Conversational UI approaches pushing AI development forward
Find out about the latest AI conversational UI approaches available for developers today from Apple, Amazon, Google, IBM, and Facebook.
Conversational UI oriented applications and Chat-Bot development tools are now emerging from all of the leading platform vendors including IBM, Google, Amazon, Facebook, and Microsoft. There are a wide variety of implementing applications that leverage them that can include AI, search, natural language processing and business process components. This new generation of tools complements existing chat-bot tools.
In some senses, the idea of the conversation UI goes back as far as the Turing test that proposed the challenge of creating AI conversational agents that could not be distinguished from humans. In 2014, a program called Eugene Goostman, convinced a group of humans that he was a 13-year old Ukrainian boy. The fundamental technology is improving in its recognition and getting better integration into application platforms that make it practical for building enterprise and consumer applications.
A key component for delivering this kind of capability are the frameworks for interpreting conversational style commands and presenting information in a useful manner. Some of the other leading contenders in creating lifelike conversations in chatterbox challenges designed to fool humans have included Cleverbot, Elbot, and Ultra Hal. The initiatives from the main cloud and mobile platforms providers are really just the tip of the iceberg. Other CUI interfaces are being built onto messaging platforms like WeChat, Telegram Messenger, Kik, and Slack as well.
Google has several different components in its conversational UI strategy built upon the core Google Now conversational search UI. On a physical level, the Google Home device promises a physical interface for IoT apps, Google searches, and retrieving personal information. If Google follows in the footsteps of its Android philosophy, Google is likely to open up the core party to third party consumer electronics manufacturers to drive the price down. This could build on Google CEO’s Sundar Pichai’s vision in which the computing device fades away.
Google is calling the core platform the Google Assistant. The underlying components weave together a physical microphone as part of Google Home, a messaging app called Allo, and a Web and phone based component around Google Now. Developers can connect to the Google Assistant infrastructure by adding schema.org markup to emails and Web applications. This helps users to see information in a more appropriate context.
The Microsoft Bot Framework includes components for image base search, summarizing text, answering what if questions, and creating captions. The Skype Bot developer platform promises to add in support for integrating chat applications to Microsoft office workflow as well. The core framework includes support for SMS, Skype, Slack, and Office 365 mail services. The Microsoft Cortana conversational API promises to make it easier for developers to add conversational features to Windows and Web applications.
The Facebook Messenger app has risen to become one of the most popular chat platforms with the growth in the social networking service. The company has announced that it is opening up the Messenger platform to allow developers to create AI-powered chatbots engage consumer through the platform. More than 50 million companies are already using Facebook to engage with consumers and more than 1 billion business messages are sent every month. The Facebook bot platform promises to make it easier to automate these consumer engagements.
The Messenger bot platform promises to make it easier to developers to work with the send and receive APIs for messenger that can leverage identity to send and receive text, images, buttons, bubbles and calls to action. Companies leveraging the bot platform include 1-800 flowers, Hipmunk, Staple, and sales force. It can leverage Facebook’s Wit.ai Bot engine to build conversational bots that parse messages into structure actionable data. This promises to make it easier to implement apps that work with mobile apps, home automation, and eventually even robots. These bots could also work with the Facebook “M” Personal Assistant designed to make it easier to make purchases, book restaurants, or make appointments.
Amazon’s foray into the conversational UI includes the Alexa conversational UI, and the Echo and Dot physical microphone interfaces. Consumer Intelligence Research Partners, a consumer research firm estimates that Amazon has sold about 3 million Echo devices since its launch in late 2014. Amazon is making it easier for developers to add new apps called “skills” to Alexa through the Alexa Skills Kit. A Smart Home Skill API also includes support for adding controls to home automation equipment. Amazon also wants to make it easier for consumer electronics manufacturers and IoT device makers to add Alexa to any device with a speaker and microphone using the Alexa Voice Service.
Amazon expects this to be just the beginning and is incentivizing development of new Alexa apps and integrations through a $100 million dollar fund to drive voice technology innovation. Some of the early pioneers funded have included Sutro, a connected smart pool monitor; Invoxia, a music player for the fridge, Petnet, a smart feeder for pets; and Mara, an intelligent voice based running assistant.
The IBM Watson Cloud platform is calling the CUI cognitive computing. The core platform is already being used in healthcare, finance, and business applications for a wide variety of applications. IBM made headlines as an early pioneer in this field by winning the Jeopardy game show. The platform includes image recognition components with the AlchemyVision API, emotion analysis, and tone analysis from social media integrations, in addition to its strong natural language processing capabilities.
Apple’s Siri interface provides a CUI for limited functionality within the company’s highly curated platform. In the long run this may ensure a better user experience. Apple has still not announced an official API, but limited access is available through the Search features built into iOS. Developers can also craft their applications in ways that make their apps more accessible to Siri in a limited way as programmer Nevan King explained.
But critics like programmer Marco Arment believe this may hold Apple developers back from fully taking advantage of the evolution of a rich conversational driven ecosystem. He said, “Today, Amazon, Facebook, and Google are placing large bets on advanced AI, ubiquitous assistants, and voice interfaces, hoping that these will become the next thing that our devices are for. If they’re right — and that’s a big if — I’m worried for Apple.”
Learn how edge devices can leverage AI technology.
Simplified standard IoT API frameworks make talking to devices easier
Learn how Java visionaries predicted the the IoT evolution 20 years ago