2. /
  3. Innovations
  4. /
  5. Chatbots und Conversational User


How computers learn to understand us

A chatbot, which is short for chatterbot, is a program for communication activities. These computer programs imitate written or spoken human language and thus simulate a conversation with a real person. They can be integrated into existing system landscapes as mobile applications without any extensive installation requirements.

The acceptance levels and proliferation of chatbots – as a form of conversational user interface (CUI) – have been boosted due to the rise of messenger apps. Six of the top ten apps used around the world are now instant messenger or chat programs like Facebook Messenger, Slack, Telegram, text messages, etc. Further examples of CUIs include digital assistants, which really are "on the tip of everyone's tongue". Known models include Alexa, Cortana, Google Assistant, Siri and Watson.

A central element of the chatbot technology is natural language processing (NLP). This part of computer linguistics deals with the extraction of relevant information and intention from requests. The goal here is to make written or spoken user input understandable on the machine side.
For the purpose of identification, chatbots generally make use of a knowledge base with answers. During the next step the question is interpreted. This takes place using semantic recognition patterns, which are used to form the so-called intent. A chatbot identifies the intention of the person asking the question and continually learns to generate more appropriate content when producing output. An increasing number of companies in the automotive and banking sectors are now using chatbots. The goal of this is to increase customer satisfaction, on the one hand, but also to be able to offer additional services from providers in different sectors, for example, the IoT industry.

However, at the end of the day technical progress often relates to making things simpler and more "time-saving". Chatbots and other forms of CUIs provide the possibility of limiting "App Overload".

[Translate to Englisch:]

Ein Chatbot, kurz für Chatterbot, ist ein Programm für Kommunikationsaufgaben. Diese Computerprogramme imitieren geschriebene oder gesprochene menschliche Sprache und simulieren so eine Konversation mit einer realen Person. Ohne hohen Installationsaufwand fügen sie sich als mobile Anwendungen leicht in die bestehende Systemlandschaft ein.

Die Akzeptanz und Verbreitung von Chatbots – als eine Form von Conversational User Interface (CUI) – wurde in den letzten Jahren durch den Aufstieg der Messenger-Apps begünstigt. Sechs der Top 10 weltweit genutzten Apps sind mittlerweile Instant Messenger oder Chatprogramme, wie Facebook Messenger, Slack, Telegram, Text Messages, etc. Weitere Vertreter von CUIs sind digitale Assistenten, die ebenfalls sprichwörtlich in aller Munde sind. Bekannte Modelle sind beispielsweise Alexa, Cortana, Google Assistant, Siri oder auch Watson.

Ein zentrales Element der Chatbot-Technologie ist das Natural Language Processing (NLP). Dieser Teilbereich der Computerlinguistik beschäftigt sich mit der Extrahierung relevanter Informationen und Intention von Anfragen. Ziel ist es, menschliche schriftliche oder gesprochene Benutzereingaben maschinell verständlich zu machen.
Zur Identifikation greifen Chatbots üblicherweise auf eine Wissensbasis mit Antworten zurück. Im weiteren Schritt erfolgt die Interpretation der Frage. Dies geschieht über semantische Erkennungsmuster, woraus sich der sogenannte Intent formt. Ein Chatbot erkennt die Intention des Fragenden und lernt immer passendere Inhalte für eine Ausgabe zu generieren.

Immer mehr Unternehmen der Automotive Branche oder auch im Bankenumfeld setzen auf den Einsatz von Chatbots. Dabei ist die Erhöhung der Kundenzufriedenheit das Ziel sowie die Möglichkeit, zusätzliche Services branchenübergreifender Anbieter bereitzustellen, beispielsweise der IoT-Industrie.

Am Ende des Tages dreht es sich beim technischen Fortschritt darum, die Dinge einfacher und zeitsparender zu gestalten. Chatbots und weitere Formen von Conversational User Interfaces bieten die Möglichkeit den „App Overload“ einzudämmen.

Chatbots are often the first port of call when it comes to the dialogue between a company and its customers, or between employees and company data. Possible use cases include understanding and sorting incoming inquiries from users, and then answering these automatically. A significant benefit of machine-supported communication is the saving of time. Chatbots provide the person asking the question with feedback in real time for most routine inquiries.

Our approach and expertise

One significant challenge regarding the hyped-up chatbot technology is to make it an even more intelligent communications partner for an ever-increasing range of topics. CUIs like Alexa or Cortana are already quite successful in this regard. However, their area of application is still heavily dependent on the subject matter and the quality of the database.

Our goal is to make conversations with our digital helpers even more natural in future. When a chatbot is unable to process user input, a standardised reply is usually given. EXXETA wants to remove "I didn't understand you" from the repertoire of chatbots. For this reason we are also looking into the question of how "intents" can be generated automatically from existing records and further processed. If these intents – drawers for storing the intentions of the users – are maintained more comprehensively, the conversation between CUI and human will become more natural.

This innovative approach is based on the analysis and semantic bundling of unrecognised intentions. A series of questions, which produce such an error message, would be bundled together in relation to the content and then used to generate recognition patterns. At the same time, existing records that the chatbot should use should include valuable information for ensuring optimum intent structure. This would then be used to generate new intents automatically, without the need for an additional editorial outlay or information. Each conversation with a human would make a chatbot designed in this way more intelligent while also enabling new topics to be handled. In future, natural conversations should become the central interface between human-machine interaction. And EXXETA is already working on many projects with the vision of making this a reality.


Rich Com­mu­ni­ca­ti­on Ser­vice (RCS):
the future of SMS

EX­XE­TA is part of Google's Rich Com­mu­ni­ca­ti­on Ser­vice (RCS) Business Messaging Program. This gi­ves us ear­ly in­sight into this in­no­va­ti­ve tech­no­lo­gy. In this way, we can pre­sent this new chan­nel for cust­o­m­er com­mu­ni­ca­ti­on to you and de­ve­lop it to­ge­ther with you for your tar­get groups.

With RCS, you can ex­tend your SMS com­mu­ni­ca­ti­on with your cust­o­m­ers and in­clu­de rich con­tent for­mats such as quick an­s­wers, vi­de­os, dy­na­mic con­tent, etc. You can reach your cust­o­m­ers with new mar­ke­ting stra­te­gies di­rect­ly on their smart­pho­nes wi­thout ha­ving them to in­stall ad­di­tio­nal apps. In ad­di­ti­on, RCS tech­no­lo­gy al­lows you to build an au­to­ma­ted con­ver­sa­tio­nal user ex­pe­ri­ence. With the help of deep links, you can down­load Com­pa­ni­on Apps and ac­cess con­tent.

Get in touch

EXXETA AG - Sascha Vollmer

Sascha Vollmer


+49 1522 2619555