AI Nudges in Bancassurance: a Framework for Personalising Protection at Scale in the Digital and Green Economy

Abstract

Asia–Pacific continues to exhibit a material protection gap despite rapid digitisation in banking. This
paper proposes and expert-validates through consultation with 11 industry professionals a four‑layer
framework for deploying AI‑enabled, behaviourally‑informed nudges in bancassurance—integrating
Data Foundation, Intelligence Engine, Engagement Orchestration, and Governance. A systematic
literature review (2010–2025) and expert interviews (N=11; executives, regulators, academics) inform
design choices. We explicitly incorporate empirical realities—a typical ~21% median nudge effect
alongside high null rates (~38%) (Hummel and Maedche, 2019, p. 53)—by embedding experimentation
(A/B, bandits), suitability/affordability controls, explainability, and market‑specific policy toggles. Two
utility demonstrations (travel insurance self‑serve; cyber protection with advisor copilot) illustrate
operational feasibility and ethics‑by‑design. We further align the artefact to sustainability objectives
(paper‑avoidance, compute‑energy proxies, green‑product uptake) and cross‑jurisdictional obligations
(e.g., FEAT/PDPA, PDPO, CSRD/ESRS). Contributions include an integrated framework, method
transparency (PRISMA + traceability), and a methodologically rigorous blueprint for future empirical
validation in production environments