🚀 Multi-Agent Marketing Campaign Creator

Automated workflow using AI agents: Researcher → Copywriter → Art Director → Manager

This project is a multi-agent AI workflow automator that generates a complete marketing campaign brief from a simple product description.
Using Groq’s OpenAI-compatible LLM API, each specialized agent performs a different role—just like a real marketing team:

• 🧠 Research Agent – analyzes trends, audience, insights
• ✍️ Copywriter Agent – writes taglines, messages, headlines, body copy
• 🎨 Art Director Agent – creates visual concepts & image prompts
• 📋 Manager Agent – assembles everything into a structured final brief

This project demonstrates agent orchestration, workflow automation, modular design, and real-world LLM application development.

---

🧩 Features

✔ Automated end-to-end AI workflow
✔ 4 specialized AI agents collaborating sequentially
✔ Generates:

• Research summary
• Positioning angles
• Tagline + campaign messaging
• Headlines + ad copy
• Image prompts (for DALL·E, Stable Diffusion, etc.)
• Final campaign brief (Markdown)

✔ Uses Groq LLMs (fast & free-tier friendly)
✔ Uses OpenAI-compatible API calls
✔ Modular, readable Python architecture
✔ CLI interface for interactive use

---

📂 Project Structure

multi_agent_workflow/
│
├── main.py                     # CLI entry point
├── requirements.txt
├── .env                        # Contains GROQ_API_KEY (not pushed to GitHub)
│
├── src/
│   ├── config.py               # Settings, model names, API keys
│   ├── llm_client.py           # Groq API client
│   ├── workflow.py             # Orchestrates all 4 agents
│   │
│   └── agents/
│       ├── research_agent.py       # Audience insights & research
│       ├── copywriter_agent.py     # Tagline, headlines, messaging, copy
│       ├── art_director_agent.py   # Image prompts & visual concepts
│       └── manager_agent.py        # Final campaign brief assembly
│
└── README.md

▶️ Usage

Run the program: python main.py

It will ask for:

• Product name
• Product description
• Target audience
• Campaign goal
• Tone
• Channels (e.g., instagram, tiktok, email)

Example

• Product name: EcoSip Reusable Bottle
• Product description: Stylish insulated bottle that keeps drinks cold 24 hours
• Target audience: young professionals who care about sustainability
• Campaign goal: drive online sales for summer collection
• Tone: fresh, energetic, eco-friendly
• Channels: instagram, tiktok, email

Output Includes

✔ Research insights ✔ Copywriting content (tagline, headlines, body copy, CTA) ✔ Image prompts (3–5 visual concepts) ✔ A polished final campaign brief

🧠 How the Workflow Works

• Research Agent: Generates insights, pain points, trends, and positioning angles.
• Copywriter Agent: Creates messaging including tagline, headlines, and body copy.
• Art Director Agent:Produces 3–5 detailed prompts for AI image generators.
• Manager Agent: Combines all previous outputs into a final campaign brief in Markdown.

🎯 Why This Project Matters

This project demonstrates:

• Multi-agent orchestration
• Practical workflow automation
• Prompt engineering for specialized agent roles
• Modular Python project structure
• LLM-powered content generation
• Realistic simulation of a professional marketing workflow
• Great for showcasing AI engineering, automation, creative pipelines, and multi-agent systems.

🛠️ Future Enhancements

• Add JSON schema validation for agent outputs
• Optional PDF export of final brief
• Streamlit or Gradio web UI
• Plugin for automatic image generation (DALL·E, SDXL, Flux)
• Save outputs to /outputs/ directory

Author

Created by Syed Waleed Ahmed