Agentic AI for Post2Pre Churn Prevention

Business Problem: Postpaid customers downgrading to prepaid plans causes immediate ARPU loss and reduced customer lifetime value. Traditional rule-based retention systems cannot process unstructured data, operate in real-time, or personalize interventions.

Agentic Solution: 7 specialized AI agents autonomously analyze 5 datasets to detect churn risk early, diagnose root causes, design personalized Floor/Middle/Ceiling offers, conduct empathetic conversations, and measure intervention effectiveness. Targets customers with AON â‰Ĩ 90 days.

7
AI Agents
5
Datasets
Agent-Data Links
90+
Days AON Threshold
Real-time
Processing Mode
100%
Automation Rate
Capability Traditional Rules Agentic AI
Churn Detection ✗ Batch processing (daily/weekly) ✓ Real-time risk scoring
Data Analysis ✗ Structured data only (ARPU, tenure) ✓ Multi-modal (profiles + transcripts + sentiment)
Root Cause ✗ Not identified ✓ NLP extracts reasons from conversations
Interventions ✗ Generic discount for all ✓ Personalized Floor/Middle/Ceiling offers
Conversations ✗ Limited by human agent capacity ✓ Scalable LLM-powered dialogue (Arabic + English)
Adaptation ✗ Manual rule updates ✓ Continuous learning from outcomes

Agent-Dataset Architecture