Remix

Make your own version of SalesGuard AI: South India Retail

Edit the prompt below and generate a customized version in seconds.

Original by Praveen Thala

Describe your app

🧠 MASTER PROMPT: SalesGuard AI (South India Edition)
🔷 ROLE DEFINITION

Act as:

Senior AI Architect
Machine Learning Engineer
Data Scientist
MLOps Engineer
Full Stack Developer (React + FastAPI)
Product Designer

You are responsible for designing and implementing a production-ready AI SaaS platform.

🔷 SYSTEM CONTEXT

You are building:

SalesGuard AI — an intelligent retail decision-support system tailored for small and medium retail businesses in South India, with multilingual support (Tamil, Telugu, Kannada, English).

The system must:

Forecast sales accurately
Predict risks (loss/spike/drop)
Provide actionable insights
Offer interactive simulation tools
Deliver explainable AI outputs
Support real-time data streaming
Be deployable as a SaaS product
🔷 EXECUTION STRATEGY

Follow:

Step-by-step execution
Modular architecture
Clean, production-level code
Explain before implementing
Generate diagrams where needed
Optimize for scalability and real-world usage
🚀 PHASE-WISE EXECUTION INSTRUCTIONS
🔷 PHASE 1: REQUIREMENT ANALYSIS
Tasks:
Clearly define:
Problem statement
Objectives
Target users (Retail Owner, Analyst, Admin)
Add South India retail context:
Seasonal festivals (Pongal, Diwali, Ugadi, Onam)
Small shop constraints
Local language usage
Output:
Structured documentation
🔷 PHASE 2: DATA PIPELINE DESIGN
Tasks:
Build scalable pipeline:
Data ingestion (CSV, API, streaming)
Schema validation
Missing value handling
Outlier detection
Data transformation
Include:
Support for:
Walmart
Rossmann
Online Retail
M5 dataset
Output:
Python pipeline code
Data flow diagram
🔷 PHASE 3: FEATURE ENGINEERING
Tasks:

Create:

Time-based features
Lag features
Rolling statistics
Seasonal indicators
Promotion & holiday flags
Add:
Local festival features (South India specific)
Output:
Feature engineering module
🔷 PHASE 4: MODEL TRAINING
Models:
ARIMA
Prophet
XGBoost
LightGBM
Tasks:
Train models
Hyperparameter tuning
Evaluation (RMSE, MAE)
Output:
Training scripts
Model comparison table
🔷 PHASE 5: ENSEMBLE ENGINE
Build:
Weighted averaging
Stacking model
Auto-selection logic
Output:
Ensemble code
Performance improvement metrics
🔷 PHASE 6: DECISION INTELLIGENCE
Generate:
Sales loss alerts
Demand spike/drop alerts
Inventory optimization
Pricing recommendations
Promotion strategies
Output:
Human-readable insights engine
🔷 PHASE 7: EXPLAINABLE AI
Implement:
Feature importance
Custom SHAP (without library)
Waterfall visualization logic
Output:
Explainability module
🔷 PHASE 8: ADVANCED FEATURES (MANDATORY)
Implement:
1. Real-Time Streaming
WebSocket-based live sales ticker (update every 2s)
2. What-If Simulator
Interactive sliders (price, promo)
Instant forecast recalculation
3. Anomaly Detection Timeline
Z-score based detection
Visualization with bands
4. Model Monitoring
Drift detection
Auto-retraining triggers
5. Competitive Benchmarking
Radar chart (6 metrics comparison)
6. Smart Receipt Scanner (Mock UI)
OCR-based input flow (UI only)
7. PDF Report Generator
One-click export (board-level insights)
🔷 PHASE 9: BACKEND DESIGN
Use:
FastAPI
APIs:
/upload-data
/train-model
/predict
/get-insights
/simulate
/stream
Output:
Clean modular backend code
🔷 PHASE 10: FRONTEND DESIGN
Use:
React + Tailwind
Chart.js / Recharts
Build:
Dashboard
Forecast graphs
Alerts panel
Simulation UI
Explainability panel
Add:
Multilingual toggle (EN/Tamil/Telugu/Kannada)
🔷 PHASE 11: SYSTEM ARCHITECTURE
Design layers:
Data Layer
Processing Layer
Model Layer
Ensemble Layer
Decision Layer
API Layer
Presentation Layer
Output:
Architecture diagram (text + explanation)
🔷 PHASE 12: CODE GENERATION RULES

All code must be:

Modular
Scalable
Production-ready
Well-commented

Generate:

Python scripts
API backend
React components
🔷 PHASE 13: DEPLOYMENT GUIDE
Include:
Backend deployment (Render / AWS / Docker)
Frontend deployment (Vercel / Netlify)
CI/CD pipeline
Model monitoring setup
🔷 PHASE 14: OUTPUT & EVALUATION

Deliver:

Forecast results
Accuracy metrics
Insights
UI dashboard
🔷 PHASE 15: ITERATIVE IMPROVEMENT

Continuously:

Improve features
Tune ensemble
Optimize UI/UX
Enhance explainability
🔷 CONSTRAINTS
Must work for small retail shops
Must support low-resource environments
Must be multilingual
Must be real-world deployable
Avoid unnecessary complexity
🔷 FINAL OUTPUT FORMAT

For each phase, generate:

Explanation
Code
Diagram (if applicable)
Output preview

Free to use. No credit card required.