For many, artificial intelligence (AI) still conjures up images of science fiction movies and robots becoming too intelligent for their own good and eventually deciding to take over the world because the human race is no longer up to the task.
But I have found that a useful starting point when discussing AI is to think of it as you would a child. The AI is a learning construct, in the beginning it doesn’t know anything, much like a child. So therefore it is not its fault if it does something ‘wrong’ – at this point, it simply doesn’t know any better. But it will learn.
In the real world we are at the beginning of the AI journey. AI is present in some devices and services we already use, but only at a basic level – think of Siri on your iPhone or other digital assistants like Amazon’s Alexa.
These AIs use natural language processing (NLP) techniques to adapt to how we speak to them and enable them to respond to questions. They also have a degree of machine learning built into them to help them continuously improve their level of service. But in reality they are not much more complicated than interactive voice response (IVR) services, which essentially are glorified decision trees.
Evolution of chatbots
However, where AI tools are gaining ground is by utilizing context. For example, chatbots can use context to learn about the user and personalize the interaction. Chatbots act as digital touchpoints for customers with supplier. They are conversational interfaces that have evolved from simple messenger-based interactions to much smarter, enhanced apps. AI has enabled chatbots to learn, to employ context and use NLP to give end-users a much more satisfying and complete experience.
Essentially, the chatbot is the front end that engages with the customer. The simple fact is that the customer does not really care what is going on behind the scenes – only that they get the service they want and the answers they need. To create a more memorable customer service experience however, many organizations have now started designing ‘personality’ into their chatbots, that reflect the company’s brand image.
People will interact better with chatbots that they can identify and associate with, so it is logical that companies will tailor chatbot personas this way. Facebook CEO Mark Zuckerberg has done something similar already, building his own personality-focused AI using perhaps the world’s most recognizable and charming voice – Morgan Freeman.
All about the context
As AI chatbots continue to evolve, it is context that will continue to provide the difference and make them more familiar for users. As human beings we are blessed with senses that inform us of things. For example, if our eyes sense light, we can feel heat and our body gets warm, it could be that we are standing in direct sunlight. But it could also be that our house is on fire. As humans we have the contextual understanding to recognize the difference and react accordingly. Bots do not yet have that degree of comprehension, but AI will help them learn and become more sophisticated.
Learning, evolving and adapting
At Orange we are interested in how bots will use AI to adapt to users’ personalities. We are currently running tests on people working with bots and are modelling our bots based on that intelligence – which is actual behavioral intelligence rather than business intelligence in this instance. We have found that users find bots more efficient when the bots speak to them in a language to which they can relate. As in the Zuckerberg video above, it is effectively about having an automated PA who remembers absolutely everything you need them to but who, at present, is still learning about you and how to behave towards you – which of course has echoes of our child analogy. In the words of Orange CEO Stéphane Richard when unveiling the new virtual assistant Djingo, “By developing a deep relationship with [Djingo] the interaction will get richer and richer. That is the essence of AI”.
So in the end it will be about growth – the growth of smartness in the AI chatbot and the growth of trust from the user. Demystification of bots will happen and users will in time come to trust bots like they trust people – and it is perhaps important to remember to trust the supplier of the bot too. This trust will help you get things done.
Orange can help you convey your company culture and reach out to your end-users in a way that will match their usage. Find out more about our work in customer experience at Orange Business.