Conversational user interfaces (aka chatbots or bots) come in two flavors, directed (aka rule-based) and natural language processing (NLP or AI-based).

NLP-based Chatbots Are More Human-like

They can converse with humans on a semi-personal level. They can have conversations without the need of typing. They take advantage of deep learning to interpret your intent and attempt to express their response based on that interpretation. Although they are constantly improving, they are limited by the data sources used to train them. As a result, the outcomes can sometimes be less than satisfactory.

Directed Chatbots Guide the User Better

As a result, your organization can easily manage the chatbot without the need for developer involvement. Although not as flashy, these are in many situations the better alternative. Products that use directed chatbots tend to be simpler and easier for domain experts to manage, and as a result far more scalable than products that require programming expertise.

Which Is Best Suited for What You Want It To Do?

This video explores the pros and cons of each approach. It also investigates  how business analysts and those who define digital solutions are essential for defining, designing, and implementing either of these modern forms of human-machine interaction.