ET Intelligent Cost Optimization

An autonomous AI cost-control platform that goes beyond dashboards by monitoring enterprise operations data, detecting inefficiencies, and surfacing corrective actions with measurable financial impact.

---

Project Overview

This project brings together multiple enterprise agents inside one premium web experience:

• Analyze blended cost signals across spend leakage and infrastructure waste
• Detect duplicate payments and vendor anomalies through Spend Intelligence
• Recommend shutdown or consolidation actions through Resource Optimization
• Prevent operational penalties with SLA Guard
• Surface reconciliation mismatches through Financial Ops
• Track quantified monthly and yearly impact
• Maintain an activity timeline through continuous monitoring
• Route action proposals through an approval-driven Action Queue

---

Key Features

• Multi-agent architecture across spend, infrastructure, SLA, and finance workflows
• AI-assisted optimization recommendations with fallback rule-based reasoning
• Continuous monitoring loop with automated activity logging
• Approval workflow for action proposals
• Financial impact surfaced across every major module
• Premium editorial UI inspired by luxury enterprise website layouts
• Deployment-ready sample datasets included in the repository

---

Agent Suite

1. Analyze

• Combines spend and resource insights into one overview
• Highlights duplicate-payment exposure and infrastructure waste
• Summarizes corrective opportunities in one executive view

2. Spend Intelligence

• Detects duplicate payments and suspicious spend patterns
• Estimates recoverable duplicate-payment exposure
• Surfaces affected vendors and transaction clusters

3. Resource Optimization

• Identifies idle or underutilized infrastructure
• Suggests shutdown or consolidation actions
• Estimates monthly waste and savings opportunity

4. SLA Guard

• Flags high-risk and breached SLA tickets
• Suggests escalation or rerouting actions
• Quantifies penalty exposure before the financial hit lands

5. Financial Ops

• Detects reconciliation mismatches
• Surfaces variance exposure and likely root causes
• Suggests corrective finance actions

6. Action Queue

• Aggregates actions from multiple agents
• Supports approval-driven execution flow
• Persists approval state across restarts

7. Activity and Impact

• Logs automated monitoring runs in the Activity timeline
• Tracks projected monthly and yearly value in Impact

---

Tech Stack

Frontend

• React
• Axios
• Lucide React
• Custom CSS

Backend

• FastAPI
• Pandas
• Groq API
• Uvicorn

Deployment

• Frontend: Vercel
• Backend: Render

---

Project Structure

et-intelligent-cost-optimization/
|-- backend/
|   |-- agents/
|   |-- llm/
|   |-- action_builders.py
|   |-- action_engine.py
|   |-- action_factory.py
|   |-- audit.py
|   |-- main.py
|   |-- monitoring.py
|   `-- workflow.py
|-- data/
|   |-- finance_reconciliation_sample.csv
|   |-- resource_usage_azure_sample.csv
|   |-- sla_operations_sample.csv
|   `-- transactions_sf_vouchers_sample.csv
|-- frontend/
|   |-- public/
|   `-- src/
|-- DEPLOYMENT.md
|-- README.md
`-- requirements.txt

---

Live Deployment

• Frontend: https://et-intelligent-cost-optimization.vercel.app
• Backend: https://cost-ai-backend.onrender.com
• Backend Docs: https://cost-ai-backend.onrender.com/docs

Note: the backend may take a few seconds to wake up on free hosting.

---

Getting Started

1. Clone the repository

git clone https://github.com/rivu-intel45/et-intelligent-cost-optimization.git
cd et-intelligent-cost-optimization

2. Set up the backend

python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt

Create a backend environment variable:

GROQ_API_KEY=your_groq_api_key_here

Run the backend:

uvicorn backend.main:app --reload

3. Set up the frontend

cd frontend
npm install

Create frontend/.env:

REACT_APP_API_BASE_URL=http://127.0.0.1:8000

Run the frontend:

npm start

---

Datasets Used

This project uses cleaned, deployment-safe sample datasets derived from larger public or real-world-inspired sources:

• Azure resource utilization traces for infrastructure optimization
• San Francisco vendor payments for spend intelligence
• Helpdesk and SLA operations data for breach-risk monitoring
• Finance reconciliation data for variance and mismatch analysis

Large raw files are intentionally excluded from GitHub, while smaller deployable sample files are included in the repository.

---

Continuous Monitoring

The backend includes an automated monitoring loop that periodically:

• checks duplicate-payment patterns
• scans for idle resources
• evaluates SLA risk
• reviews finance discrepancies
• logs findings into the Activity timeline

This helps the app behave like an autonomous operations intelligence platform rather than a static dashboard.

---

Deployment Notes

• Backend dependencies are listed in requirements.txt
• Frontend production API configuration uses frontend/.env.production
• Approval state is persisted in data/actions_db.json
• Required backend sample CSV files are validated on startup

See DEPLOYMENT.md for deployment details.

---

Future Improvements

• Move approval persistence from JSON storage to a database
• Add authentication and user roles
• Add richer executive reporting and charting
• Expand downstream workflow automation
• Tighten production CORS and infrastructure hardening

---

Author

Built by Rivu Intel45 as an enterprise AI cost optimization project focused on autonomous monitoring, financial impact, and action-driven operations intelligence.