What are chatbots and how do they work?

By Published On: November 23rd, 2023Categories: chatbots

In recent years, the term “chatbot” has become increasingly common due to the growing interest and public attention towards artificial intelligence and automated systems. With the hype surrounding ChatGPT, many people are now familiar with the term, but what does it actually mean?

Essentially, a chatbot is an artificial intelligence or computer program that is capable of conducting human-like conversations and answering inquiries in a natural manner. Chatbots can be utilized in various industries and applications, such as customer service, e-commerce, healthcare, and education.

Today, chatbots are also being used for other tasks such as internal communication within a company or automating business processes. The popularity of chatbots is constantly increasing, as companies use them as an effective tool to improve customer communication and experience.

Some companies also use chatbots for marketing and sales activities to generate leads and build customer loyalty.

What types of chatbots are there?

There are various types of chatbots, including rule-based chatbots that are based on predefined rules and scripts, and AI-powered chatbots that can learn and improve through the use of machine learning and natural language processing.


LLM chatbots

LLM chatbots, or Large Language Model chatbots, are advanced artificial intelligence systems that use large-scale language models, such as GPT-3, to understand and generate human-like text in conversations. These chatbots leverage extensive pre-training on diverse datasets to comprehend and respond to a wide range of user inputs. With their ability to grasp context, infer meaning, and generate coherent and contextually relevant responses, LLM chatbots are capable of engaging in more natural and dynamic conversations, making them valuable tools for various applications, including customer support, information retrieval, and interactive user interfaces.

Rule-based chatbots

These chatbots are based on predefined rules and scripts and can only respond to specific questions or requests. However, they can quickly and effectively handle simple requests, such as answering frequently asked questions or checking order status.

AI-powered chatbots

These chatbots use machine learning and natural language processing to continuously learn from experiences and feedback, and adjust their responses and actions accordingly. They can also conduct human-like conversations and solve complex requests or issues.

Hybrid chatbots

These chatbots combine rule-based and AI-powered approaches to offer the best of both worlds. They can quickly respond to simple requests while also providing human support when a complex question or issue arises.

Voice-based chatbots

These chatbots use speech recognition technology and text-to-speech software to enable communication with users through voice. They can be deployed on various devices such as smartphones, smart speakers, or computer systems.

Messaging chatbots

These chatbots operate through messaging platforms such as Facebook Messenger, WhatsApp, or Slack. They can be directly addressed by users and typically respond quickly and effectively.

Intelligent assistants

These chatbots are usually designed to perform a variety of tasks and provide a personalized experience for the user. For example, they can help plan daily schedules, create shopping lists, find information, or even assist the user in navigating through a website or application.

Artificial Intelligence (AI)

Thanks to advances in Artificial Intelligence, it has become very easy for humans to interact with automated systems or conversational interfaces such as chatbots or voice assistants. These systems are designed to communicate with users in a natural way and effectively and accurately process their requests. AI-based chatbots and voice assistants are continually being developed to better meet the needs and requirements of users.

According to MarketsAndMarkets, the global market for conversational AI is expected to grow from $6.8 billion in 2021 to $18.4 billion in 2026, representing a CAGR of 21.8%. This is due to the increasing adoption of conversational AI technologies in various industries, such as customer service, healthcare, and retail.

Furthermore, the use of chatbots has increased by 67% between 2018 and 2020. This increase is due to the growing automation of customer service and sales processes, where chatbots play an increasingly important role. Chatbots are also used in other areas such as recruitment and e-learning to improve efficiency and user experience.

Der Hype um Chatbots


The hype around chatbots gained momentum in 2016, when messaging apps had more active monthly users than social networks for the first time.

Today, chatbots play a significant role in business automation processes.

If you’ve missed the train so far and want to jump on board, you first need to understand what chatbots are and how they work.

In addition, you probably have, like many of our clients, some questions about chatbot implementation, technology, etc., which we will also address.

“Chatbots and voice assistants are more than a new tool. They are about a full-scale transformation process that will permanently change the way we gather information and manage knowledge.”

alexAlexander Wahler, CEO Onlim
Onlim Logo

Onlim has been active in the chatbot industry since 2016 and has now implemented several hundred successful chatbot projects for medium and large companies from various industries.

We are happy to share our experiences with you in article The Best Chatbot Use Cases From Different Industries.

What are chatbots exactly?

Chatbots are a form of automated service, essentially a robot with which customers can communicate via a chat interface. They rely on Artificial Intelligence (AI), which operates through a set of defined rules and parameters.

Essentially, you engage in a virtual conversation with a robot that can perform certain tasks for you or answer your questions.

How does a chatbot work?

Chatbot Customer-Service

A chatbot functions similarly to a customer service representative. When a customer starts a chat for help, a AI agent responds. For example, the inquiry could be: “What time does your store open tomorrow morning?”

The bot then responds based on the information available to it: “Our store opens at 9am tomorrow and closes at 5pm.”

Nowadays, chatbots can also handle complex inquiries such as:

“My printer isn’t working. I’m getting error code 505. What can I do?”

“I’m looking for a room for 2 adults with my dog for Easter weekend in Neusiedl.”

“How much does a day ski pass cost for adults with a child?”

“I’m looking for a blue carbon mountain bike for women for less than €3,000.”

