Skip to main content
The Flows module lets you build conversational flows visually — think of it as a canvas where you design the exact path your AI agent should follow during a conversation. Instead of relying solely on the AI’s free-form responses, Flows give you granular control over how interactions unfold step by step. Use Flows when you need:
  • Structured conversations: Guide the user through a specific sequence (e.g. onboarding, troubleshooting, data collection).
  • Human escalation paths: Define clear rules for when and how a conversation should be transferred to a human agent.
  • API integrations: Call external services at specific points in the conversation to fetch or send data.
  • Lovi Tools: Leverage built-in Lovi capabilities within the flow for advanced actions.

Getting Started

When you open the Flows section for the first time, you’ll see an empty state inviting you to create your first flow. Click the + New flow button in the top-right corner (or the + Create my first flow button in the centre) to get started.

Creating a New Flow

A dialog will appear asking you to configure the basics of your flow:
FieldDescription
Flow name (required)A descriptive name for your flow (e.g. Greeting and policies, Order tracking).
Description (optional)A brief explanation of what this flow does. Useful for keeping your team aligned.
AI agent (required)The AI agent this flow will be linked to. Select one from the dropdown.
Flow typeChoose how you want to start building.

Flow Types

TypeDescription
Custom (empty)Start from scratch with just a trigger node. Full creative freedom to build your own flow.
Human escalationA pre-configured template designed for handing off conversations from the AI agent to a human agent. Ideal for support workflows.
Once you’ve filled in the details, click Create flow to enter the flow editor.

The Flow Editor (Canvas)

After creating a flow, you’ll be taken to the flow editor — an interactive canvas where you design your conversational tree using drag-and-drop nodes and connections. At the top of the editor you’ll see:
  • The flow name and the linked AI agent.
  • A toolbar with the available node types.
  • A Save button to persist your changes.
At the bottom-left, you’ll find canvas controls: zoom in, zoom out, fit to screen, and lock/unlock the canvas.

Node Types

Nodes are the building blocks of your flow. Each node represents a step in the conversation. You add them from the toolbar at the top of the editor and connect them to define the conversation path.

1. Agent Node

The Agent node represents a point where the AI agent speaks or performs an action. Use it to:
  • Send a message to the user.
  • Process information using the AI engine.
  • Make decisions based on the conversation context.
This is the most common node — it’s the AI doing its job within the flow.

2. User Node

The User node represents a point where the system waits for user input. Use it to:
  • Capture a specific response from the user.
  • Branch the flow based on what the user says or selects.
  • Collect data (names, emails, preferences, etc.).
Think of it as a “pause and listen” step in the conversation.

3. API Node

The API node lets you integrate external services directly into the flow. Use it to:
  • Fetch data from a third-party system (e.g. order status, account info).
  • Send data to an external service (e.g. create a ticket, update a CRM).
  • Trigger webhooks at specific conversation points.
This is how your flow connects to the outside world.

4. Lovi Tools Node

The Lovi Tools node gives you access to Lovi’s built-in capabilities within the flow. Use it to:
  • Perform platform-specific actions (e.g. tag a conversation, assign to a department).
  • Use advanced Lovi features that go beyond simple text responses.
  • Automate internal operations as part of the conversational flow.

Building a Flow

Here’s the general workflow for building a conversational flow:
  1. Start with the trigger: Every flow begins with a trigger node that defines when the flow starts.
  2. Add nodes: Drag nodes from the toolbar onto the canvas.
  3. Connect nodes: Draw connections between nodes to define the conversation path. Simply click on a node’s output port and drag to another node’s input port.
  4. Configure each node: Click on a node to open its settings panel and define what it does (messages, conditions, API calls, etc.).
  5. Save: Click the Save button to persist your flow.

Tips for Effective Flows

  • Keep it simple: Start with a minimal flow and expand as needed. Overly complex trees are hard to maintain.
  • Test frequently: After saving, test the flow through your linked AI agent to make sure the conversation feels natural.
  • Use branching wisely: User nodes can branch into multiple paths — use this to handle different user intents without making the flow too deep.
  • Combine with AI: Flows don’t replace the AI — they guide it. The AI agent still handles the nuance; the flow provides the structure.

Best Practices

  • Name your flows clearly: Use descriptive names so your team knows what each flow handles at a glance.
  • One flow per objective: Instead of building one massive flow, create separate flows for distinct goals (e.g. one for onboarding, another for support escalation).
  • Document with descriptions: Use the description field when creating a flow — your future self (and your colleagues) will thank you.
  • Review and iterate: Flows are living documents. Check your analytics to see where users drop off and refine the paths accordingly.

Quick Reference

ActionHow
Create a new flowClick + New flow in the top-right corner
Add a nodeClick the node type button in the toolbar
Connect nodesDrag from one node’s output to another’s input
Configure a nodeClick on the node in the canvas
Save changesClick the Save button
Zoom in/outUse the controls at the bottom-left of the canvas
Fit to screenClick the fit-to-screen icon at the bottom-left