A chatbot on the company website relieves the customer service team. But is it worth it for your company? An overview of types of chatbots, chatbot providers, and costs.
They usually wait for visitors in the bottom right of the browser window and start conversations with friendly requests such as: “Ask me a question” or “I’m happy to help you”. Chatbots in customer service answer questions day and night in a matter of seconds and with stoic patience. This makes them practical helpers in customer service for companies, who can save costs and relieve employees.
But for which companies is it worthwhile to have their own chatbot for the company website and how do medium-sized companies find suitable software for this? This text will help you make a decision, and show you examples of chatbots, chatbots in customer service – does it work?
Programming your own chatbot for the company website has long been a huge project. With the success of AI chatbots based on language models such as ChatGPT from OpenAI or Gemini from Google, the reputation of chatbots in customer service is also improving.
Companies can now set up a customer chatbot with little time and personnel investment if they opt for off-the-shelf software. There are now dozens of these to choose from.
Two technologies are currently contributing significantly to the success of chatbots for companies:
- natural language processing (NLP), which allows bots to understand and interpret human questions,
- and generative artificial intelligence (AI), which creates individual answers to customers’ questions.
The combination makes chats sometimes feel as if a human is typing instead of the chatbot on the company website. and give an overview of chatbot providers.
Types of Chatbots in Customer Service
“In the first conversation, I usually have to determine together with the companies what the word chatbot actually means to them,” says Markus Mölter, managing director of the digital agency MoSeven from Kerpen-Sindorf near Cologne. Because there are different types of chatbots:
Rule-based chatbots
These are dialog systems that recognize keywords and provide prescribed answers based on set rules. They can only answer questions that are stored in their database. The advantage of these chatbots in customer service is that they can be programmed relatively quickly and then only need to be updated if answers change.
However, conversations with them do not feel very natural for users. Instead of free text fields in which customers type their questions, rule-based chatbots often offer topics or questions to choose from via a button to click on. Typical applications are chatbots that help with making appointments or that customers can use to inquire about the status of their orders in online shops. As soon as customers ask a question outside of the specified topics, the dialogue system reaches its limits.
AI chatbots
These dialogue systems generate a unique answer for each customer query. This makes conversations with them feel more natural for customers than the stiff statements of rule-based chatbots. There is one disadvantage, however: although the AI system can in principle generate an answer to every question, the texts are based on probabilities and can be wrong.
AI systems need to be trained to give the right answers. Most chatbot providers use pre-trained AI models that are then adapted to the company. This reduces the training effort.
Hybrid chatbots
These are dialogue programs that use both AI and rule-based response systems. One advantage of this software is that companies can, for example, narrow down sensitive topics where they do not want to trust generative AI and instead prefer predefined statements.
For which companies is a chatbot worthwhile?
Two questions help entrepreneurs assess whether a chatbot is worthwhile for their website:
- How many customer inquiries does the company receive via its website?
The simple rule of thumb here is: the more work customer questions cause in the company, the more relief a bot can provide. “Chatbots are particularly worthwhile when the same questions keep cropping up,” says Frederic Kerber, head of the Innovative Retail Laboratory at the German Research Center for Artificial Intelligence. - What questions do customers ask?
Kerber recommends creating a list of the 20 most common customer inquiries to get an overview and assess whether a chatbot can answer these questions.
Is the chatbot compatible with other software?
The most common questions also provide information about whether the chatbot for the company website should be linked to existing systems, for example to provide information on the status of orders. This is why entrepreneurs should make sure that the chatbot software is compatible with the interfaces of the existing systems. IT service provider Mölter advises involving the company’s current service providers. “Medium-sized companies often have no idea what their service providers can do. It’s worth checking whether the website service provider also offers chatbot solutions, for example.”
Setting up a chatbot for your company website
The services offered by chatbot providers range from low- or no-code workspaces, in which companies can click together simple dialogs themselves, to multi-stage introductions to the system – depending on how complex the chatbot’s customer service tasks are. AI chatbots generate their answers from knowledge databases into which companies can upload documents such as FAQs or product descriptions. “In my experience, compiling this information is the most time-consuming step,” says Mölter. Because, for example, specialist departments have to revise the texts.
How reliably or creatively should the chatbot respond?
While more rule-based bots reliably provide accurate answers, AI chatbots surprise customers with very natural dialogues. “When customers use one of these old-generation chatbots, many are put off. They find the interaction with such machines terrible,” says Swiss chatbot consultant Sophie Hundertmark. “The customer experience is so much better that I would rather risk mistakes than put up with the bad dialogues.” IT entrepreneur Mölter, on the other hand, points out that the effort involved in using AI chatbots is greater in the long term.
How much effort is required to maintain the chatbot?
AI chatbots learn as they are used, but they need to be monitored. “It can happen, for example, that the AI provider changes the language model under the hood, for example because there is a newer version,” says Kerber. This can lead to a chatbot on the company website suddenly behaving differently. “If I decide on a system like this, I should make sure before I introduce it that the company has the capacity for long-term maintenance and that the system is easy to use even for laypeople.”
