Strategi Transformasi Digital untuk Efisiensi Biaya Perbankan Indonesia: Integrasi TAM, Keamanan Data, dan Agentic AI
Penelitian ini menganalisis bagaimana penerimaan teknologi, keamanan data, dan prioritas transformasi digital berkaitan dengan efisiensi operasional perbankan digital di Indonesia. Kerangka Technology Acceptance Model (TAM) diperluas dengan Data Security untuk menjelaskan continuance intention dan trust, kemudian dikaitkan dengan Operational Cost Efficiency (OCE) berbasis persepsi pengguna. Data dikumpulkan melalui survei daring terhadap 500 nasabah aktif perbankan digital di Jakarta, Surabaya, Medan, dan Makassar pada akhir 2025 hingga awal 2026. Hubungan antarkonstruk diuji menggunakan Partial Least Squares-Structural Equation Modeling (PLS-SEM), sedangkan Best-Worst Method (BWM) digunakan untuk memprioritaskan faktor strategis berdasarkan penilaian 10 pakar. Naskah melaporkan seluruh hipotesis signifikan pada p < 0,05; jalur Data Security terhadap Trust merupakan efek numerik terkuat yang tersedia (β = 0,565; p < 0,001), dan Data Security dilaporkan memoderasi hubungan perceived ease of use dengan continuance intention. Empat prioritas BWM teratas yang dilaporkan adalah Kepemimpinan Digital (0,2255), Strategi Digital Adaptif (0,1786), Keamanan Sistem Terintegrasi (0,1650), dan Otomatisasi Agentic AI (0,1425). Temuan menunjukkan bahwa efisiensi digital memerlukan keselarasan antara adopsi pengguna, keamanan, kepemimpinan, dan otomatisasi; Agentic AI diposisikan sebagai enabler strategis yang memerlukan tata kelola dan human oversight. This study examines how technology acceptance, data security, and digital-transformation priorities are associated with operational efficiency in Indonesian digital banking. The Technology Acceptance Model (TAM) is extended with Data Security to explain continuance intention and trust, and is subsequently linked to user-perceived Operational Cost Efficiency (OCE). Data were collected through an online survey of 500 active digital-banking customers in Jakarta, Surabaya, Medan, and Makassar from late 2025 to early 2026. Relationships among constructs were tested using Partial Least Squares-Structural Equation Modeling (PLS-SEM), while the Best-Worst Method (BWM) prioritized strategic factors based on judgments from 10 experts. The manuscript reports that all hypotheses were significant at p < 0.05; the strongest numerically reported path was Data Security to Trust (β = 0.565, p < 0.001), and Data Security was reported to moderate the relationship between perceived ease of use and continuance intention. The four highest BWM priorities reported were Digital Leadership (0.2255), Adaptive Digital Strategy (0.1786), Integrated System Security (0.1650), and Agentic AI Automation (0.1425). The findings indicate that digital efficiency depends on alignment among user adoption, security, leadership, and automation. Agentic AI should therefore be treated as a strategic enabler that requires governance and human oversight rather than as an independently validated driver of bank-level accounting efficiency.