Visual Flow Architect
The Chatty Flow Architect provides an interactive drag-and-drop canvas that lets you map out deterministic conversational journeys, lead qualification paths, and custom automated tool triggers.
When a visitor interacts with a flow-enabled bot, the state machine executes node-by-node, falling back to autonomous generative AI whenever the visitor strays from the predefined tree.
Canvas Node Types
Trigger Nodes
Defines entry points into the flow: first widget open, button click, page URL match, or specific keyword intent.
Message & Question Nodes
Presents structured messages, multiple-choice buttons, quick replies, or text input prompts to capture visitor responses.
Condition & Branching
Evaluates variables (e.g. visitor_lead.budget > 5000 or device === "mobile") to fork conversational pathways.
Tool & Handoff Nodes
Executes backend operations: check calendar availability, trigger a webhook, capture a lead, or transfer to a live human agent.
AI Flow Copilot
Instead of manually drafting every branch, you can describe your goal in natural language to the Chatty AI Flow Copilot:
"Build a lead qualification flow for a dental clinic. Ask for the patient's insurance provider, check if they need emergency treatment, propose a 30-minute consultation if yes, or direct them to pricing if no."
The Copilot generates the complete React Flow JSON graph with node coordinates, conditional edges, and fallback triggers automatically.
Flow Architecture API
Save and update flow definitions programmatically:
curl -X POST https://api.chatty.personaliai.com/api/flow/generate \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"bot_id": "YOUR_BOT_ID",
"prompt": "Create an enterprise customer support escalation flow."
}'