By Mark Troester, VP of Strategy at Progress
Most of us can recall more than one conversation with a chatbot that has gone very wrong. These conversations are typically short, fruitless and tend to end abruptly with the caller being either deferred to a human customer agent or with the interaction finishing all together.
The concept of chatbots as customer service agents is definitely built on the right premises. In this day and age, a customer expects that a number of enquiries can – at least to some degree – be dealt with in a ‘smarter’ and more automated way. What has been holding this shift in customer experience and engagement back is the limitations in the way chatbots have been programmed and subsequently in what they can do. Thanks to Artificial Intelligence (AI) and cognitive learning, this is already changing fast with the rise of a new generation of chatbots that are able to handle complex tasks and deliver a personalised experience to customers.
Chatbot 1.0 didn’t quite cut it
In the past few years, chatbot technology has seen rapid development and innovation. Many things have changed in the ‘chatbot world’ since the launch of Eliza in mid-1960s which took a stab at mimicking human conversation by matching user prompts to scripted responses.
The reason behind this enthusiasm for the benefits of chatbots for customer engagement is that chatbots have emerged as a great option for providing a 24/7 self-service solution to address a host of customer support requirements. Customers can get simple queries answered in real time which means that support staff can focus on more complex tasks instead of having to field high volumes of repetitive inquiries. In theory, this sounds like a good way to save time and resources, for customer and business alike. In practice, however, customer conversations with chatbots often hit a brick wall very soon after they start.
This is mainly because of the fundamental problems in the way traditional chatbots have been designed from scratch. Up to now, the intelligence of these chatbots has been ‘menu driven’ and, therefore, limited by predefined decision trees that allows chatbots to understand and be able to perform a limited set of actions. This leads chatbots to pretty much behave as answering machines that eventually source calls to human call agents or even worse that fail to understand what the customers are actually looking for.
This last case is what has been causing frustration amongst customers. Having to ask the same question over and over as the chatbot ‘doesn’t get it’, together with an overall inability to take action or progress with the conversation beyond simply logging in data, leads to disappointment and the discouragement of customers from wanting to engage with a chatbot. The end result is for these interactions to escalate to human engagement.
Cognitive chatbots are here to stay
It is clear from the above that, as trust and accuracy are at the core of every customer interaction with businesses, clunky pieces of self-service software that still require customers to look for customer support agents can lead to reputational damage.
The solution is a cognitive bot that can be deployed to a customer’s channel of choice – website, social or mobile. Cognitive chatbots are much more autonomous and transactional. This technology, drawing on AI and machine learning, enables chatbots to be trained to learn from each conversation, meaning that the more they are used, the smarter they become. Cognitive algorithms allow for a conversational user interface (UI), allowing the chatbot to move beyond answering questions to actually conducting transactions, whilst offering optimised and seamless customer experience.
To do this, chatbots need to be equipped with the intelligence to prepopulate and understand elements like the customer’s name, account number, prior inquiries and requests so the conversation does not have to start from zero every time the customer reaches out. For this to be achieved, the chatbot needs to be integrated with backend systems, something that so far has been a complex and costly operation. However, severless backend-as-a-service platforms have simplified these processes and spared organisations the need to write server-side logic.
‘Smart’ customers want ‘smart’ chatbots
In a time when innovation is expected to run in the bloodstream of every organisation, customer service is a crucial touchpoint with the customer to showcase the latest technologies. Whilst existing chatbots have been a step in the right direction, it is time that businesses took the next step and made the most of what cognitive chatbots can offer. Cognitive bots that are fast and easy to deploy can be a real game changer and an invaluable brand ambassador to customers.