New ICT priority: systems that can talk and think

New conversational technology has reached the mainstream for corporate ICT groups. Businesses from nearly every vertical market should get ready to deploy these so-called “chatbots” both internally and to facilitate interactions with customers.

In the last decade, millions of consumers have started to use virtual personal assistants (VPAs) also known as chatbots—such as Amazon Alexa, Apple Siri, Google Assistant and Microsoft Cortana. Starting with answering simple questions like “what’s the weather like outside?” these tools have become an important new technology domain called “conversational platforms.” What works for consumers—interacting with a computer faster via speech than by using a mouse, keyboard, or touch interface—is about to become the latest wave of corporate computing.

The chatbot wave intensified in 2016 with Facebook’s announcement of a developer-friendly platform to build chatbots on Facebook messenger.

This first generation of natural language business chatbots for example helped call centers with straightforward interactions such as solving customer service issues. Instead of clicking on a menu of choices or speaking predetermined commands, callers can speak as if having a conversation with another human.

These early chatbots worked well mostly for requests that could be communicated in a very structured, step-by-step process and be fulfilled by relatively simple answers--such as travel reservations. They got users comfortable talking to a device as if it was another human, sparking interest in more free-form conversation.

In the last few years, communication-style applications have become far more sophisticated, powered by conversational forms of artificial intelligence (AI) and Machine Learning (ML). The new generation of chatbots can have a fully interactive dialogue with users. They can remember and learn from past conversations, recalling a user’s preferences from one interaction to the next. They can take commands and answer queries that are unstructured, applying conditional reasoning (“if x, then y”) and respond by executing a function, presenting content or asking for additional input. They can even veer off to discuss a different issue, then use memory to return to the original topic.

Case in point: Amy, a virtual assistant developed by x.ai, which is a fast-growing startup. Amy can schedule meetings and calls by issuing and responding to emails. No need to use precise, stilted phrasing. Amy can interpret and act on language as unstructured as: “we should connect sometime.”

Systems that can think and talk to users based on AI and ML are spreading fast. ICT groups need to understand how the new technology can improve internal operations and facilitate dialogue with customers in order to start building prototypes that will one day fulfill the needs of users at scale and at unprecedented speed.