Survival of the fittest: Battle of the bots ensues
4 min read
12 January 2017
From Facebook’s launch of enterprise chatbots on its Messenger app, to Google and Amazon introducing virtual assistants into the home, 2016 has given rise to the battle of the bots.
Industry experts believe the battle of the bots has grown intense – and that the technology itself is here to stay. Gartner also predicted that by 2019 some 20 per cent of brands will abandon mobile apps, while the average person will have more conversations with chatbots than with their spouse by 2020.
But, not all bots are created, nor do they perform, equally. We predict that in 2017 this will go one step further with the battle of the bots, where the “good” bots will set best practices, while the “bad” bots will fail.
So what are the critical success factors of a “good” bot? It all comes down to customer experience. This sentiment is echoed by Forrester Research: “Today’s customers reward or punish companies based on a single experience – a single moment in time. This behaviour was once a Millennial trademark, but it’s now in play for older generations. It has become normal.”
So what does a bot need in order to survive this battle of the bots? We believe it takes:
1) Conversational artificial intelligence
A successful bot must be capable of holding an intelligent, two-way conversation. It must be able to understand, like a human, the context of the conversation as the user changes the subject, or when they use colloquialisms. Most bots are not currently sophisticated enough to do this. For instance, it may be able to respond to the question “what temperature is it in London?”, but is unable to maintain the context of the conversation to respond to a secondary question of “what about in Bristol?”.
2) Cognitive artificial intelligence
While traditional speech recognition systems understand what people say, it is one step further for the bot’s reasoning to take action or even predict the needs of a consumer. Natural language systems are now far more sophisticated, helping bots understand what people mean and want to do.
3) Human assisted artificial intelligence
This is how Nuance refers to supervised AI. Using human customer service agents as bots “partners” accelerates machine learning and, most importantly, ensures that bots are learning “the right things” from humans. Microsoft’s Twitter chabot earlier this year demonstrated how bots can learn negative tendencies from the wrong engagement.
4) Omnichannel integration
A successful bot won’t be a standalone application, but rather common tools that act as a central cognitive brain and are deployed across a number of channels – messaging applications, phone systems, mobile applications, web searches, chat applications and on social media.
Taking an integrated, omni-channel approach will ensure a consistent experience for customers, regardless of which channel they use. This also reduces the cost of siloed technology stacks for brands.
5) Intelligent authentication and security
Integrating biometrics security – whether that be voice or selfie recognition – provides an easy and natural way for consumers to authenticate their identity without the need for pins or passwords. For instance, by speaking a passphrase like “My voice is my password”, consumers can be confident that their account is secure without the need to remember pins or the responses to a series of security questions, from their mother’s maiden name to their most recent transaction. And beyond the customer experience, biometric security also improves security over legacy authentication methods and helps reduce fraud.
Sebastien Reeve is director of product management at Nuance Communications