Building an Automation Workflow

This guide walks through building an automation in the visual workflow builder — from a blank canvas or a preset, to a tested, active flow that reacts to incoming messages on its own.

1. Open the automation builder

From the dashboard sidebar, open Automation, then start a new workflow at /user/automation/create. This opens the visual, node-based canvas where the whole workflow is built by connecting blocks.

2. Start from a preset or a blank canvas

You can start two ways:

• Preset — choose from 11 ready-made workflow presets covering common scenarios (things like auto-replying to a keyword, sending a welcome sequence, or following up after a period of inactivity). A preset gives you a working structure you can edit rather than starting from nothing.
• Blank canvas — start empty and build the flow node by node, useful once you know exactly what logic you need.

3. Add a start node

Every workflow begins with a start node — the entry point that defines what kicks the automation off. Presets already include one; on a blank canvas, this is the first block you place.

4. Add trigger nodes to react to incoming messages

Add keyword/trigger nodes so the workflow reacts to what a contact sends — for example, firing when an incoming WhatsApp message matches a specific keyword or pattern. This is what turns the workflow from a one-off send into something that responds automatically to real conversations.

5. Add wait nodes for delays

Drop in wait nodes wherever the flow needs to pause before continuing — for example, waiting a few hours before a follow-up message, instead of sending everything back-to-back instantly.

6. Add condition nodes to branch the flow

Use condition nodes to split the workflow down different paths based on data about the contact or their reply — so, for instance, contacts who reply "yes" go one way and everyone else goes another, all within the same workflow.

7. (Optional) Add AI nodes

For steps that should be handled by AI rather than a fixed script, add:

• ai_send — sends a message generated by the AI at that point in the flow.
• ai_generate — generates content (for use elsewhere in the workflow) using AI rather than a static template.

AI nodes are useful when a reply needs to sound natural and vary contact-to-contact, rather than reusing the exact same wording every time. See AI WhatsApp Bot for how the underlying AI layer and knowledge base work.

8. Test before activating

Before turning the workflow live, run a test to confirm it behaves as expected — triggers fire correctly, conditions branch the way you intended, and messages read the way you want them to. Testing catches logic mistakes while the workflow can't yet reach real contacts.

9. Activate the workflow

Once you're satisfied with the test run, activate the workflow. It now runs continuously in the background, reacting to incoming messages and contact behavior without further manual action from you.

Next steps

• Automation Workflows — the full feature reference, including all available node types and preset details.
• AI WhatsApp Bot — how AI nodes connect to bot personas and knowledge bases.