Revenue Intel

Overview

AI-powered revenue intelligence platform for SaaS businesses with multi-model predictive analytics, real-time Claude AI insights across every dashboard, Monte Carlo scenario simulation, and customer-level natural language Q&A.

Key Features

• Real-time Claude Sonnet insights on 5 dashboards — executive briefings, risk analysis, funnel coaching, scenario guidance, and revenue narratives with natural language customer Q&A
• Multi-model predictive stack: XGBoost churn prediction with SHAP explainability, Logistic Regression for expansion propensity, and Random Forest deal scoring
• Monte Carlo simulation engine with triangular distributions and confidence intervals for what-if scenario planning across churn, conversion, expansion, and pricing levers
• Weighted multi-factor health scoring (Usage 35%, Engagement 25%, Sentiment 20%, Financial 20%) with component-level drill-down and intervention triggers
• Full-stack architecture with FastAPI serving ML models, DuckDB for high-performance OLAP analytics, and SaaS industry benchmarking against median NRR, LTV:CAC, and CAC payback

Challenges & Solutions

The project addressed complex challenges including orchestrating real-time Claude AI insights across five analytical dashboards with context-aware prompts tailored to each page's data, building a multi-model ML pipeline spanning XGBoost churn prediction, Logistic Regression expansion scoring, and Random Forest deal probability, implementing a Monte Carlo simulation engine with 1000+ iterations and triangular distributions for robust revenue forecasting with confidence intervals, designing a weighted health scoring system that synthesizes usage telemetry, engagement signals, NPS sentiment, and financial indicators into actionable risk tiers, and building a performant OLAP analytics pipeline with DuckDB for real-time aggregation across large SaaS datasets.