APPLIED AI
Your High-Risk Deadline Moved to December 2027. Your Chatbot's Did Not.
Relief on the high-risk file arrived nine days ago. The risk now is that banks read one deferral and stand down a programme that still has an obligation biting today.
CATEGORY
Generative AI, agentic systems, governance, and the LLM tooling that lands in banking workflows.
APPLIED AI
Relief on the high-risk file arrived nine days ago. The risk now is that banks read one deferral and stand down a programme that still has an obligation biting today.
APPLIED AI
Fifteen years of model risk doctrine was rewritten in April, and the one paragraph that will matter most to a 2026 risk committee is the one that says which models the guidance does not cover.
APPLIED AI
Immutable is not the same as trusted. As instant settlement and probabilistic AI reshape banking, the uncertified ledger becomes the weak link, and a certifiable blockchain becomes the audit spine.
APPLIED AI
Payments are now the core of bank technology strategy. This is the operating model, risk frontier and revenue playbook G-SIBs need to lead the agentic, real-time cycle.
APPLIED AI
Retail banking got its API standard a decade ago. Corporate banking never did — and now that AI agents and the Model Context Protocol have turned integration into a runtime decision, the gap has stopped being a nuisance and become a fault line. A concrete failure mode, the strategic choice it forces, and what a corporate banking API standard has to mandate.
APPLIED AI
Regulators have started scanning the horizon out loud. Reading the 2026 emerging-technology landscape for banks across three converging vectors — personalised AI intelligence, synthetic financial crime, and programmable finance — and a framework for turning weak signals into supervisory action before harm scales.
APPLIED AI
Agentic AI is operational infrastructure in 2026. This index measures it the way banks measure capital and credit: a six-dimension readiness score across autonomy tiers, the control plane, regulatory evidence, unit economics, organisational readiness, and global regulatory alignment.
APPLIED AI
Agentic treasury co-pilots are moving from controlled pilots into production in 2026. The pattern is consistent across CIB treasuries: agents read ISO 20022 cash data, call bounded tools, and rebalance liquidity inside policy bands — with SR 11-7, DORA, and EU AI Act controls wrapped around them.
APPLIED AI
ISO 20022 in 2026 is the autonomic nervous system of treasury. pain.001 and pacs.008 carry richer data than any MT message ever did. Nearly half of banks remain off-track for the November 2026 SWIFT MT/MX cut-over, structured addresses become mandatory, and agentic treasury cannot exist without MX-native APIs. This article maps the engineering reality, from CBPR+ validation rules to a real pain.001 fragment with PstlAdr fields.
APPLIED AI
BankStatementParser is an open-source Python toolkit that turns CAMT, PAIN.001, MT940, OFX/QFX, CSV, and scanned PDFs into a unified transaction model treasury can audit — deterministic parsers, LLM fallback, OCR, balance verification, and review.
APPLIED AI
CloudCDN is an open-source blueprint for the AI-native edge — a zero-trust MCP gateway with 42 tools, atomic Durable Objects rate limiting, WebAuthn passkeys, signed URLs, SLSA Level 3 provenance, and 3,185 tests at 100% coverage, mapped to DORA, BCBS 239, and Basel III.
APPLIED AI
A 2026 autonomous treasury readiness index — measuring agentic treasury workflows, programmable liquidity coverage, tokenised deposit integration, real-time payment orchestration, and automated cash control as one operating-model fabric.