Is ChatBOT a synonym for great Customer Experience?
Boom… BOT revolution. Since Facebook has announced a massive use of artificial intelligence in its systems, people like me who lives in the iperdigital world only hear of this kind of technologies and their unlimited applications. At this stage the aim of Facebook is to put its customers in touch with brands that are present in the social network (or at least this is the first step of a project that may have a great scalability), but the BOT owns the potential for the new “APP”, available for developers free to create solutions as it happens in any online store.
Let’s try to better understand the BOT value and its potential. Nowadays the number of textual interactions is infinitely exceeding the voice ones and even the social ones. People communicate by writing messages. This is true also for the interactions between customers and brands, which today are at 35% digital. The keystone is to switch on the conversational level also the fruition of contents, which radically changes the relational model between brands of every type. Obviously it assumes the birth of new service models towards customers, which permit to provide information no more in the traditional way, but by processes in which the content is in the form of conversation.
So, what does a chatBOT consist of? By a technical point of view, a chatBOT is a program that has (or it aims at having…) a vocal or textual conversation depending on the context to which it is applied. It consists typically of an engine, sophisticated as desired, that rules the conversation and is linked to tools capable of interpreting the natural language (semantic engines). The more the conversational engine is sophisticated, the more it will be capable of sustaining complex conversations; the more the semantic component is accurate and committed to the context in which the engine is used, the more it will improve the level of comprehension of contents.
Of course, in the customer experience context the chatBOT arouse great interest, although through interpretations that are often extremely reductive compared to the ambition of the project. As stated above, the end point is to permit the fruition of contents through a conversational process. A process of this kind not only is contextualized to a chat, but to any “channel” utilized by user: chat, push notification, email, or voice conversation. Regarding inbound processes, it is quite clear that many conversational phases may be automated and delegated to chatBOT (for instance, the customer’s profiling and the individualization of his needs), as well as recurrent actions (document collection, acquisition of new customers, products comparison and support to frequent issues, and so on…). Concerning the outbound marketing, once the customer is clustered and engaged, it is possible to handle the levels of conversation and follow-up automatically.
The quality of a conversational solution is valued generally according to the customer’s journey mapping and by its capacity of interpretation. In the customer engagement segment I consider to be fundamental to investigate also how to handle with the moment in which the artificial intelligence has to pass the turn and engage a human hand. At this step, the most of engines fail, as they do not guarantee the “continuity” of experience and context expected by a customer. We are often used to change the way of interaction by shifting from an interface to another, or to talk to brand representatives little or not at all contextualized in the conversation started by BOT. It leads to frustration that culminates in the “bad customer experience” and brings, at 69% of cases, to the change of dealer/provider/bank. The other significant issue linked to this “momentum” is the mapping. Systems that are little or not at all integrated supply heterogenous data, that do not permit to measure correctly the efficiency of the platforms and to detect optimisations.
Let’s try to imagine where a system of this type could be set reasonably quickly.First of all, the e-commerce world would be revolutionised.The linking of a chatBOT together with a system of payment in APP, would permit to contextualize a search, to explore the contents and to conclude the purchase in the application of the retailer, or even inside the container of others, for instance messenger (also because WhatsApp, the would-be first choice, at present do not permit third party procedures to create processes like this). Only in few cases in this context it would be necessary a human assistant, completely bypassing the normal process of web purchase, by using a totally innovative UX in accordance to the customer’s expectations.
On the other hand, let’s think of a new context completely different as the banking one, and assume a new mortgage online request. Maybe by a new prospect. At first, the customer starts an open dialogue with the BOT to comprehend the needs and to define the suitable product. Then, if the customer (who can be pressed with outbound automated processes, more or less intrusive) chooses to investigate further, he collects the necessary documentation, performs an onboarding process to open the account and defines the loan conditions. So far, everything is done through BOT. At last, it could be necessary to involve a consultant who verifies the dossier and agrees the details with the customer. The whole thing is obtained by using always a single conversation process based on push notifications and mobile messaging.
The processes described above aren’t anything too complex from the technologic point of view, and they aren’t even difficult to understand by the customer. The UX furnished is more simple and intuitive than can be imagined, and ensure economic savings and demonstrable improvements in the customer satisfaction.
The end point towards it aims is traced, an unified access for the customers of a brand that allows to get into the contents, to pay or search for a service, through a simple conversation in a natural language. Having regard to the time in which technologies use to spread, we would not expect 10 years before we will be surrounded by services of this kind.
Our Solution: www.eudata.com