A complete guide to chatbots: past, future, the perspectives

A complete guide to chatbots: past, future, the perspectives

A chatbot is a digital dialogue system that has natural language skills. Such a bot can recognize text and voice messages and react to them independently.

A flashback to the history

The virtual psychotherapist Eliza is historically the first chatbot. It was developed between 1964 and 1966 by the German-American computer scientist Joseph Weizenbaum.

Chatbots experienced their first major hype in the late 1990s. It is now impossible to imagine the spectrum of digital communication forms without them. For example, chatbots are used on websites or in instant messaging systems to advertise products and services, answer customer questions, or take orders. Other bots welcome new visitors to a website, take care of concerns independently or forward them to a human customer service representative. From a technological point of view, speech recognition software such as Siri and Alexa is also a bot.

Chatbots - speech recognition, deep learning, and AI

A chatbot examines textual or linguistic inputs from users and converts them into appropriate reactions using rules and routines set by those programming them.

The first bots were only able to identify individual keywords and carry out associated actions. Even today some chatbots are structured by simple questions and answers. However, the current trend is towards bots as self-learning systems based on Artificial Intelligence (AI), Natural Language Processing (NLP), and Deep Learning. Such a bot can understand context-dependent meanings and derive complex interactions from them - it constantly learns from the feedback of the users.

The technical dialogue system bundles all information in a single channel and uses it to process customer inquiries. In particular, a chatbot takes on standardized and/or repetitive tasks in marketing, sales, and customer service, thereby relieving employees.

All communication processes are stored permanently, securely, and anonymously following the General Data Protection Regulation (GDPR) that is used for further analysis. Companies, therefore, benefit from chatbots in particular because they increase the efficiency of processes. Thanks to the intelligent software, the dialogue system learns to be more independent day by day.

What types of chatbots are there?

Not all of them are similar. To answer the question "What is a chatbot?" taking a look at the possible variants of this technology helps to give the correct answer. Even small technical changes can make a big difference here. 

You can differentiate the following types of chatbots:

• Rule-based

• Intelligent

• Application-specific chatbots

• Rules-based chatbots

Rule-based chatbots talk to the customer with predefined answer options. However, if the user asks a question that was not considered in advance during programming, the chatbot won’t be able to understand you and therefore, to answer. Thus, a rule-based digital assistant is only recommended for very simple and standardized processes. They are suitable for smaller marketing campaigns.

However, new digital technologies such as artificial intelligence and machine learning have produced new methods that significantly expand the function of rule-based chatbots and that’s why now they could be related to the intelligent chatbots.

The pros and cons of every type

The pros and cons of rule-based chatbots


  1. Ideal for small marketing campaigns and simple tasks (e.g. filling out forms, answering simple questions);
  2. Once developed, could run on all platforms;
  3. Suitable for short dialogues (e.g. ordering a pizza);
  4. Available around the clock.


  1. Initial programming causes high acquisition costs;
  2. Cannot be used immediately;
  3. Longer development time;
  4. Little automation possible;

The advantages and disadvantages of AI-based chatbots


  1. They can understand complex sentences and lead dialogues;
  2. They always keep track of the conversation and context;
  3. Ideal for an optimal user experience;
  4. After development, they can be used on all platforms;
  5. High savings potential through better efficiency in the processes;
  6. You can build your database and they will always keep learning;
  7. They deliver data in real-time;
  8. High potential for generating insights;
  9. They are available around the clock and answer customer inquiries effectively and quickly;
  10. Inexpensive in the long run.


  1. Initial programming causes high acquisition costs;
  2. Cannot be used immediately;
  3. Longer development time.

Application-specific chatbots

Application-specific chatbots mix rule-based with intelligent dialogues being enhanced with graphic interfaces.  An example of these bots is calendar graphics that lead to an appointment or a booking within a click. The customer does not have to enter the date manually and saves time.

The pros and cons of application-specific chatbots


  1. Ideal for online banking, consulting applications, insurance, hotels or restaurants;
  2. They offer the user more convenience in using websites or forms.


  1. The system needs a long period before it can be intuitively operated;
  2. The high-cost factor in development and maintenance;
  3. Every device and platform requires new programming.

How chatbots work

It works like a technical dialogue system that can be used for communication via text or voice input. The first chatbots were purely text-based. However, with the increasing development of speech synthesis/ speech recognition, communication with many of them can take place with a mix of text and language or purely with speech. Chatbots are integrated directly into the messenger - providers such as Facebook Messenger or Telegram already offer likewise interfaces.

On the technical side, a chatbot can be imagined as an if-then principle. To process and answer a request, chatbots use knowledge bases and recognition patterns for the respective questions and answers. That means they have a defined set of keywords, questions, answers, and topics. You send a message to a bot, which it then processes. Processing is based on the if-then principle. That means: You define beforehand which answers your chatbot should give to which question. An example of this principle: If a user writes “Show me your location”, the bot has to show the directions.

Chatbots – the hype or the assistance which will only emerge in the future?

The huge hype about chatbots is over. And that is a good sign for the success of the technology: It is increasingly used in customer service and is revolutionizing the customer experience there. Compared to apps, websites, or traditional call centers, conversational AI is proving to be more and more beneficial for both customers and companies - and is thus approaching the mainstream. But what will happen in the coming years?

The market research company Gartner estimates that by 2020 around 85% of customer contacts will be done without human contact, i.e. with self-learning chatbots.

For this vision to become reality, the developers of chatbots now have to overcome a critical point: Many technological innovations fail precisely at this point, where the aim is to transfer technology from an innovation-driven niche into the mainstream.

Fired by previous successes, chatbot makers are pushing the technology forward more than ever.

Conversational AI is becoming more and more popular in customer service: On the customer side, technology has the potential to continuously improve service. For the company, on the other hand, it reduces the costs that are expended for processing customer concerns. Chatbots, therefore, offer considerable advantages over websites, apps, and traditional call centers, which companies can adopt using various approaches.

For small companies, the improved service through the use of Conversational AI and the associated increased user loyalty play a decisive role: for example, they can raise the conversion rate.

For large companies, the implementation of chatbots in customer service, on the other hand, promises considerable cost minimization.

Chatbots also offer great potential in the direct-to-consumer sector –users benefit from enormously increased speed. Processing a wide variety of issues takes an average of three to eight seconds.

Chatbots serve as an interface between information, people, and machines. To do this, they can be integrated into various infrastructures, including today's social networks such as Facebook or Twitter and mobile apps. Specialized bots take on, for example, the administration of appointments, carry out navigation tasks, personalize online research and online shopping, organize the ordering of an Uber taxi, or fulfill various business functions. Virtual assistants such as Siri, Alexa & Co., on the other hand, are intended to fulfill the widest possible range of tasks and potentially serve all of their users' needs.

Although chatbots still have great development potential in terms of technology, numerous technology groups - including Apple, Google, Amazon, and Facebook - now consider their development to be an important strategic business area. Many applications that have been taken over by apps will likely be executed by chatbots in the foreseeable future.