AI, chatbots and the customer dialogue of tomorrow

August 17, 2016

Google CEO Sundar Pichai’s prediction last spring that we are headed into an “AI first” world gives us some idea of where companies will be focusing their digital efforts going forward.

The current focus on customer journeys, omni-channel, digital transformation, platformisation and making data available is all about facilitating the construction of the businesses, innovation and customer experiences of tomorrow. What is truly exciting is what we do next.

AI Quick Facts:
  • The concept of artificial intelligence was coined in 1955 by John McCarthy, a researcher at Dartmouth, and was defined as follows: “…[when] machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.”
  • The research company Research and Market estimates that the AI market will grow from USD 60 million in 2014 to USD 6 billion in 2020
  • According to the Turing test, designed by Alan Turing in 1950, a machine is considered intelligent if a person engaged in conversation with that machine cannot determine whether she is conversing with a human or a machine. I recommend seeing the Oscar-winning film Ex Machina for an exploration of just this issue.

An obvious area of application for artificial intelligence is the effort to individualise customer dialogue in the consumer market. Achieving an omni-channel experience where customers and companies are able to carry on a coherent dialogue across multiple channels, or where customers can order products in one channel and pick them up or return them in another, is one thing. But for companies operating in a mass market and that are trying to create an individual relationship with each customer and to exhibit empathic and contextual understanding of the customer’s situation, it is difficult to imagine anything other than AI as the way forward. This poses a greater challenge than autonomous cars or beating human players in games of chess or Go, but we are on our way.

Advances in fields like language comprehension, image comprehension and sentiment analysis have contributed to the resurgent AI trend. One of the hottest trends this past spring was chatbots. Automated rule-based services that can help people with everything from weather forecasts, buying shoes and booking meetings, to financial advice and/or a virtual friend to talk to. Communication takes place via chat platforms that most consumers already use. For example, Facebook Messenger, Slack, SMS or email – meaning that you don’t have to download an app, you can just send an ordinary message to the chatbot and it will respond.

The more artificially intelligent the bot is, the more sophisticated the level of dialogue it will be able to manage. Most chatbots haven’t reached this level yet, but many players, including Facebook and Microsoft, believe that in the future, chatbots will be so smart that based on the data they have about the consumer and what the consumer communicates using natural language, they will be able to understand context and to hold personal, relevant conversations with customers. So what is interesting about chatbots is not to be found in the text interface itself, but in the algorithms behind it. The interface can also be human speech. According to the Meeker 2016 trend report, voice searches account for around 20% of all mobile searches using Google in the United States today. And the share is growing rapidly, mainly due to technological advances like improved speech recognition. The Chinese search giant Baidu expects a minimum of 50% of all searches to be by voice or image by the year 2020. In other words, voice-controlled bots such as Apple’s Siri, Amazon Echo or Microsoft’s Cortana are on the rise. In fact, Amazon Echo was the best-selling speaker in the United States in 2015.

Looking to the future, the next big step will be for the very concept of the “device” to fade away. Over time, the computer itself—whatever its form factor—will be an intelligent assistant helping you through your day. We will move from mobile first to an AI first world.

– Sundar Pichai, VD Google

One exciting case is Dartmouth-Hitchcock, an American health network that, with the help of Microsoft and Cortana Analytics (a part of Microsoft’s Azure cloud platform), has developed an application called Imagine Care, making it possible for patients to search for hospitals and for doctors to stay in touch with their patients after discharge, and to proactively warn them about health problems and recommend preventive actions before problems manifest. The solution includes monitoring patient blood pressure, measuring patient weight on the bathroom scale, and even analysing the patient’s voice during telephone calls with medical staff for indications as to their stress level and emotional state. Imagine what a boost it would be for any customer service organisation if it had a digital AI assistant that was able, in every support case, to help staff consider factors such as the customer’s emotional state, situation and previous dealings with the company – and which could even anticipate the customer’s behaviour and recommend an appropriate way of handling that customer. For example, what would be the best way to proceed to convert this disappointed customer in particular into a satisfied ambassador.

All of the Big Five – Google, Microsoft, Apple, Facebook, Amazon – are clear drivers of the latest AI trend. Several of these tech giants have even made AI technology available to the public and to other companies in order to accelerate the development cycle and their own learning process, and to otherwise reinforce their market position. A few examples within the machine learning and recognition field include IBM’s Watson, Google’s TensorFlow and Microsoft’s Cortana Analytics; examples in the chatbot field include Facebook’s Messenger chatbot platform and Microsoft’s Bot Framework.

AI First is a world that is still under development. It will not emerge overnight – challenges such as empathy, learning and contextual comprehension are not easily overcome. Indeed, artificial intelligence has been around as a field of research since the 1950s, and has seen an upsurge in media coverage several times – without ever seriously taking off. But this time, chances for success are better: significantly more computing power, better algorithms and platforms, and greater interest in investments by both companies and public authorities. So we can expect to see many new success stories going forward – but of course, many less successful stories as well. Such as Microsoft’s intelligent twitterbot Tay, which turned racist and was shut down just 16 hours after launch. But examples like these are not evidence of failure, just part of the learning process. Not just for Microsoft, but for our own businesses as well.

In order to remain competitive going forward, companies should begin staying abreast of new technologies within AI and what these technologies can do for them. They should also be finding relevant use cases and experimenting their way to success. It’s time to think beyond the omni-channel framework.

Further reading:

Image by Michele M. F under CC BY-SA 2.0