Let’s look at a typical use case, the end user asking for support. Often in front of the communication is a chatbot. The user asks and the chatbot responds using AI/LLM based knowledge bases. The chatbot can then continue to “flow”, for example when it is not satisfied with the automatic response and wants to connect with a human. Before connecting the end user and the operator, the chatbot will ask for some data, show the privacy policy for acceptance, etc. Sometimes the user data is also sent to an external CRM like Hubspot, and sometimes the chatbot responds to the user with data obtained via some Company’s API.
The way you build today is very different from the way you built yesterday. We have moved from an intent-based design approach to a block-based one. The heavy use of a programmatic approach in complex intent responses has moved to a visual, no-code approach for developing responses. And that eventually happened, chatbots have met the no-code movement by using a lot of visual graphical representation of complex logic, based on concepts like nodes, blocks, actions, conditional branching, etc.
Chatbots have gone from intent-based to block - based . A block is a piece of logic that can be interconnected in various ways to other blocks, making your chatbot's logic complex.
For the no-code or citizen developer , it is very important to equip yourself with a flexible tool that uses the same visual language that you use to draw logic on a piece of paper.
Intentions weren’t enough for these kinds of complex flows. Intents were easy to adopt when building a question-answer bot, but suddenly found their limits as chatbots became more complex.
