Automation Workflow

Set up automated workflows to streamline your communication processes.

Overview

Voxvaani’s automation capabilities help reduce manual work by handling repetitive tasks automatically — from AI-powered responses to automated WhatsApp replies. Automation reduces response time, ensures consistency, and frees up human agents for complex inquiries.

⚠️ Test Before Enabling

Always test automation rules in a controlled environment before enabling them in production. An incorrect keyword trigger or overly broad rule can auto-reply to every incoming message, potentially confusing customers. Run each rule with a test contact or internal number first, verify the response is appropriate, then enable it.

Types of Automation

WhatsApp Automation

WhatsApp Automation lets you set rules that trigger automatic replies:

Keyword Triggers — Auto-reply when a message contains specific words. Example: When a message contains “price”, auto-reply with your pricing brochure and a link to schedule a call.
Message Type Triggers — Respond based on message category (text, image, document, location). Example: When an image is received, auto-reply “Thanks for sharing! Our team will review this and get back to you.”
Time-Based Rules — Different responses for different hours. Example: During business hours (9 AM-6 PM), reply “A team member will respond within 1 hour.” Outside business hours, reply “We’re currently closed. We’ll respond first thing in the morning.”

Advanced Keyword Configuration:

Exact match: Trigger only when the message equals a keyword exactly (e.g., “STOP” or “UNSUBSCRIBE”)
Contains: Trigger when the keyword appears anywhere in the message (e.g., “price” matches “What is the price?”)
Case sensitivity: Toggle case-sensitive matching for precise control
Multiple keywords per rule: Group related terms (e.g., “price”, “cost”, “pricing”, “rate”)

Escalation Fallback: Every automation rule should define a fallback action when the rule doesn’t fully resolve the inquiry. Options:

Create a support ticket and assign to a human agent
Forward the conversation to the WhatsApp Inbox for manual handling
Notify a designated team member via email or dashboard alert

Conversational AI

AI Agents handle natural conversations automatically:

Create an AI agent with custom instructions (system prompt) defining its role, tone, and knowledge boundaries
Define the scope — what topics the agent can handle and when to escalate to a human
Provide context documents or knowledge base URLs the agent can reference
Integrate with IVR flows for voice interactions or chatbot for text interactions

Example AI Agent Configurations:

Use Case	Agent Name	Instructions
Appointment Booking	”SchedulerBot"	"You are a helpful scheduling assistant. Ask for preferred date/time, check availability, and confirm bookings.”
Order Status	”OrderTracker"	"You are an order tracking assistant. Ask for the order ID, look up its status, and provide shipping ETA.”
FAQ Handler	”HelpDesk"	"You answer common product questions using the provided knowledge base. Escalate to a human if the answer isn’t in the knowledge base.”

Chatbot Automation

The Chatbot automates lead capture and ticket creation:

Configure AI provider (OpenAI or Gemini) and provide the API key
Define conversation flows: greeting, qualification questions, information collection
Enable automatic ticket creation — conversations that cannot be resolved by the bot generate a support ticket with full transcript
Enable lead capture — bot-qualified leads are added to the CRM with conversation summary and tags

Example Chatbot Flow:


User: "Hi"
Bot: "Hello! I can help with pricing, product info, or technical support. What do you need?"
User: "pricing"
Bot: [Captures as "pricing inquiry" lead. Asks qualifying questions about company size and use case.]
Bot: "Thanks! I have all the info I need. A sales rep will reach out within 24 hours."
[Creates lead with tags: pricing, qualified, chatbot]

Automation Audit

Review Cadence

Automation rules drift over time as your business, products, and customer needs change. Schedule regular audits:

Frequency	What to Review
Weekly	Automation reply logs — spot-check 10 random interactions for appropriateness
Monthly	Keyword trigger effectiveness — identify new keywords customers use that are not covered
Quarterly	Full automation audit — review every active rule, update responses, remove unused rules
Annually	Strategy review — evaluate whether automation coverage aligns with business goals

Audit Checklist

Are all active automation rules still relevant?
Are reply templates up to date with current pricing, links, and policies?
Are there new common questions that should be automated?
Is the escalation path still correct? (Teams, agent assignments, etc.)
Are there rules that overlap or conflict with each other?
What percentage of conversations are fully resolved by automation vs. escalated to human agents?
Are AI agent responses accurate and within defined scope?

Best Practices

Start simple — automate one process at a time and validate before adding more
Test automation rules thoroughly with a test contact or internal number
Monitor automation effectiveness weekly through reply logs
Combine automation types for complex workflows (e.g., WhatsApp keyword trigger > AI agent response > fallback to support ticket)
Always have a human escalation path — no automation should be a dead end
Use clear, professional language in auto-replies — avoid overly casual or robotic tones
Tag conversations handled by automation for reporting and auditability

Example: Customer Support Automation

This flow handles the most common support scenario: a customer sends a message, automation attempts to resolve it, and only unresolved issues reach a human agent. This reduces agent workload by handling routine inquiries automatically while ensuring complex issues get proper attention.

Tips

Review automation performance weekly
Update rules based on common customer questions discovered in support tickets
Set clear boundaries for what automation handles vs. human agents
Log automation interactions so you can audit and improve rules over time
Keep automation reply text concise — long auto-replies feel impersonal and are often ignored