You can provide a ServisBOT skin to a new or existing Dialogflow bot with a few simple steps. The main difference is that the process to create the bot should start in Dialogflow. Managing the intent will happen there too. The main benefit of wrapping a Dialogflow bot with a ServisBOT experience is the ability for the bot to use markup, and how quickly it is to get a front end up and running.
<TimelineMessage> <List selectable="true" title="Who is your favourite dwarf?" style="large-icon"> <Item id="0" title="Grumpy" iconUrl="https://imgplaceholder.com/48x48/ffffff/333333/fa-image" /> <Item id="1" title="Happy" iconUrl="https://imgplaceholder.com/48x48/ffffff/333333/fa-globe" /> <Item id="2" title="Sleepy" iconUrl="https://imgplaceholder.com/48x48/ffffff/333333/fa-book" /> <Item id="3" title="Bashful" iconUrl="https://imgplaceholder.com/48x48/ffffff/333333/fa-car" /> <Item id="4" title="Sneezy" iconUrl="https://imgplaceholder.com/48x48/ffffff/333333/fa-building" /> </List> </TimelineMessage>
You can train a DialogFlow agent converse more like a human by enabling smalltalk. This allows the agent to answer a variety of questions from ‘How are you?’ to ‘Thanks for your help’.
When using smalltalk, it is advised to not have any intent names start
smalltalk., as the smalltalk intent and the actual intent may clash.
To get started, you will first need to open the smalltalk section on the left navigation bar.
Once at the smalltalk page, click the ‘enable’ button and click ‘save’ at the top.
By default, DialogFlow provides a variety of built-in smalltalk responses. The answers to different questions are customizable by editing the categories below. Multiple response variations can be entered for each utterance if desired.
Once ready, you can test your agent by typing in utterances into the console on the right of the page.