Chatbot Rules Engine

The chatbot rules engine lets you define automated responses that fire when specific conditions are met. Rules are evaluated in priority order: the first matching rule sends its response. Use rules to answer FAQs, collect information, or route conversations automatically.

Overview

Each rule has two parts:

• Condition: when the rule triggers (keyword match, regex, always-on)
• Response: what the chatbot sends (text, quick replies, or info collection form)

Rules can be toggled on/off without deleting them, making it easy to temporarily disable a rule during testing.

Rule Types

Keyword Match

Triggers when the visitor's message contains a specific word or phrase. Matching is exact: use multiple keywords with separate rules for broader coverage.

{
  "conditions": {
    "type": "keyword",
    "value": "pricing"
  }
}

Regex Match

Triggers when the visitor's message matches a regular expression pattern. Use this for flexible patterns like email addresses, order numbers, or phone numbers.

{
  "conditions": {
    "type": "regex",
    "value": "order-[A-Z]{2}-\\d{4}"
  }
}

Always

Triggers on every message when no other rule matches first. Use for a fallback response when the chatbot cannot answer the question.

{
  "conditions": {
    "type": "always",
    "value": ""
  }
}

No Operator Available

Triggers when no operator is currently online. Useful for collecting information during off-hours.

{
  "conditions": {
    "type": "no_operator_available",
    "value": ""
  }
}

Quick Replies

Present the visitor with clickable button options. Each quick reply sends the button text back as the visitor's message, which can then match downstream rules.

{
  "response": {
    "type": "quick_replies",
    "message": "What would you like help with?",
    "quickReplies": [
      "Pricing",
      "Technical Support",
      "Talk to a human"
    ]
  }
}

Info Collection

Collect structured information from the visitor with a form-like experience. Each field has a label, internal field name, and required flag.

{
  "response": {
    "type": "collect_info",
    "message": "Please provide your details so we can help faster.",
    "collectFields": [
      {
        "label": "Full Name",
        "field": "name",
        "required": true
      },
      {
        "label": "Email Address",
        "field": "email",
        "required": true
      },
      {
        "label": "Order Number",
        "field": "orderNumber",
        "required": false
      }
    ]
  }
}

Operating Hours

Configure the chatbot to automatically switch behavior based on operating hours. When operating hours are enabled, the chatbot can use the no_operator_available condition during off-hours.

GET /api/chatbot/config

{
  "enabled": true,
  "greetingMessage": "Hi! How can we help you today?",
  "fallbackMessage": "I couldn't understand that. Let me connect you with an operator.",
  "operatingHoursEnabled": true,
  "operatingHoursStart": "09:00",
  "operatingHoursEnd": "17:00"
}

Update the config to set greetings, fallback messages, and operating hours:

PUT /api/chatbot/config

{
  "greetingMessage": "Welcome to support!",
  "operatingHoursStart": "08:00",
  "operatingHoursEnd": "20:00"
}

Rule Priority

Rules are evaluated in ascending priority order. Lower priority numbers are checked first. Use priority to control which rule wins when multiple rules could match the same message.

{
  "name": "Order Status Lookup",
  "enabled": true,
  "priority": 0,
  "conditions": {
    "type": "regex",
    "value": "order-[A-Z]{2}-\\d{4}"
  },
  "response": {
    "type": "text",
    "message": "I'll look up that order for you right away."
  }
}

API Reference

Rules

Config

All endpoints require authentication. Deleting rules requires admin role.

Related

• Knowledge Base: articles that power AI responses
• Proactive Engagement Triggers: automatically engage visitors based on behavior
• Triggers: widget-level trigger configuration