Conversational AI — A New Wave Of Voice-Enabled Computing

Conversation is the key to the next wave of AI infrastructure investments. Photo: Thomas Szynkiewicz

In 2014, Andrew Ng predicted that by 2020 more than half of all web searches would be non-text and instead they would instead be image and voice-based. While voice has yet to overtake the keyboard in terms of search traffic we are seeing a new inflection point in the voice market the rise of Conversational AI. Moving beyond simply voice commands and by combining the latest in voice recognition and language parsing technology with text-based smarts of interactive chatbots that has been developing rapidly over the last few years, Conversational AI promises to be a rich new vein of technology innovation.

A range of new digital services are emerging as companies explore this new landscape of voice-first interactive applications that have the capacity to provide rich information as well as learning from interactions with users.

Last October, Gartner named conversational interfaces the top trend in digital commerce and predicted that by 2020, 70% of all companies will have tried conversational platforms with expanded sensory channels such as augmented reality and 25% of companies will have systems in production.

Voice Applications, just like mobile apps on phones extend the functionality of a growing number of voice-enabled devices such as smart speakers in the lounge room. And there’s now a large and deep market for voice apps — estimated to be over 50,000 between the two main platforms Amazon Alexa Skills (Amazon's name for voice apps) Store and Google Actions (Google's name for voice apps). 

Amazon claims that 100 million devices have now been sold with Alexa installed and Google has claimed theirs to be over 1 billion. In the more specific and newer marketplace however of smart speakers, Amazon and Google appear to have forged ahead from Apple and others players.

Is Home Voice now a Two-Horse Race?

Apple’s acquisition of Pullstring a marketplace for voice-app developers in February perhaps acknowledges the battleground for in home voice-access is now a two-horse race. Their own device the Apple HomePod may have been losing traction in consumer interest over the last year to rivals as this chart of relative Wikipedia page views (below) illustrates. Pullstring’s main focus according to their website prior to the acquisition was “Voice Applications for Amazon Alexa.’’

Using Wikipedia page views as a guide to consumer mind-share (which has been shown in research to correlate to market share for new products) the data indicates Google and Amazon are neck-and-neck but between them dominate the segment commanding about 78% of the markets attention.

Facebook’s new Portal smart displays launched late last year with Amazon's voice-controlled intelligent personal assistant service Alexa built-in and independent smart speaker maker Sonos now supports Amazon, Apple and Google voice applications.

A report by Consumer Intelligence Research Partners last year suggests Apple’s HomePod had a 6% market share in the smart speaker market in the U.S. Some, including Forbes columnist John Koetsier, wondered if the actual figure could be significantly lower than this, as the revenue from a 6% market share does not line up with quarterly revenue result. 

Whoever wins the battle for the voice in the lounge-room (and the car, hotel and aeroplane) another frontier where there’s lots of investment and activity is voice infrastructure, applications and cloud services.

A New Generation of Voice Tech Companies Emerge

The evolution of the human-computer-interface has always been seen as a catalyst for enabling seismic industry changes and huge opportunities for business growth in the tech sector. Each major new wave of technology innovation can be characterized by changes to the way in which we interact with computers.

The move from arcane text-only interfaces to graphical, easy-to-use personal computers was a key success factor for both Apple and Microsoft in their early days. And the multi-touch mobile interface on Apple’s iPhone distinguished it from other early smartphone competitors and led it to becoming the most successful new technology product ever created.

Voice is part of a broad and ongoing quest for more natural user interfaces or NUIs as they’re known involving touch, gesture, voice and even thought. Voice technology has been around for a while but with the growing use of sophisticated AI techniques is now moving into a new era — real time “conversation”.

Conversational AI is an area of growing interest among investors and corporations alike. In addition to Apple’s acquisition, tech giants Microsoft and SAP have each also acquired Conversational AI companies in the last year. Semantic Machines was acquired by Microsoft in 2018, and SAP acquired the Paris-based Recast.AI which is now the basis of SAP Conversational AI - their chatbot platform for developers.

Conversational AI Blossoms in Verticals & Infrastructure

Over 50 Conversational AI startups have been formed in the last five years making them a new and relatively young cohort of tech startups that is evolving rapidly. Many of the newer companies in this space established in the last year or so are being formed with a specific industry or functional focus such as:

• Retailer: Heyday, Montréal, Quebec, Canada 
• Travel & Hospitality: Dialogue Shift, Berlin, Germany
• Recruitment: Curious Thing, Sydney, Australia

Conversational user interfaces are moving beyond text chat on a keyboard to real-time interactive voice applications with feedback loops that enables AI models to learn from and improve the advice they give interactively.  

As this category expands and evolves, expect to see a number of research-backed natural language processing and infrastructure services companies such as Arria NLG (London), Sydney-based Appen (Sydney) and Pulse Labs (Seattle) become increasingly strategic and valuable in this next stage of the voice ecosystem.