The rule-based chatbot NLP-based chatbots
Opportunities
  • Responses structured on multiple levels to provide more meaningful and complete answers
  • High precision in understanding user queries
  • Users can ask their questions anonymously (important for sensitive or personal queries)
  • Free text entries that were not understood serve as learning input for the chatbot
  • Text input can be evaluated by the corresponding functional team
Strengths
  • Topics that the user can select are clear from the outset
  • Users can be guided to a destination preferred by the company
  • Natural way of conversation
  • Fast and simple operation
  • Positive impact on UX
  • Wow effect is triggered
Freedom for users Users must follow a path for the answer Users can simply write their requests
Understanding of a request Users always receive a 'correct' answer through the buttons NLP understands and interprets on average 90% of all queries correctly
Degree of automation Low, as the number of answers is limited High, as more and more questions can be answered
Expectation of chatbot Buttons make it clear what the user can expect Chatbot needs to communicate very precisely what it knows and what it doesn't (onboarding)
Trend detection Highly limited due to lack of free text input High because users can simply write any request