Agentic Email Agent (Lab Demo)

This repository contains a Jupyter Notebook demo from my Agentic AI Programming learning journey.
It demonstrates how to build an agentic workflow that can generate, evaluate, and send cold outreach emails automatically.

---

What this project demonstrates

• Agent Workflows → how different AI agents collaborate.
• Multi-Agent Collaboration → three sales agents with different personas (professional, engaging, busy).
• Tool Integration → using external tools like Resend to actually send emails.
• Decision Making → agents generate drafts, a manager agent selects the best one.
• Handoffs → selected email is passed to another agent for subject line, HTML formatting, and sending.

---

Why this project matters

Sales outreach is one of the most repetitive tasks in business.
This demo shows how Agentic AI can:

• Automate the generation of personalized sales emails.
• Allow multiple “agent personas” to contribute ideas.
• Choose the most effective email automatically.
• Format and send it seamlessly without human effort.

In real-world terms: this is a mini SDR (Sales Development Rep) powered by AI.

---

Project Structure

notebooks/
   └── agentic_email_demo.ipynb   # Main notebook demo

requirements.txt                  # Dependencies
README.md                         # Documentation

---

Requirements

Install dependencies (adjust based on your environment):

pip install -r requirements.txt

Create a .env file:

RESEND_API_KEY=your_api_key_here

---

Usage

1. Open the notebook notebooks/agentic_email_demo.ipynb in Jupyter or VS Code.
2. Run the cells step by step.
3. The notebook will:
   • Generate multiple cold emails.
   • Pick the best one.
   • Format it into HTML.
   • Send it to your inbox (via Resend).

---

Example Workflow

1. Three agents generate email drafts.
2. Manager agent compares them and selects the best.
3. Email manager agent:
   • Writes a subject line.
   • Converts to HTML.
   • Sends the email.

All automatically ⚡

---

Tags

agentic-ai · ai-agents · multi-agent-systems · cold-email · sales-automation · resend · openai · python · jupyter-notebook

---

Notes

This repository is for educational purposes.
It’s part of my ongoing work as an Agentic AI Engineer, showcasing how agent frameworks and tool integration can solve real-world business problems.

Future improvements could include:

• Connecting to CRMs (HubSpot, Salesforce).
• Adding personalization with RAG pipelines.
• Multi-channel outreach (email, LinkedIn, SMS).