What service does the chatbot provider offer?
That’s why it’s worth checking what service and support the chatbot provider promises. “It makes a big difference whether you have to call the USA in English or have a small provider with customer service in German,” says chatbot consultant Hundertmark. Of course, another important criterion in this context is how much the chatbot costs.
How does the chatbot fit into data protection?
The EU’s AI regulation will come into full force in August 2026. Among other things, it stipulates transparency obligations for providers and operators of AI systems. “This includes the obligation to indicate when customers are speaking or chatting with an AI,” says Achenbach. Due to the GDPR, however, it is already necessary to provide sufficient information about the type of data processing. This also includes an indication that AI is being used.
Another important criterion is that the chatbot is compatible with data protection. The company usually concludes a data processing contract with the chatbot manufacturer for this purpose. “If data is processed outside the EU, you also have to ensure that the data is protected there in the same way as it would be on European servers,” says lawyer Matthias Achenbach from the Berlin law firm HK2 Rechtsanwälte, which specializes in IT law. He also advises that companies should ensure in the contract with the software service provider that the use of customer data to train the AI language models is excluded.
Chatbot providers at a glance
According to the chatbot manufacturers, all chatbots listed are GDPR compliant. The chatbot costs refer to the cheapest service package for companies.
Chatbot4you
What can the chatbot do? For customer service or lead generation, for example, companies can create rule-based chatbots in the browser without having to write code. The system is self-learning and can improve answers with the help of AI.
Integrated AI systems: ChatGPT
Costs: from 45 euros/month
Link: Chatbot4you.io
chathero.ai
What can the chatbot do? Companies can create dialogs using drag and drop and independently create a knowledge database for the AI by uploading documents. The chatbots can be linked to other applications for lead generation or appointment booking.
Integrated AI systems: ChatGPT
Costs: from 39 euros/month
Link: chathero.ai
OMQ
What can the chatbot do? The chatbot recognizes what users mean – even if the wording is more difficult to understand. It can be connected to various applications for actions such as address changes or subscription cancellations.
Integrated AI systems: ChatGPT
Costs: from 150 euros/month
Link: omq.ai
melibo
What can the chatbot do? The dialogue system is specialized in e-commerce and answers questions about products, order status, availability or returns. Companies receive onboarding and can maintain their chatbot themselves.
Integrated AI systems: Including GPT-4
Costs: from 450 euros/month
Link: melibo.de
Kauz.ai
What can the chatbot do? Rule-based dialogues are possible, as is the integration of AI. The chatbot can be linked to other applications to collect feedback, send emails and arrange appointments. Companies receive onboarding and can maintain their chatbot themselves.
Integrated AI systems: ChatGPT plus rule-based system, the operation of your own models is also possible
Costs: from 499 euros/month plus 1500 euros setup fee
Link: kauz.ai
dialog bits
What can the chatbot do? Companies can set up chatbots themselves, for example for customer service or marketing, and upload documents or videos as sources of information for the AI. The AI trains itself with the information.
Integrated AI systems: ChatGPT or on-premises (on your own hardware)
Costs: from 795 euros/month
Link: dialogbits.com
moin AI from Knowhere gmbh
What can the chatbot do? The chatbot asks questions when it is unclear whether it has understood the user. The chatbot learns independently and can be linked to applications such as Shopify. Areas of application: in marketing, for example for lead generation, and in customer service.
Integrated AI systems: Own AI models
Costs: From 850 euros/month plus setup fee
Link: moin.ai
botario
What can the chatbot do? Chat system for customer support that also helps to transfer customers to employees in live chats. Companies can create dialogs themselves using drag and drop.
Integrated AI systems: GPT models from Google, Open AI, Microsoft and on-premises
Costs: from 949 euros/month
Link: botario.com
Assono
What can the chatbot do? Rule-based chatbot that can be expanded with AI responses. Connection to product databases is possible, for example.
Integrated AI systems: IBM Watson Assistant
Costs: from 890 euros/
month
Link: assono.com
botfriends
What can the chatbot do? The chatbot answers customer questions, advises on products and can be linked to various systems, for example to complete orders or arrange appointments. Companies can create bots themselves and receive onboarding.
Integrated AI systems: Users can choose from different language models
Costs: 1190 euros/month
Link: botfriends.de
Parloa
What can the chatbot do? The chatbot can be used in self-service, for example when customers want to change their data.
Integrated AI systems: Based on ChatGPT, among others
Costs: On request
Link: parloa.com
Chatchamp
What can the chatbot do? Specialized in e-commerce, customers can use the bot to track packages or complain about goods, for example.
Integrated AI systems: ChatGPT and other models from Open AI
Costs: On request
Link: chatchamp.com
Userlike
What can the chatbot do? In addition to the website chatbot, the company also offers messenger chatbots for WhatsApp, Facebook, and Instagram.
Integrated AI systems: Including GPT-4
Costs: On request
Link: userlike.com
