core feature. With the approach we've used, we can embed any number of multi-turn dialogues as part of a bigger conversation, allowing us to build uniquely sophisticated conversational interfaces. But my vision for our NLU engine is much broader, which is why I've always seen this approach as a basic feature, not a game changer. With LEX and other players on the market, this further provides smaller companies the incentive to innovate beyond what large companies are building. We're nimble and in a position where we can quickly shift and adapt.

With such a competitive environment, we're set to see quite a few innovations in the coming months and years. Indeed, 2017 will undoubtedly prove to be quite an exciting year in the AI and tech industry, which I look forward to.


要查看或添加评论,请登录

Alex Boudreau的更多文章

  • A first look at Gemini

    A first look at Gemini

    It was only a matter of time for Google to respond to OpenAI's apparent dominance in the GenAI space. According to the…

    1 条评论
  • GenAI in Clinical Trials

    GenAI in Clinical Trials

    As you read the title, you might be asking yourself how Generative AI applies to Clinical Trials. After all, GenAI can…

    5 条评论
  • Exploring ways to enhance K12 learning using a purpose built search engine

    Exploring ways to enhance K12 learning using a purpose built search engine

    After a few years of helping out my son with his studies, I’ve noticed that resources provided to students often tend…

  • Impactful Innovation

    Impactful Innovation

    Recently, I've had quite a few people ask me what I love about working at Pindrop. It's indeed a good question to ask…

  • What I learned building an ML powered enterprise chatbot platform

    What I learned building an ML powered enterprise chatbot platform

    For the past few months my team and I have been hard at work building an enterprise grade machine learning powered…

  • AI made in Sydney

    AI made in Sydney

    Do your Chatbots use a NLU engine that generates optimised compact feature vectors for higher performing machine…

    2 条评论
  • I built a self-aware Messenger bot

    I built a self-aware Messenger bot

    With Facebook’s announcement of the Messenger platform, chat bots have made the front page in the media and triggered a…

    4 条评论
  • Building Scalable Microservices with Google Cloud Platform

    Building Scalable Microservices with Google Cloud Platform

    Over the past few years I've seen the Google Cloud Platform evolve tremendously. In fact, in its early days, there was…

  • How to delight your app customers?

    How to delight your app customers?

    The app market is highly competitive and is quickly becoming a winner takes all space (or nearly). In fact, according…

  • Introducing personalised shopping

    Introducing personalised shopping

    2015 is a year of experimentation, a year where customer experience is finally seen as the standout imperative for…

    2 条评论