Skip to content

Chapter 3 — Setup

Organizations & Projects

Creating a VoiceBot organisation, provisioning a conversation project, and configuring project-level call handling, recording, chatbot endpoints and knowledge base integration.

Creating a new VoiceBot organisation

Organisations represent distinct VoiceBot configurations and are created through the BlockFlow interface.

  1. Open the Organization tab.
  2. Click Create Organization in the upper right.
  3. Fill in the basic organisation information and assign the users who belong to it.

Figure 1 — Create Organization dialog

Figure 1 — Create Organization dialog.

Creating a conversation project

Once the organisation is configured, projects can be created from the default Projects page using the Create Project button. The dialog requires the following details:

Field Purpose
Project name Name of the new VoiceBot environment, for example “Customer Support Bot — English”.
Language The language the VoiceBot uses for speech recognition, synthesis and natural language understanding.
Organization The organisation within which the project resides.
Conversation Monitor project ID The identifier the Conversation Monitor uses to associate and log the project's conversations. The Conversation Monitor provides the human-in-the-loop functions: call recording, post-processing and editing, transcription, LLM fine-tuning and TTS improvement.
Description Short description outlining the purpose or scope of the VoiceBot environment.
Mode Stage during development; migrated to Production once the project is production-ready.
Status Maintenance metadata — Building Scenario, Testing, or Production.
Active Toggles between Live and Maintenance mode in production. In Maintenance mode calls are redirected to the configured fallback number.

Figure 2 — Create new VoiceBot Project dialog

Figure 2 — Create new VoiceBot Project dialog.

Project settings

Administrators configure call-handling behaviour, recording parameters, chatbot communication endpoints, knowledge base integration and monitoring settings. The dialog is organised into the sections below; Save stores the updated parameters and Reset reverts every field to its default or last saved value.

Redirect message properties

  • Maximum call attempts per user — how many calls a user may initiate within a defined time window (6 hours) before calls are redirected to human agents.
  • Chatbot timeout — maximum waiting time for a chatbot response before a fallback action triggers.
  • Redirect VDN number — the number or extension calls transfer to when the chatbot fails to respond, the conversation cannot proceed, or the project is in maintenance mode.
  • Success tag — marks successful completion of a scenario, for example appointment creation.
  • Redirect message — played to the user when a redirection is triggered by timeout or failure.

Message recording

Defaults for capturing the caller's voice input; both can be overridden per node in a scenario.

  • Recording duration — maximum time the user may speak during an interaction segment.
  • Recording silence — silence threshold after which recording stops automatically.

Chatbot communication

  • Conversation Logger API URL — stores conversation logs, exchanged messages and metadata.
  • Chatbot API — primary endpoint for sending user input and receiving AI-generated responses.
  • Chat UI API — handles user message delivery and chatbot reply processing in the conversational interface.
  • Health Check Monitoring API — verifies chatbot availability and responsiveness.
  • Initiate Call API — triggers outbound calls initiated by the VoiceBot.

Knowledge base integration

  • Knowledge Base Refresh API — reloads or refreshes the knowledge base when updates occur.
  • RAG URL — default retrieval-augmented generation endpoint for queries against indexed knowledge.
  • Neo4j — connection endpoint for the graph database used for structured knowledge representation and GraphRAG queries.

Figure 3 — BlockFlow project configuration

Figure 3 — BlockFlow project configuration.