AI: A New Partner in Our Everyday Journey

Artificial Intelligence is the science of making machines think and act like humans AI is all about building machines that can think, learn, and solve problems like humans. The purpose of AI is not to replace humans but to assist us in saving time,improving decisions, and handling routine work so we can focus on Creativity and Innovation

As the quote goes,“AI is not here to replace the humans, but to work with us and make us better”

Do You Know How Machines Learn Like Us?

Deep Learning

Deep learning is a special branch of Artificial Intelligence that teaches machines to learn from data, just like humans learn from experience.It uses a structure called neural network, inspired by how our brain’s neurons work.

Think of AI as a student. A student learns by studying examples, practicing, and correcting mistakes Machines learn in a very similar way. This learning process is what makes AI smart.

There are three main types of machine learning technology that guide AI.

Supervised Learning

In this method, machines learn from labeled data —> kind of like a teacher giving a correct answer in advance

Example: If you give a computer 1000 photos labeled “Tiger” and “Lion”, it will learn the different and then correctly identify a photo

Unsupervised Learning

Here, there’s no teacher. The machine tries to find the pattern on its own

Example: If you feed a program a lot of customer shopping data, it may group customers with similar buying habits without anyone telling it what the groups mean.

Reinforcement Learning

In this method, Machines learn by trial and error, receiving rewards for correct actions and penalties for mistakes — like learning to play a game.

Example: Training a robot to walk, teaching AI to play chess, or helping self-driving cars make safe driving decisions.

As the saying goes, “Machines are fast,but humans are wise”

Manual Workflow for CRUD Operations in Agile CRM

What is Agile CRM

Agile CRM is a cloud-based platform that helps Businesses manage their interactions with Customers and leads, and in simple terms , it helps companies work smarter ,save time ,and build stronger Relationships with Customers.

How to Get Agile API Credentials and Access Token

Notes: The purpose of this guide is to show how to generate Agile API credentials and access tokens , which are required to fetch information from Agile using APIs. Without an access token, external applications cannot securely connect to Agile crm or retrieve data

Step 1: Developer documentation URL : https://github.com/agilecrm/rest-api?utm_source=chatgpt.com

Click your profile (top-right) → Admin Settings → Developers & API → copy REST API Key (use the REST key for server REST calls)

Authentication :

This is an HTTPS-only API. Authentication will be performed based on the email address of the user and their respective API Key.

Email and API key should be passed via HTTP Basic Authentication. Email address as username and their respective API Key as password.

If you want more details about this , refer to the developer documentation I mentioned Below

Developer documentation URL : https://github.com/agilecrm/rest-api ?

Step 2:

Set the postman setup

make it all method in Postman URL: https://{domain}.agilecrm.com/dev/

Go to the contact and move to the Authorization, make it Auth Type set to Basic Auth

And "previous" - referring to getting the Rest API token from the previous steps, where you generated the Rest API token

Click the send

You can get the response

We performed CRUD operations using a UI and Python code

Create operation: How to create a Contact using Agile UI

In Agile CRM, click contacts on the left-hand navigation bar to open the contact module and then click add contact on the top right corner to add a new lead.

In Agile , the Lead module is used to capture customer information such as name , phone number , last name, and company name

Create operation: How to create a Contact using Python code

Now, let’s implement a POST function in Python to Create a new lead in Zoho CRM.

import requests
import base64

def create_agilecrm_contact():
    #  Replace with your Agile CRM details
    domain = "YOUR_DOMAIN"
    email = "YOUR_EMAIL"
    api_key = "YOUR_API_KEY"

    # Step 1: Build Auth Header
    base_url = f"https://{domain}.agilecrm.com/dev/api"
    auth_str = f"{email}:{api_key}"
    auth_header = {
        "Authorization": "Basic " + base64.b64encode(auth_str.encode()).decode(),
        "Accept": "application/json",
        "Content-Type": "application/json"
    }

    # Step 2: Contact data (example)
    contact_data = {
        "first_name": "Arun",
        "last_name": "Kumar",  # REQUIRED
        "email": "arun.k@example.com",
        "company": "XYZ Pvt Ltd",
        "phone": "+91-9876543210"
    }

    # Step 3: Build payload
    payload = {
        "properties": [
            {"type": "SYSTEM", "name": "first_name", "value": contact_data.get("first_name", "")},
            {"type": "SYSTEM", "name": "last_name", "value": contact_data["last_name"]},
            {"type": "SYSTEM", "name": "email", "value": contact_data.get("email", "")},
            {"type": "SYSTEM", "name": "company", "value": contact_data.get("company", "")},
            {"type": "SYSTEM", "name": "phone", "value": contact_data.get("phone", "")}
        ]
    }

    # Step 4: Send POST request
    url = f"{base_url}/contacts"
    response = requests.post(url, headers=auth_header, json=payload)

