Route user requests through conversation flows

Botpress runs conversation flows that interpret what a user is asking for and direct the conversation down the matching path, such as a support request, a sales question, or an account lookup. A Beam agent reads the flow's output as a conversation progresses, applies the routing rule a team has approved, and triggers the next step, such as pulling account data or creating a ticket. It logs the path each conversation took. Requests the flow cannot classify with confidence are routed to a person instead of forced down a default path.

Retrieve answers from a knowledge base

Botpress can connect a bot to a knowledge base of documents so it answers user questions with sourced content rather than a fixed script. A Beam agent reads each user question, retrieves the matching content through Botpress's knowledge base search, and returns the answer along with its source. It logs which questions were answered and which had no confident match, so gaps in the knowledge base can be identified and closed over time. Questions with no matching source, or where the retrieved content conflicts with itself, are routed to a person to answer directly.

Trigger downstream workflow actions

Botpress flows can call external actions, such as looking up a record, updating a system, or sending a notification, as part of a conversation. A Beam agent reads when a flow reaches an action step, applies the approved rule for what that action should do, such as updating a customer record with information collected in the chat, and executes the action through Botpress. It records the outcome of each action. Actions involving a change the team has not pre-approved, such as a refund or account cancellation, are routed to a person to complete.