AI-Powered Passenger Experience in Airlines: A KPI-Driven Framework with Mixed-Methods Evidence

Abstract

This article develops and tests a KPI-driven, Balanced-Scorecard–aligned framework that links AI modalities to enterprise outcomes across the passenger journey. The primary objective is to quantify and explain how dynamic offers, recommender systems, conversational AI, biometrics, and predictive IRROPS affect customer experience (NPS, CSAT, CES), monetization (ARPP, conversion), operations (IRROPS time, OTP), and finance/ESG (RASK, CASK, CO₂/ASK). Using a mixed-methods, explanatory-sequential design, we analyze multi-system KPIs and A/B or difference-in-differences rollouts, followed by executive interviews on governance (privacy, fairness, robustness, explainability). Findings show dynamic ancillary pricing delivers +17–58% conversion and +10–43% revenue per offer; conversational AI reduces waiting time up to 80% for routine intents but requires hybrid escalation to sustain NPS; recommenders raise CTR (~+15%) yet need stronger causal links to repeat booking and CLV; revenue-management accuracy improves +14–22%, supporting yield stability. Evidence remains limited on causal bridges to RASK/CASK and CO₂/ASK. Theoretically, we formalize a five-layer KPI architecture and position digital maturity as a measurable moderator. Practically, we recommend a Foundations → Pilot → Scale → Optimize roadmap with per-pax/per-ASK denominators, instrumentation (A/B, DiD, uplift), and responsible-AI gates (consent, bias, robustness, model cards). The framework enables airline leaders to convert AI initiatives into decision-grade, auditable value.