The birth and subsequent adoption of a technology often observes a curve known as the “Hype Cycle”, as defined by Gartner. Conversational technologies such as virtual assistants and chatbots, especially those using Artificial Intelligence components such as Natural Language Processing, are perfect examples of this cycle.
The reappearance of chatbots in the middle of the past decade gave rise to oversized expectations that inevitably led to a widespread disappointment with conversational technology, a phase known as the trough of disillusionment. Currently, the market has managed to get out of the trough and face its next phase, the slope of enlightenment. This stage is based on a greater knowledge of the technology, its advantages and its limitations. Thanks to past experiences, we can now prevent inflated expectations and we better know now how to approach the implementation of new conversational solutions with caution. Likewise, the chatbots themselves are also strengthened from previous phases and they now evolve in an era of technological maturity.
Over the past years all suppliers of AI-based solutions for customer service have witnessed the peak of chatbots, as well as the growth of expectations regarding conversational interfaces. The self-service eco-system has transitioned from virtual assistants based on FAQ’s or article repositories, through to NLP-powered IVRs and integration in messaging Apps, until reaching the performance of several complex transactional processes within one unique chatbot. This means that nowadays, chatbots not only provide informative contents which are not limited to redirecting the user to the section of the web or app where they can effectively carry out their operations. Chatbots can achieve much more.
Today’s Chatbots are able to receive user information from a CRM, provided that the user has previously accessed his private area. This ability allows you to personalize your customer care for example, addressing the user by name, referring to the products he/she has subscribed to, or resuming a pending issue. Chatbots also know how to collect specific information during the conversation or pro-actively request it if necessary. Thanks to the management of these variables or parameters, the conversation is always contextualized and statistics can be filtered according to variables. The user does not have to repeat information he already communicated (either implicitly or explicitly) which in turn can make them feel personally taken care of during a quick and efficient conversation. The variables are the basis for the Chatbot to change from being a mere explanatory communication channel into becoming an integrated transactional application.
Naturally the link between the chatbot and the internal applications has to be made in accordance with the security policies and avoiding that a failure in a transaction could affect the functioning of the Chatbot in general. The in-between element, which is responsible for both security and isolating each process from another, is usually a webhook. It is a piece of code that acts both as a translator and a filter when transferring information between the Chatbot and the internal system, and vice versa. It can include different layers of additional security, depending on the data’s confidentiality requirements. With regards to security, we must underline that the Chatbot should not manipulate any sensitive information but only forward it and/or show it in the user interface. Besides, some technology suppliers include in their platform a variables and logs obfuscator tool in order to guarantee the data privacy of the clients and their end-users.
Currently, use cases that include transactional processes are as diversified as the needs of companies are: booking a meeting room, locating the nearest ATM, recovering a password, tracking an order, checking an account balance, confirming the status of a flight, requesting a duplicate of an invoice and so on. They all have in common simplifying the access to operations that until then, could only be managed from the Contact Center and through a human agent. In addition, some of these transactions are linked to each other or integrated into decision trees, which grants the conversation with the chatbot even a greater fluency and briskness.
The benefits of a transactional chatbot are obvious: on one hand, users have a simple conversational and personalized application available, which can also be integrated into their favorite contact channels (Whatsapp, Facebook Messenger, native app, web, etc). On the other hand, companies make the most of their already digitized processes, while relieving their customer service departments of the most frequent and repetitive operations.