    # Step 5: Handle response
    if response.status_code in (200, 201):
        try:
            return response.json()
        except Exception:
            return {"message": " Success (no JSON)"}
    else:
        raise Exception(f" API Error: {response.status_code} - {response.text}")


#  Example usage
if __name__ == "__main__":
			print(" Created Contact:", create_agilecrm_contact())

Get operation: How to get a Contact using Agile UI

Click Contact on the left hand Navigation Bar to view all the Contacts

Get operation: How to get a Contact using Python code

Let’s create a GET function in Python to retrieve leads from Agile CRM.

import requests
import base64

def get_all_agilecrm_contacts():
    #  Replace with your Agile CRM credentials
    domain = "YOUR_DOMAIN"
    email = "YOUR_EMAIL"
    api_key = "YOUR_API_KEY"

    # Step 1: Auth Header
    base_url = f"https://{domain}.agilecrm.com/dev/api"
    auth_str = f"{email}:{api_key}"
    auth_header = {
        "Authorization": "Basic " + base64.b64encode(auth_str.encode()).decode(),
        "Accept": "application/json",
        "Content-Type": "application/json"
    }

    # Step 2: API request
    url = f"{base_url}/contacts"
    response = requests.get(url, headers=auth_header)

    # Step 3: Handle response
    if response.status_code == 200:
        return response.json()
    else:
        raise Exception(f" API Error: {response.status_code} - {response.text}")


# Example usage
if __name__ == "__main__":
    print("All Contacts:", get_all_agilecrm_contacts())

Update operation: How to update a Contact using Agile UI

Click contacts on the left-hand navigation bar, and click on one contact

To update a contact, click on the required contact from the screen below.

In this screen, update the contact of your choice

Update operation: How to update the property or information using Python code

Let’s create an update function in Python to modify an existing contact in Agile CRM.

import requests
import base64

def update_agilecrm_contact(contact_id: str, updated_data: dict):
    #  Replace with your Agile CRM credentials
    domain = "YOUR_DOMAIN"
    email = "YOUR_EMAIL"
    api_key = "YOUR_API_KEY"

    if not updated_data:
        raise ValueError(" Error: updated_data cannot be empty.")
    
    if not contact_id:
        raise ValueError(" Error: 'contact_id' is required to update a contact.")
    
    if not isinstance(contact_id, str):
        raise ValueError(" Error: 'contact_id' must be a string inside double quotes.")

    # Step 1: Auth Header
    base_url = f"https://{domain}.agilecrm.com/dev/api"
    auth_str = f"{email}:{api_key}"
    auth_header = {
        "Authorization": "Basic " + base64.b64encode(auth_str.encode()).decode(),
        "Accept": "application/json",
        "Content-Type": "application/json"
    }

    # Step 2: Prepare Payload
    payload = {
        "id": contact_id,
        "properties": [
            {"type": "SYSTEM", "name": k, "value": v}
            for k, v in updated_data.items()
        ]
    }

    # Step 3: API Request
    url = f"{base_url}/contacts/edit-properties"
    response = requests.put(url, headers=auth_header, json=payload)

    # Step 4: Handle Response
    if response.status_code in (200, 201, 204):
        try:
            return response.json()
        except Exception:
            return {"message": " Contact updated successfully"}
    else:
        raise Exception(f" API Error: {response.status_code} - {response.text}")


#  Example usage
if __name__ == "__main__":
    contact_id = "1234567890"  # Replace with actual Agile CRM contact ID
    updated_data = {"first_name": "Ramesh", "company": "XYZ Pvt Ltd"}
    print("Updated Contact:", update_agilecrm_contact(contact_id, updated_data))

Delete operation: How to delete a Contact using Agile UI

Click contacts on the left hand Navigation Bar, and click on one contact . Click the three dots and select delete

Delete operation: How to delete a Contact using Python code

Let’s Create a delete function in Python to remove a Contact from Agile CRM.

import requests
import base64

def delete_agilecrm_contact(contact_id: str):
    #  Replace with your Agile CRM credentials
    domain = "YOUR_DOMAIN"
    email = "YOUR_EMAIL"
    api_key = "YOUR_API_KEY"

    if not contact_id:
        raise ValueError(" Error: 'contact_id' is required to delete a contact.")
    
    if not isinstance(contact_id, str):
        raise ValueError(" Error: 'contact_id' must be a string inside double quotes.")

