This paper examines the role of agentic artificial intelligence across three strategic domains of banking: the enhancement of customer experience, the mobilization of low-cost deposits, and the management of balance-sheet mismatches. The study is a narrative conceptual review. Sixteen sources published between January 2024 and March 2026 were retained from searches of Google Scholar screened against stated inclusion criteria; the selection is purposive rather than exhaustive and is reported as such. On this base the paper examines the capabilities that distinguish agentic systems, environmental perception, reasoning, multi-step planning, memory, adaptive learning, and multi-agent coordination, and traces their implications in each domain. The analysis suggests that intelligent agents, through continuous personalization, retention of interaction context, and orchestration of multi-step processes, turn customer experience from a reactive exchange into a sustained, goal-directed relationship. In deposit mobilization, the paper specifies a four-link chain running from agentic capability through reduced friction, trust, and retention to low-cost balances, argues the theoretical warrant for its final link through account primacy, and reports that this link remains untested; a countervailing mechanism, in which lower friction also lowers switching costs and raises deposit mobility, is set out alongside it. For mismatch management, the integration of transactional, behavioral, and market data supports cash-flow forecasting, liquidity monitoring, stress testing, and treasury decisions, although no deployment in a named institution has yet been documented and the domain has not moved past architectural proposal. The paper’s conceptual contribution is a learning loop connecting customer interaction, financial behavior, deposit mobilization, and treasury decision-making. Sustained value creation, the paper concludes, depends on phased deployment, data quality, explainability, bounded authority, human oversight, and fiduciary alignment of agent incentives with customer interests.