A personal and technical reflection on my first three months at the Government Savings Bank (GSB): why I came, the birth of GSBAIL (Government Savings Bank AI Laboratory), and the talk ‘GSB AI: Beyond the Model’ on 19 November 2026. Covers reinforcement learning, RLHF, RLVR/GRPO, the Kao model family (KaoOCR, KaoLEAF, KaoChat, KaoASR, KaoNPL), the QR Maha Heng × Nokkrasip merchant-loan pilot, and the case for sovereign, local, and ethical AI in a social bank.