In the context of customer service, automation was once a byword for inefficiency, synonymous with impersonal, laboured transactions, riddled with errors and delays.
Think of the self-service tills in supermarkets that can still wreak havoc and the soulless phone exchanges that greeted you when trying to sort out a discrepancy with your bank. Left on hold and directed to endless menu options, all too often the reality of the service experience was the antithesis of the improved speed and efficiency the technology was designed to deliver.
We know that the concept isn’t going anywhere, it just needed to get smarter and it has. According to Gartner, by 2020, 85% of customer relationships will be managed in the enterprise without any human interaction at all.
Thanks to the explosion in mobile technology, consumers are more digitally empowered than ever, and more demanding of higher quality automated interaction as a result. The industry has responded accordingly with boosted capabilities that have driven efficiencies and improved the customer experience, exemplified by traction of robotic process automation in banking.
Transformed by artificial intelligence (AI) and machine learning, automation has become intrinsic to the customer experience, rather than being perceived purely as the cost cutting poor relation to human service. Uniquely primed to simplify the user interface through its ability to recognize and emulate human speech, AI has evolved the more rudimentary avatars into dynamic and multi-dimensional chatbots, with far deeper understanding of the consumer’s status, objectives, and nuances of personality.
As they integrate with apps such as Facebook Messenger, Kik, WhatsApp and Slack and understand what customers are saying in real-time, chatbots are increasingly primed to deliver the personalized and intuitive experience needed to pique the interest of the time-poor, but digitally savvy, customers. Through more relevant and contextual content, their ascendancy proves that automation and personalization need not be mutually exclusive.
With businesses facing heightened pressure to ensure customer interaction is seamless and missing any kind of friction, the reliance on chat-based interfaces is snowballing, set to account for almost 40% of transactions in the call centre by 2020, if industry predictions bear out.
Their evolution is particularly evident with the next generation call automation systems. Fuelled by more sophisticated cognitive capabilities, these customer interactive voice response systems can deliver advanced conversational interaction by detecting the caller’s intent and by using natural language for a more authentic dialogue, while in some cases, even predicting the user’s responses to which are then processed accordingly.
In a similar vein, companies are harnessing text messaging bots that can handle common questions faster than waiting for a human advisor to become free as another way of driving efficiencies, while introducing more conversational and natural interaction into the customer experience. Soon texting your bank to get an account balance update, for example, will become the norm.
Here, we see how AI can augment the skills and capabilities of people, as this blend of human operators with automated bots becomes a preferred choice for businesses wanting to retain the human touch and soft skills in conjunction with the technological prowess.
Interestingly though, we’re not far off seeing bots equipped to do it all, now able to deliver on the soft skills as well. On the cusp of of being equipped with emotional intelligence, chatbots being able to recognise a customer’s emotional state and respond to the feelings appropriately during the conversation, is a development that will signal the next major disruptive phase in business communication. This next interface level will therefore render human intervention redundant in certain service capacities, or at least obsolete from a certain level in the service chain.
It’s clear that potential is rife, but ultimately reaping benefits boils down to the amount of data you’re able to collect and use, and how it is handled. Crucially, we know that AI - for all the opportunities that it presents - isn’t infallible. The unrestrained nature of many of the systems around robotic process automation, for example, demands rules and structure if they are to fully thrive without becoming a loose cannon. Tools and solutions that can read metadata and visualise the output of machine learning systems to interpret anomalies and create rules in an accessible and easy to understand way, have never been more important.