"You cannot not communicate" - Paul Watzlawik.
Impersonal communication takes place permanently: through gestures, facial expressions and the pitch of the voice. The clearest way of human communication, however, remains what is said, language.
Today, the communication between each other is no longer the classic face to face conversation, but via mobile devices. Through chats, messengers, FaceTime, etc., dialogue-based human-human communication has reached a new dimension - faster, more direct, omnipresent.
The first precursor of today's speech programs already existed in the 1960s, when Weizenbaum developed a speech program that simulated a psychotherapeutic session. Weizenbaum refuted his own thesis that communication between man and machine always remains superficial. The then found Eliza effect describes that people tend to attribute empathy and feelings to a machine as soon as it "independently" interacts with us. In the meantime there are similar dialogue systems on the market that are supported by Artificial Intelligence (AI) - called chatbots or language assistants. Among the leading providers on the market are Facebook Messenger Bot, Google Dialog Flow and IBM Watson.
These intelligent chatbots, which can be controlled via natural language, are entertainers, helpers and sources of information in a variety of situations. The simple, convenient and natural input of text via speech has become an equal and accepted alternative to a Graphic User Interface (GUI).
A Conversational Interface (CI) allows to transform the spoken or written into Natural-Language Processing (NLP), that means methods for the machine processing of natural language to understand the so-called intent (intention) and important entities (keywords, statements). The interaction takes place in Conversational User Interfaces (CUI). They are more social and inviting in communicating with the user by interacting via voice, chats or other natural language-oriented interfaces - as opposed to buttons, images, videos and menus in classic GUIs.
A CUI can be categorized into simple-bots, intelligent text-based assistants and language assistance systems such as Apple Siri, Amazon Alexa and Microsoft Cortana, in which commands and queries are sent to databases, categorized and analyzed and then made available to the user as high-quality answer.
LANGUAGE ASSISTANCE SYSTEMS:
Via voice recognition, they provide information about weather forecasts, upcoming appointments, home automation and music playing from streaming services.
A voice assistance system orients and focuses on what the user wants. As part of weak Artificial Intelligence, it has a task for which it is trained. A language assistance system (such as Google Home, Apple Siri and Microsoft Cortana) is activated with a signal word. NLP converts the spoken into text. Within a very short time, the request is analyzed (i.e. the intent is extracted), evaluated and an answer is generated or an action is executed. The aim is to personalize the interaction between man and machine. Operation is simple and usually smooth. The personalized user experience is the key value. It becomes unique.
Beside the possibility to operate such an assistance system via natural language, a CUI can also occur in form of a chat, more precisely a chatbot. Text input can be used to communicate with the system. These AI chatbots have text inputs and response masks. A program that uses strong Artificial Intelligence to constantly expand knowledge about the user and his needs - Machine Learning (ML). The neuronal nerve cell network of the human brain is simulated and connections are adapted and strengthened according to the situation. The "further learning" of the system takes place via information and programmed basis. It can be predicted whether rational results will occur again. The so-called Deep Learning, a branch of Machine Learning, enables the system to draw conclusions independently and is thus more granular and more advanced than ML. Answers are generated via a full-text analysis that accesses large knowledge databases.
To build a good Conversational Experience (CUX), a company first has to deal with what it wants to achieve.
Introducing a product or new service?
Generate leads or sell something?
The counterpart, the user expectation, i.e. the intention of communicating with a chatbot or language assistance system, is also an important aspect in the development of CUX in order to filter the relevant input and enrich the AI with the necessary information. Static information is enriched and modified with dynamic data by editing it. Dynamic information is integrated into the bot from databases, external resources or CRM (Customer Relationship Management) systems, usually via its API. The advantage is that this information is not limited - but there is a lot of complexity.
The way we communicate with machines will change significantly in the coming years. The Artificial Intelligence behind language assistants and chatbots can now even develop humor. Initial difficulties in speech recognition have subsided rapidly and are constantly being optimized. The parallel further development of clouds, apps, backend and mobile technologies have contributed to the fact that CUI (speech-based interfaces) with well-configured API service layers can be easily used.
The bots are already playing their part in consistent customer journeys with unique user experiences and have established themselves as a constant companion thanks to their flexible yet robust architecture. Users are categorized using socio-demographic variables and virtual scenarios to define the right approach. Smart homes, talking cars and more are no longer a rarity. CUIs are also accepted in the B2B sector, such as Amazon Alexa for Business or IBM Watson. Especially in service and support, the use of language-based interfaces is more efficient since the knowledge database is always and everywhere accessible. Nevertheless, the B2B sector, especially in the sales sector, is hesitant about CUI, at least at the moment. In order to conduct a meaningful conversation, the corresponding knowledge database is (still) missing. (Source: https://medium.com/@lswanson/why-b2b-needs-knowledge-bots-instead-of-chat-bots-9eb271c7f542)
The current legal situation with regard to the new Data Protection Regulation (DSGVO) in Germany is difficult. Questions like how to interpret the use and storage of interactions with AI is a main topic in upcoming discussions.
In the B2C area, however, Conversational User Interfaces can be seen as a milestone. They enable seamless communication and interaction between man and machine.