You can use contexts to handle the flow of the conversation. Contexts are used to help dialogflow detect the correct intent. Output contexts are the contexts created after the intent was detected, and input context is a context required for an intent to be detected.
Think of dialogflow contexts, as the real context of a conversation, there are some phrases that you can expect after an interaction.
Here's an example using 3 intents for this conversation:
- "Default welcome"
- "Updating name selection"
- "Updating name"
Conversation
Intent detected: Default welcome
Agent says: "Select an option to update [options are {name, age, color, place}]"
Output context: "select-option"
User says: "name"
Input context: "select-option"
Intent detected: Updating name selection
Agent says: enter your name
Output context: "update-name"
User says: "My name is Xyz"
Input context: "update-name"
Intent detected: Updating name
Agent says: "record updated"
In addition to this, there are follow-up intents which are used for this scenarios, one of the advantages is that in follow-up intents, the contexts are created automatically.
Parameters
You can get all the information to be updated within the same intent. For this, you can add the paramaters as required and take advantage of slot filling. Notice that you can specify the define prompts to be used when asking for these parameters.
You would need to add training phrases that help indetifying different parameters given:
During the conversation, when this intent gets detected, the agent will try to match all the required parameters, and will ask for missing information. At the end, you will have access to this information:
Consider that this kind of intents can quickly become hard to manage. Try using well defined entities for your parameters, and lots of training phrases (notice that in one training phrase you can collect more than one parameter, e.g., "My name is Xyz and I'm 27").