Have you ever asked Siri or Alexa about the weather forecast in your area? They are also a type of bot, namely voice assistants, and can give a natural-sounding answer to us humans.

What happens in the invisible area for us?

questions and answers

To answer this question, we first need to divide chatbots into two categories.

As mentioned above, all chatbots are a form of artificial intelligence and complex programming. However, there are two types that differ primarily in their essential component. This can be either artificial intelligence and machine learning or structured questions and answers.

Structured by questions/answers.

Chatbots based on structured questions and answers are somewhat less complex than those that use machine learning to fully harness the power of artificial intelligence.

The visible chat interface looks very similar to us. Therefore, users often cannot recognize any difference.

These chatbots have a slightly smaller knowledge base and limited abilities.

In a rule-based chatbot, possible user inputs and corresponding answer options are defined in advance. For example, if a question is asked that was not defined in advance, the chatbot cannot help with answering that question. They can only provide correct output for specific instructions.

This means that the questions asked must correspond to the established programming. Let’s take the weather bot as an example again.

The question “Will it rain tomorrow?” can be easily answered. However, if the bot is not programmed to answer the question “Will I need an umbrella tomorrow?“, it can lead to confusion.

The bot would probably respond with “Sorry, I didn’t understand the question” and can only be as smart as the programming behind it.

If you can only interact with a chatbot through buttons or a predetermined menu and it may be difficult to understand you, you are probably talking to this type of chatbot.

This type of chatbot is often implemented on messenger platforms for marketing purposes, as users do not necessarily need to interact much with the bot. They are good for distributing newsletters or other content, generating leads, or conducting surveys.


Learning through machine learning & knowledge graphs

If you are looking for a customer service chatbot for your company, then this is the right type of automation for you.

Unlike rule-based bots, chatbots that are based on machine learning are able to understand and process natural language. This is done using natural language processing and natural language understanding (NLU).

They understand natural language and therefore do not require such specific commands. They automatically link umbrellas and raincoats, so a deviation from the standard question will not confuse them.

This means that chatbots based on machine learning become smarter with each interaction. The effort behind these automated assistants is, of course, much greater.

Now let’s take a look at the role that knowledge graphs play in this.

How can chatbot conversations be improved?


Two very important factors for the overall performance of a chatbot are the structure and quality of the data available to answer questions. This is where Knowledge Graphs come into play.

A Knowledge Graph is a synonym for a special type of knowledge representation. It stores facts in the form of edges and nodes in a graph. In addition, most knowledge graphs also store the schema of the data. Knowledge graphs realize their full potential, especially in large and complex data structures.

When used for chatbots, a Knowledge Graph offers two direct benefits: improved data integration and a simultaneous improvement in conversations.

New data sources can be more easily integrated, as they only need to be brought into a specific format and schema. Knowledge graphs also offer more flexibility in expanding existing knowledge.

It is also possible to link multiple knowledge graphs together. In this way, a modern managed knowledge database for companies can be built, which can be accessed in natural language, for example, via APIs and interfaces.

How do you know if a chatbot is suitable for your company?

Chatbot & Business

Chatbots are suitable for many industries and use cases. They are known for being able to provide excellent customer service, but there are also some examples in marketing and sales.

Here are some questions as a guide to analyze your current situation and potential benefits of a chatbot:

  • How often does your company have direct customer contact?
  • Does your customer service receive frequent recurring inquiries?
  • How intense are your customer demands?
  • Which communication channels do your customers prefer?
  • What added value can a chatbot create for your company?

What does a chatbot cost?

Unfortunately, this question is not as simple to answer as “How much does a bottle of Fanta cost?”, because the investment you make in a chatbot depends on various factors.

Things like the complexity of the bot, its AI capabilities, its architecture, technical integrations, infrastructure, support during and after launch, and more need to be considered if you want to estimate or compare the costs of a chatbot.

While you can sometimes start for free with very simple bot builders, you may pay setup fees for chatbots built with more advanced bot builders, which can range from €1,000 to €15,000+ and monthly fees between €100 and €5,000+.

We would be happy to provide you with information about our products and prices upon request.

How does the implementation process look like?


This also depends on the factors that are crucial for the costs of the chatbot. At Onlim, we define and implement custom content and all necessary interfaces to data sources together with our customers. In addition, preconfigured modules are available through our conversational AI platform. After the technical setup is completed, thorough testing and optimization are performed before the chatbot goes live. Customers can then manage all chatbot content, analytics, and more through our SaaS platform.

How long does the implementation take?

Our current record is 3 weeks from commissioning to go-live. However, on average, we estimate a turnaround time of 6-10 weeks after commissioning.


What’s next?

If your company is ready to join the intelligent chatbot revolution, schedule a non-binding consultation with Onlim.

We are experts in conversational AI and offer an intelligent chatbot and voice assistant solution.

It is based on state-of-the-art natural language processing (NLP) and knowledge graph technology, includes standardized data and interface connections, and can be installed on websites, apps, messengers, phone systems, or Amazon Alexa.

In addition, our solution can be easily operated and managed without programming knowledge.

In a first step, we can help you identify the best use cases for chatbots in your company.

More Knowledge For Chatbots And Voice Assistants

In the whitepaper “More Knowledge For Chatbots And Voice Assistants” you will learn how meaningful conversations between humans and machines are made possible in automated customer communication through so-called Knowledge Graphs.


Request whitepaper
More articles