Customer Intelligence Dashboard

The Challenge

Businesses typically discover a customer has churned only after they stop paying — by then it's too late to intervene. Standard analytics dashboards show you what happened last quarter; they offer no foresight about which high-value customers are quietly drifting toward a competitor right now. Without a way to score churn risk in real time and route that signal to the right team member, retention becomes guesswork rather than strategy.

The Solution

A live analytics dashboard that pairs two trained ML models — a CLV predictor and a churn classifier — with AI agent skills that translate raw scores into plain-English action plans. When a customer crosses a risk threshold, the system doesn't just flag them; it drafts a sales call brief, a marketing outreach template, or a data-analysis memo so the right team member can act within minutes, not days.

Key Capabilities

• 🔮 CLV Prediction GradientBoosting model (R² = 0.957) predicts each customer's lifetime value from purchase history, frequency, and recency — segmenting the portfolio into tiers automatically.
• ⚠️ Churn Classification RandomForest classifier (91% accuracy) scores every active customer's 30-day churn probability so teams can prioritize outreach to the accounts most likely to leave.
• 📞 Sales Agent Skill For at-risk high-CLV customers, the Sales Agent writes a personalized call-prep brief — account summary, risk factors, and suggested talking points — ready to forward to a rep.
• 📣 Marketing Agent Skill Generates targeted campaign briefs for churn-risk segments — subject lines, offer rationale, and send-time recommendations based on the customer's engagement pattern.
• 📊 Real-Time Dashboard Interactive Plotly/Dash visualizations update as new customer data flows in — churn score distributions, CLV heatmaps, and segment-level trend lines all in one view.