AdShield Intelligence Platform

AdShield Intelligence: The Fraud-Proof Campaign Optimization Engine

Inspiration

Digital marketers spend over 80% of their work week manually pulling reports across Google, Meta, and LinkedIn while losing up to 30% of their ad spend to silent bot traffic and click fraud. We built AdShield Intelligence to solve both problems: merging Agentic AI, ML Cyber Defense, and Zero-Knowledge Blockchain Proofs into a single fraud-proof growth platform.

What it does

AdShield Intelligence is an autonomous marketing analytics and fraud prevention engine that:

Aggregates Multi-Channel Data: Consolidates live GA4, Meta, and LinkedIn metrics into a single real-time dashboard.
Auto-Calculates Key Metrics: Tracks performance metrics continuously: $$\text{ROAS} = \frac{\text{Revenue}}{\text{Ad Spend}} \qquad \text{CAC} = \frac{\text{Total Cost}}{\text{Conversions}}$$
Intercepts Ad Fraud (ML): Uses real-time anomaly detection to catch bot attacks using Z-score deviations: $$Z = \frac{x - \mu}{\sigma}$$ If live CPC or traffic ($x$) deviates by $Z > 3.0$ from the moving average ($\mu$), an automated defensive circuit-breaker triggers.
Predicts & Reallocates (AI): Ranks ad copy variations and auto-reallocates budget to high-performing campaigns.
Verifies On-Chain (ZK-Blockchain): Logs cryptographically hashed proofs ($\text{keccak256}$) of spend and audit reports on-chain without exposing raw financial data.

How we built it

Frontend: Next.js 14 (App Router), TypeScript, TailwindCSS, Recharts, and RainbowKit/Wagmi for Web3 wallet connectivity.
AI & Machine Learning: Python FastAPI microservice with Scikit-Learn (Isolation Forest for anomaly detection) and LangChain + OpenAI for creative scoring and agentic recommendations.
Blockchain Layer: Solidity smart contract (AdSpendRegistry.sol) on Polygon Amoy testnet for immutable audit logging.

Challenges we ran into

Schema Normalization: Unifying messy, inconsistent data formats from Meta, Google, and LinkedIn into a single standardized schema.
Privacy-Preserving On-Chain Verification: Structuring Zero-Knowledge proofs so ad spend and ROAS could be verified on a public ledger without exposing customer PII or raw company revenue.

Accomplishments that we're proud of

Successfully intercepting simulated bot traffic in real time with our ML model and preventing wasted ad spend.
Building a working end-to-end Web3 + AI system that seamlessly combines modern marketing UI with on-chain cryptographic proofs.

What we learned

Cybersecurity is a marketing issue: Bot traffic isn't just an IT headache—it directly destroys campaign ROI if left unchecked.
Agentic AI drives real action: Turning raw data charts into autonomous, actionable AI recommendations saves teams hours of manual analysis every week.

What's next for AdShield Intelligence Platform

Live API Integrations: Transition from simulated webhook pipelines to full OAuth production integrations with Meta Ads, Google Ads, and TikTok.
Automated Smart Contract Bounties: Enable automated smart-contract payouts for affiliate marketers based on ZK-verified conversion proofs.n

Built With

docker
fastapi
langchain
polygon-amoy-testnet
rainbowkit
react
recharts
tailwind
typescript
viem
wagmi

Created by

## My Contribution

* Conceived and defined the project idea, objectives, and overall system architecture.
* Conducted requirement analysis and identified the key challenges in digital marketing analytics.
* Designed the complete UI/UX using React, TypeScript, Tailwind CSS, and ShadCN UI.
* Developed responsive frontend pages, dashboards, authentication, and campaign management modules.
* Built RESTful APIs using FastAPI and implemented a scalable backend architecture.
* Designed and managed the PostgreSQL database schema and relationships.
* Implemented secure JWT-based authentication and role-based access control.
* Developed AI/ML models for click fraud detection using Isolation Forest.
* Built predictive models for ROAS and campaign performance using Scikit-learn and XGBoost.
* Integrated AI recommendation features to provide campaign optimization suggestions.
* Created interactive dashboards with real-time charts, KPIs, and campaign analytics.
* Implemented report generation and data export features.
* Added notification and alert systems for fraud detection and campaign performance.
* Containerized the application using Docker and Docker Compose.
* Performed testing, debugging, performance optimization, and security validation.
* Prepared technical documentation, project reports, API documentation, and deployment guides.
* Managed version control using Git and followed clean coding practices throughout the project.
* Coordinated project planning, task scheduling, and milestone tracking from development to deployment.

HARISH D

Updates

HARISH D started this project — Aug 06, 2026 12:05 PM EDT

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.