    # Step 1: Auth Header
    base_url = f"https://{domain}.agilecrm.com/dev/api"
    auth_str = f"{email}:{api_key}"
    auth_header = {
        "Authorization": "Basic " + base64.b64encode(auth_str.encode()).decode(),
        "Accept": "application/json",
        "Content-Type": "application/json"
    }

    # Step 2: API Request
    url = f"{base_url}/contacts/{contact_id}"
    response = requests.delete(url, headers=auth_header)

    # Step 3: Handle Response
    if response.status_code in (200, 204):
        return {"message": f" Contact {contact_id} deleted successfully"}
    else:
        raise Exception(f" API Error: {response.status_code} - {response.text}")


#  Example usage
if __name__ == "__main__":
    contact_id = "1234567890"  # Replace with an actual Agile CRM contact ID
    print(delete_agilecrm_contact(contact_id))

Turning Python Code into an MCP Server

What is the MCP Framework

The Model Context Protocol (MCP) is an open framework designed to help AI assistants (like ChatGPT or other LLMs) Interact with external tools , APIs , and data sources in a secure and standardized way.

If you’re building an AI agent in Python ,you can use the MCP framework

Expose your custom Python functions as MCP tools

Let AI models call those functions via MCP

Connect your AI assistant to external APIs (CRM,Databases,Cloud Services)

Example: Simple MCP Server in Python

import asyncio
from mcp.server.fastmcp import FastMCP

# Create MCP server instance
mcp = FastMCP("example-server")

# Register a simple tool
@mcp.tool()
async def greet(name: str) -> str:
    """Say hello to someone"""
    return f"Hello, {name}! 

# Start MCP server
if __name__ == "__main__":
    asyncio.run(mcp.run())

Here’s what happens:

We created an MCP server called example-server

Defined a greet function as a tool

Any AI assistant that supports MCP can now call this tool

📂 A typical MCP folder structure to follow

agile-crm-mcp
│
├── .env  (Environment variables eg:AGILE_DOMAIN, AGILE_EMAIL, AGILE_API_KEY, etc.)
├── requirements.txt  (Project dependencies eg:MCP SDK, requests, dotenv, etc.)
│
├── leads.py  (Handles Agile CRM Contact operations CRUD APIs)
├── agile_crm.py  (Core Agile CRM API integration logic)
│
├── agile_mcp_server.py  (MCP server exposing Agile CRM as MCP tools/resources)
├── agile_mcp_client.py  (MCP client to interact with the MCP server)

Please refer to the image below for understanding the MCP folder structure

I’ll outline the workflow and show how the modules interact at runtime.

Architecture Overview

This is an Agile CRM MCP (Model Context Protocol) Server that provides natural language interface to Agile CRM operations through AI-powered routing.

File Structure & Responsibilities

1. agile_crm.py - Core CRM API Client

Purpose: Base HTTP client for Agile CRM API interactions

Key Features:

Handles authentication using Basic Auth with base64 encoding

Provides HTTP methods: get(), post(), put(), delete()

Centralized response handling with error management

Constructs API URLs dynamically based on the domain

2. leads.py - Contact Management Layer

Purpose: High-level contact operations wrapper

Key Features:

Contact CRUD operations (Create, Read, Update, Delete)

Data validation (requires last_name for contacts)

Search functionality by email and name

Property mapping to Agile CRM's system format

Fallback search mechanisms for name-based queries

3. Agile_mcp_server.py - MCP Server Implementation

Purpose: Exposes CRM operations as MCP tools

Key Features:

Six MCP tools: get_contacts, get_contact, create_contact, update_contact, delete_contact, get_contact_by_email, get_contact_by_name

Environment-based configuration (domain, email, API key)

JSON serialization for all responses

FastMCP framework integration

4. Agile_mcp_client.py - AI-Powered Client

Purpose: Natural language interface using OpenAI LLM

Key Features:

LLM-based question routing to appropriate MCP tools

OpenAI integration with configurable models

Command-line interface for natural language queries

Automatic tool selection and argument extraction

5. requirements.txt - Dependencies

MCP framework and CLI tools

OpenAI client library

Example CLI usage

python Agile_mcp_client.py --question "Create a contact named Jane Smith with email jane@company.com"

AI: A New Partner in Our Everyday Journey

Do You Know How Machines Learn Like Us?

Deep Learning

Supervised Learning

Unsupervised Learning

Reinforcement Learning

Manual Workflow for CRUD Operations in Agile CRM

How to Get Agile API Credentials and Access Token

Set the postman setup

We performed CRUD operations using a UI and Python code

Create operation: How to create a Contact using Agile UI

Create operation: How to create a Contact using Python code

Get operation: How to get a Contact using Agile UI

Get operation: How to get a Contact using Python code

Update operation: How to update a Contact using Agile UI

Update operation: How to update the property or information using Python code

Delete operation: How to delete a Contact using Agile UI

Delete operation: How to delete a Contact using Python code

Turning Python Code into an MCP Server

📂 A typical MCP folder structure to follow

Please refer to the image below for understanding the MCP folder structure

Our product already solves this problem. Instead of spending time on implementation, you can start using our product right away.

Learn to build your own AI agent to automate Agile CRM workflow