Build conversations with clear steps and conditions.
Use the integration actions shown in your editor for connected services. Long-running delays and A/B testing are not available.
WhatsApp flow builder
Learn how to build WhatsApp workflows without code.
Choose what starts your chatbot, what it asks, and how it handles each reply. Waaru's visual editor has 28 step types for messages, forms, conditions, data updates, and team handoffs.
In one line
Meta's WhatsApp Flows are forms customers complete inside WhatsApp. Waaru's Logic flows (Chatbot) control the conversation around them: open a published form, use its answers, check conditions, and hand the conversation to your team. Review validation errors before publishing. Conversations already running keep the version they started with.
Version-pinnedLive conversations keep their version
The problem
Plan for unexpected replies.
A customer may type a date where you expected a number, ignore a button, or ask for a person. Decide what your chatbot should do in each case before you publish it.
Send unmatched replies to a fallback workflow or the team inbox.
Check each answer’s format and limit how often you ask the same question.
Published forms, templates, and sub-flows must stay pinned to reviewed resources.
Active conversations must not silently switch to a draft edited later.
Jump to a section
Choose from 28 workflow step types.
Each node is one step on the canvas. Choose the step for the message, decision, or update you need.
6 triggers for keywords, conversation entry, fallback, template replies, Click-to-WhatsApp referrals, and received media
9 message types for text, media, interactive options, catalogs, templates, WhatsApp Flows, locations, and contact cards
4 collection and data nodes for typed responses, variables, contact fields, and tags
5 routing nodes for conditions, short delays, jumps, pinned sub-flows, and explicit exits
3 inbox actions for assignment, status changes, and internal notes
1 integration node for approved connector actions
Handle replies your chatbot did not expect.
A fallback trigger handles messages that do not match another published trigger. For each question, choose what happens if the answer is valid, has the wrong format, or remains invalid after retries. When no automation matches, the message goes to the human inbox.
Native WhatsApp Flows for structured data.
Choose a published WhatsApp Flow and map its submitted values into downstream flow or contact fields. The node exposes submitted and invalid-payload outcomes so malformed responses do not continue silently.
Edit a flow without changing conversations already running.
New conversations start from the explicitly published snapshot. Existing runs resume against their pinned version, including pinned sub-flow references. Draft changes do not alter those runs. Traffic splitting and one-click rollback are not active dashboard capabilities.
Review starter flows before publishing.
Copy a starter from the library into your workspace. It stays a draft while you choose its templates and forms, review the reply paths, and fix validation errors. Publish it only after those checks.
FAQ
Frequently asked questions.
What is a WhatsApp flow builder?
The term can refer to building Meta's WhatsApp Flows, which are forms inside WhatsApp, or to building chatbot workflows. Waaru supports both: WhatsApp Forms collect structured answers, while Logic flows (Chatbot) control messages, conditions, connected actions, and human handoff. A chatbot workflow can open a published form and continue after the customer submits it.
How does Waaru handle an unexpected reply?
Use explicit fallback triggers and validation outcomes. When no published automation matches, the message enters the human inbox. This Logic Flow fallback does not call an AI model.
Does it support Meta's WhatsApp Flows (native in-chat forms)?
Yes. Select a published WhatsApp Flow, choose its screen, map supported prefill values, and store submitted answers as flow or contact variables.
Can I edit a flow that's already live?
Yes. Save a new draft version and publish it after validation. Existing runs stay pinned to their original version. Traffic splitting and one-click rollback are not active in the current dashboard.
Do I need to code anything?
No. Configure the available node types in the visual editor. API request and payment action nodes are not part of the current release.
Can a flow hand off to a human?
Yes. The handoff node pauses automation and assigns the existing conversation in the team inbox. The transcript and contact context remain attached to the thread.
Are templates included?
A starter library is available. Each copied starter remains a draft until you configure its resources, review its branches, and publish it.
What does the flow builder cost?
Deterministic flow access is free. AI features are paid separately. Each canvas supports up to 200 nodes and 400 edges.
What is the current release status?
Logic Flow V2 is available as the deterministic dashboard, API, and worker path. It remains available alongside Waaru's AI agent layer for journeys that need explicit state and branching.
