I recently published a series of YouTube videos which explain the basics of a framework called CTFS. You can use it to better understand the behavior of your Dialogflow chatbot.
It might be more useful if you watched the videos alongside the related blog posts.
What is the CTFS framework?
In the videos, I propose using the following 4 types of intents to study your chatbot conversation - candidate intents, target intents, fallback intents (similar to actual fallback intents) and surplus intents.
Once you understand the basic idea behind these different types of intents, you can use the following formula:
C = T + F + S
where C is the number of candidate intents, T is the number of target intents, F is the number of fallback intents and S is the number of surplus intents.
When designing your chatbot, your goal should be to minimize C and T at each step in the conversation, and eliminate S completely if possible.
Why use the framework?
It is certainly possible to design your Dialogflow bot without using this framework. For example, suppose you have a one-and-done FAQ chatbot where users never ask any followup questions. You can just design your chatbot using nothing more than a CSV file which has all the questions and answers.
If you choose to use the framework, it will help you do the following:
- predict which intent will get mapped with greater accurary
- avoid the mapping of the wrong intents (for that step in the conversation)
- choose good contexts for your dialog
- by reducing the number of target intents, you can study your chatbot's behavior systematically
- Set up better Chatbase funnels
- Create complex analytics
In other words, if you are building a hobby bot, or just building a chatbot for your own learning, you should skip learning about the framework.
If you are building a more commercial grade chatbot, this framework can help you build Dialogflow chatbots which have predictable behavior.