🧠 CampaignMind

AI-Powered Marketing Brief Enhancement Platform

CampaignMind transforms incomplete marketing briefs into comprehensive, data-driven campaign strategies using agentic AI agents with multi-step reasoning and real-time competitive intelligence.

---

🌟 Overview

CampaignMind is a production-ready multi-agent AI system that analyzes marketing campaign briefs and enriches them with:

• Trend Intelligence - Current digital & social marketing trends
• Competitive Analysis - Real-time competitor campaign research
• Market Landscape - Strategic positioning and whitespace opportunities
• Strategic Insights - Actionable recommendations synthesized from all data sources

What Makes CampaignMind Different?

Unlike standard RAG (Retrieval-Augmented Generation) systems, CampaignMind uses Agentic RAG:

Standard RAG	Agentic RAG (CampaignMind)
Single-shot query	Multi-step search strategies
Fixed process	Adaptive reasoning & planning
One data source	RAG + Web Search + Scraping
No self-correction	Quality evaluation & retry logic
Generic outputs	Context-aware, professional analysis

---

✨ Key Features

🤖 Agentic AI Specialists

Each agent uses sophisticated multi-step reasoning:

1. 🧾 Brief Analysis & Metadata Gating

   • Identifies gaps and determines enhancement strategy and forces required metadata confirmation before running specialist agents\

2. Choosing the execution mode

   • Works in interactive or autonomous mode

3. 🔍 Specialist Agent Pipeline

   • Trend Agent: Searches RAG → Web → Google Trends API for current data
   • Case Intelligence Agent: Finds competitor campaigns via RAG → Web → Scraping
   • Market Landscape Agent: Analyzes competitive positioning with real-time research
   • Insight Generator: Synthesizes all findings into strategic recommendations
   • Creator Agent: Generate hero ad concept ideas, tagline options and week by week calendar

4. 🗂️ Three‑Panel UI

   • Agents Team (left)
   • User Chat (center)
   • Agents Chat (right)

5. 👤 Guest & Auth Modes

   • Guest mode with browser storage
   • Supabase-authenticated mode for persistence

---

🛠️ Technologies & Tools Used

Frontend Stack

• Next.js (App Router)
• TypeScript
• Tailwind CSS
• Vercel (deployment)

Backend Stack

• Python 3.11
• FastAPI
• Uvicorn
• Pydantic
• Render (deployment)

AI / RAG

• Gemini API (OpenAI‑compatible endpoint)
• Local Vector DB for RAG
• Optional agentic tools: web search, scraping, trends

Data & Auth

• Supabase (Postgres + Auth)

---

🏗️ Architecture & Design

System Architecture

┌─────────────────────────────────────────────┐
│  Frontend (Next.js + Vercel)                │
│  - Three‑panel UI                           │
│  - Brief input + workflow control           │
└───────────────────────┬─────────────────────┘
                        │ HTTPS
                        ▼
┌─────────────────────────────────────────────┐
│  Backend (FastAPI + Render)                 │
│  - /api/v1/conversations                    │
│  - /api/v1/brief-analyze                    │
│  - Agent orchestration                      │
│  - RAG + Web tools                          │
└───────────────┬───────────────┬─────────────┘
                │               │
                ▼               ▼
        Supabase (DB/Auth)   Gemini API

Key Techniques

1. Metadata‑gated workflow before agent execution
2. Step‑by‑step or autonomous orchestration
3. Structured outputs for UI rendering
4. RAG + web search for enriched insights
5. Guest mode persistence (IndexedDB + localStorage)

---

💡 AI Techniques Implemented

1. Prompt Engineering

• Structured prompts per agent role
• JSON/Markdown output control
• Guidance routing based on brief completeness

2. Retrieval‑Augmented Generation (RAG)

• Chunked knowledge base
• Vector similarity search for context injection

3. Multi‑Agent Orchestration

• Dedicated roles for trends, cases, market, synthesis, creative
• Pipeline control with metadata and mode selection

---

📈 Technical Achievements

Performance

• Multi‑agent pipeline with step‑level outputs
• On‑demand execution in interactive mode

Scalability

• Stateless API design
• Supabase‑backed persistence for authenticated users

Reliability

• Required metadata enforcement
• Clear error handling and fallbacks

---

🧭 Execution Flow

1. Brief Analyzer
2. Metadata Confirmation
3. Trend Agent
4. Case Intelligence Agent
5. Market Landscape Agent
6. Insight Generator
7. Creator Agent

---

⚠️ Common Issues

• CORS errors: ensure FRONTEND_URL / FRONTEND_URLS match deployed frontend domain
• Supabase 500: check SUPABASE_SERVICE_KEY
• Cold start delays: Render free tier sleeps

---

🔮 Future Enhancements

• User workspaces + multi‑project management
• Saved campaign libraries
• Multi‑language output support
• Team collaboration + sharing

---

📜 License & Usage

© 2026 Su Phyu Sin Htet and Hein Thu Aung. All Rights Reserved.

This CampaignMind project was collaboratively developed by Su Phyu Sin Htet and Hein Thu Aung.
It is protected as original intellectual property.

You may view, read, and cite this repository for educational or research purposes only.
Any reproduction, redistribution, modification, or commercial use of this project, its source code, trained models, or design assets without prior written consent from both authors is strictly prohibited.

For collaboration or licensing discussions, please contact the authors directly.

---

🙏 Acknowledgments

• Microsoft AutoGen - Agent orchestration framework
• Google Gemini - LLM capabilities
• ChromaDB - Vector database
• Supabase - Database and authentication
• FastAPI - Modern Python web framework

Benchmarks

• Brief Analysis: ~3-5 seconds
• Single Agent (Standard RAG): ~6-10 seconds
• Single Agent (Agentic RAG): ~12-22 seconds
• All 5 Agents (Parallel): ~55-85 seconds
• Quality Score Improvement: 60% → 94%+

Quality Metrics

Metric	Standard RAG	Agentic RAG
Avg Quality Score	0.45	0.94
Data Freshness	Static	Real-time
Competitor Coverage	2-3	5-10
Quantitative Data	Rare	Consistent
Self-Correction	None	Automatic

You can test the project on this link

📧 Contact

Project Maintainer: Su Phyu Sin Htet
GitHub: @suphyusinhtet
Email: suphyusinhtet@gmail.com