Skip to main content

​
Quickstart Guide

This guide walks you through launching your first AI agent using the Zild Platform. By the end of this guide, you will:
  • Create a Tenant environment
  • Configure a communication channel (App)
  • Create and configure an Agent
  • Connect a Workflow
  • Test your Agent
  • Deploy to production

​
Prerequisites

Before starting, ensure you have:
  • Access to a Zild Tenant
  • Admin permissions
  • A WhatsApp, Voice, or Webchat channel available
  • API credentials (if integrating external systems)

​
Step 1 — Create or Access Your Tenant

If your organization is new to Zild:
  1. Log in to the Zild Admin panel
  2. Access your dedicated Tenant environment
  3. Confirm your organization settings
Each Tenant is isolated and contains its own:
  • Apps
  • Agents
  • Workflows
  • Conversations
  • Integrations

​
Step 2 — Configure a Channel (App)

An App connects an external communication channel to Zild.

​
Option A — WhatsApp

  1. Navigate to Apps
  2. Click Create App
  3. Select WhatsApp
  4. Configure:
    • Webhook URL
    • Authentication
    • Phone number
  5. Validate webhook connection

​
Option B — Voice (SIP / Twilio)

  1. Create a Voice App
  2. Configure:
    • Inbound webhook endpoint
    • Voice number
    • Call routing rules
  3. Test inbound call event delivery

​
Option C — Webchat

  1. Create a Webchat App
  2. Configure domain settings
  3. Embed the provided script in your website
Once configured, the App will start sending inbound events to Zild.

​
Step 3 — Create an Agent

Now you will configure your first AI Agent.
  1. Navigate to Agents
  2. Click Create Agent
  3. Define:

​
System Instructions

Describe:
  • Role (Support, Sales, Assistant, etc.)
  • Tone
  • Constraints
  • Escalation behavior

​
Knowledge

Add:
  • FAQs
  • Policies
  • Product documentation
  • Contextual business information

​
Tools (Optional)

Configure tools for:
  • CRM lookup
  • Lead creation
  • Contract generation
  • API calls

​
Escalation Rules

Define when:
  • A human agent should take over
  • A conversation should be transferred
Save the Agent configuration.

​
Step 4 — Link App to Agent

After creating the Agent:
  1. Open your App configuration
  2. Select the default Agent
  3. Save changes
Now inbound messages will route to your Agent.

​
Step 5 — Configure a Workflow (Optional but Recommended)

To automate business actions:
  1. Navigate to Workflows
  2. Create a new Workflow
  3. Choose a trigger:
Examples:
  • On Message Received
  • On Lead Qualified
  • On Conversation Closed
  • On Tool Execution
  1. Define actions:
  • Send webhook
  • Update CRM
  • Create task
  • Assign conversation
  • Send notification
Save and activate the Workflow.

​
Step 6 — Test in Sandbox

Before going live:
  1. Use a test phone number or test webchat
  2. Send sample messages
  3. Validate:
  • Agent responses
  • Tool execution
  • Workflow triggers
  • Webhook payloads
  • Conversation state
Check logs and event history to verify correct execution.

​
Step 7 — Monitor Conversations

Navigate to Conversations to:
  • View message history
  • Inspect metadata
  • Monitor agent decisions
  • Review workflow execution
  • Escalate manually if needed
This allows operational visibility before scaling.

​
Step 8 — Go Live

When testing is validated:
  1. Switch channel configuration to production
  2. Confirm webhook endpoints are secured
  3. Enable monitoring alerts (if configured)
Your AI Agent is now live.

​
Recommended Next Steps

After launching your first agent:
  • Implement conversation monitoring
  • Add performance metrics
  • Expand to additional channels
  • Create specialized agents (Sales, Support, Voice)
  • Integrate CRM and internal systems

​
Deployment Checklist

Before production deployment, confirm:
  • Tenant configuration validated
  • Apps connected and authenticated
  • Agent instructions reviewed
  • Escalation rules defined
  • Workflows tested
  • Webhooks secured
  • Access control configured

​
Need More Advanced Setup?

Continue with:
  • Platform Overview — Infrastructure details
  • API Reference — Programmatic control
  • Security Documentation — Compliance and isolation model
You now have a fully operational AI agent running on Zild.