Use this to create a conversational agent (aka chatbot) that can respond to a large variety of requests (e.g. answer employee questions about company HR policies). This will display a chatbox for the user to enter a freeform message and upload documents.
All steps in the agent are automatically presented with the contents of the user’s message along with previous user messages and agent responses from the current chat session.
You can optionally specify conversation starters in the chat message trigger. These are examples of messages that the user can send to the agent. You can also create placeholders in conversation starters using the format [[placeholder]] – when the user clicks on a conversation starter with placeholders, they will be asked to fill in the values for the placeholders before the message is sent to the agent.
Use this to create a task based agent that executes a specific workflow (e.g. suggest how to resolve a support ticket). This will display a form with the input fields that the agent needs to complete the task.
There are 3 possible input field types:
Text – freeform text
Document – requires the user to either either upload a document or enter the URL of a document that is indexed by Glean
Multiple choice – presents the user with a set of options to choose from
Input fields are not automatically presented to steps in the agent. Instead, you must explicitly tag an input field using the format [[input field name]] when you want a step to have access to it.