Skip to main content

Setup and Details

This guide walks you through the complete process of creating, training, and deploying your AI Knowledge Base Agent on WhatsApp.

Overview

AI Agents by Eazybe enable businesses to build intelligent, automated agents that operate directly on WhatsApp, trained using your own business data. These agents are designed to:
  • Handle customer conversations automatically
  • Assist in sales and support workflows
  • Learn from your existing knowledge and chat history
  • Deliver human-like responses based on real interactions
This feature represents the foundation of advanced AI agents. Businesses can later unlock deeper personalization, CRM integrations, and fully customized automation systems.

What Makes This Powerful

Unlike traditional chatbots, Eazybe AI Agents are:

Context-Aware

They learn from:
  • Your website or knowledge base
  • Your actual WhatsApp conversations

WhatsApp-Native

  • Built specifically for WhatsApp communication
  • No need for external chat tools
  • Direct integration with your existing workflow

Fast to Launch

  • Create → Train → Deploy in under 5 minutes

Continuously Scalable

  • Start simple
  • Upgrade to fully customized agents with deeper integrations

Common Use Cases

Limitations (Initial Version)

This is an early version with the following considerations:
  • Limited deep customization (available in advanced version)
  • Performance depends on quality of training data
  • Requires WhatsApp API number for deployment

Step-by-Step: Create Your AI Agent

Follow these steps to set up and deploy your AI Knowledge Base Agent:
1

Access AI Agents

Navigate to the AI Agents section in your dashboard and click on Create another agent or Add Agent.AI Agents dashboard link
2

Add Agent

Click on Add Agent in your dashboard to begin creating your AI agent.Add Agent button
3

Add Knowledge Source

Choose one of the following methods to train your agent:
  • Website URL – Enter your website URL to scrape content
  • Upload Knowledge Base – Upload documents (PDF, DOCX, TXT, CSV)
  • Connect Existing Knowledge Base – Link to your current resources
Add knowledge source optionsUpload or connect knowledge base
This helps the agent understand your business and provide accurate responses.
4

Select Use Case

Choose the primary purpose of your agent:
  • Customer Support (CS) – Handles queries, complaints, FAQs
  • Sales Agent – Converts leads, shares product info, follows up
Select use case optionsClick Next Step to continue.
5

Sync WhatsApp Chats (Most Important)

Enable Sync WhatsApp Chats to allow the agent to learn from your historical conversations. This is a critical step that enables the agent to:
  • Learn real customer tone
  • Understand frequently asked questions
  • Replicate your team’s communication style
Sync WhatsApp chats optionSelect conversations to learn fromSelect the conversations you want the agent to learn from and click Continue.
6

Configure AI Model

Click on AI Brain and select your AI agent model. Choose the appropriate model based on your requirements and expected conversation volume.AI Brain model selection
7

Select Agent Type

Confirm the role of your agent:
  • Customer Support (CS) – Handles queries, complaints, FAQs
  • Sales Agent – Converts leads, shares product info, follows up
Agent type selection
8

Select Deployment Number

Choose the WhatsApp number you want to deploy the AI agent on. This should be a connected WhatsApp Business API (WABA) number.Select deployment number
9

Customize System Instructions

Edit the System Instructions based on your specific use cases. This defines how the agent behaves and responds to customer queries.System instructions editor
10

Finalize Training

Confirm your selected data sources and click Launch Agent. The system will automatically train your agent.Launch agent button
Training typically takes just a few minutes. Do not close the browser during this process.
11

Test in Playground

Click on Playground to interact with your agent before going live. Use this opportunity to:
  • Validate response quality
  • Test edge cases
  • Make adjustments to system instructions if needed
Playground testing interfaceSources knowledge review
12

Review WhatsApp Number

Click on WhatsApp to verify the connected number.Connected WhatsApp number
13

Review Billing

Click on Billing to review your plan.Billing information
14

Publish Your Agent

Click Publish to make your AI agent live.Publish agent button

After Deployment

Once deployed:
1

Live Operation

Your AI agent is now active and will respond to customer messages on the selected WhatsApp number.Agent live dashboard
2

Internal Testing

Your team can also interact with the agent internally to monitor performance.
3

Continuous Learning

The agent responds instantly using its trained knowledge base. Monitor conversations and re-sync data sources as needed to keep the agent up-to-date.

Advanced Custom AI Agents

For businesses needing deeper customization, we offer fully tailored AI agents:
Advanced custom agents include:
  • Built using your CRM data
  • Customized to your specific workflows
  • Designed with your leadership team
  • Delivered within 4 weeks
Features include deep personalization, business-specific logic, and advanced automation flows.
Contact our team to discuss building a custom AI solution for your business. Advanced custom AI agents overview