An investor recently built an entire core banking system interfaceâcomplete with role-based logins, general ledger, and transaction viewsâin a single prompt using an AI coding agent. A year ago, this would have taken weeks. Today, it took an afternoon. Yet despite this stunning speed, the system doesn't actually work like a real bank. This gap between what AI can generate quickly and what financial institutions actually need reveals the most important truth about AI in banking: the interface layer is vulnerable to disruption, but the systems that hold money and manage risk are not. Why the 'SaaS Apocalypse' Got the Story Half Right The panic across software companies is real. When Anthropic launched a legal plugin for Claude in February 2026, it wiped roughly $285 billion off the legal tech sector in a single day. Thomson Reuters dropped 16%, RELX fell 14%, and LegalZoom sank nearly 20%. The message seemed clear: AI will replace enterprise software. SaaS is dead. Seat-based pricing is finished. But this narrative collapses when you look at what actually happened. The AI legal tool disrupted the interface layerâthe dashboards, workflows, and tools that present information to users. It did not disrupt the underlying legal research databases, compliance frameworks, or regulatory systems that law firms depend on. The distinction matters enormously for banking. What AI Can Actually Disrupt in Banking (And What It Cannot) The banking stack has three layers. The interface layer sits on topâdashboards, customer portals, loan officer tools. The controls layer sits in the middleâcompliance, fraud detection, anti-money laundering (AML), and risk management. The core layer sits at the bottomâthe actual ledger, the system of record, the financial truth that regulators audit and customers depend on. AI coding agents can vibe-code the interface layer in an afternoon. They can generate dashboards, workflows, and user experiences that look modern and function smoothly. This is why interface-layer software companies are panicking. But the other two layers tell a different story. - Interface Layer (Vulnerable): Dashboards, customer portals, loan officer tools, and workflow automation. AI agents can generate these in hours or days, making traditional interface-layer SaaS vulnerable to disruption. - Controls Layer (Harder Than It Looks): Compliance, fraud detection, anti-money laundering (AML), and risk management. Building these requires months of domain expertise, regulatory knowledge, and testing. AI can assist, but cannot replace the human judgment and institutional knowledge required. - Core Layer (Not Happening Anytime Soon): The actual ledger, the system of record, transaction reconciliation, and audit trails. Building this requires years of development, regulatory approval, and institutional trust. AI cannot vibe-code financial truth. Why Banks Won't Replace Their Core Systems with AI-Generated Code Consider what a core banking system actually does. It maintains the ledgerâthe authoritative record of who owns what money. It reconciles transactions across thousands of accounts. It survives audits from regulators who demand perfect accuracy. It handles edge cases that occur once every five years but could cost millions if handled wrong. It must be bulletproof, auditable, and compliant with regulations written by people who have never heard of AI. An AI-generated core banking system might look perfect in a demo. But as one investor noted after building 21 AI projects in 18 months: "I really tried building a system of record. The first thing I really wanted to build was an entire operating system for my venture investing activities. It was a complex requirement and I'd constantly face challenges. I deleted everything and started from scratchâthrice." Not one of those 21 projects was a system of record. Not one maintained financial state. Not one had to reconcile a ledger, survive an audit, or report to a regulator. There is a serious code quality problem with AI-generated code that gets ignored. When AI generates enormous amounts of code and nobody is reading it closely, nobody deeply understands the codebase, and there is no systematic review, you get fragility. For a weekend project, fragility is fine. For a system holding customer deposits, fragility is a catastrophe. The Real Transformation Happening in Banking Right Now While the SaaS panic focuses on interface disruption, the actual transformation in banking is happening in the controls and core layersâand it is far more profound. Banks are investing heavily in AI for fraud detection, credit risk modeling, and real-time monitoring. The AI-powered lending market was valued at $109.73 billion in 2024 and is projected to reach $2.01 trillion by 2037, growing at a 25.1% compound annual rate. This is not about replacing core systems. It is about making them smarter. AI-driven credit models analyze up to 10,000 data points per borrower, compared to 50 to 100 in traditional scoring. A UK high-street bank implemented machine learning models that identified 83% of previously unrecognized bad debt without increasing loan rejection rates. Lenders using AI-based scoring have reduced per-loan origination costs by up to 14% and cut defect rates by 40%, with a 5-day shorter loan production cycle. In investment banking, AI agents are reshaping trading and market analysis. AI-driven trading systems now account for more than 65% of all trading volume in major global equities markets. These systems process millions of data points per second, integrating structured and unstructured data to identify profitable trading opportunities in real time. Machine learning algorithms detect price patterns, anomalies, and hidden correlations that traditional quantitative models miss. How Risk Leaders Are Adapting to AI-Accelerated Banking For Chief Risk Officers, the mandate is clear: safeguard portfolio quality while modernizing risk infrastructure. AI investment in commercial banking is accelerating at a notable pace. Industry forecasts indicate that AI spending in the Americas banking sector could exceed $54 billion by 2028ânearly tripling from 2024 levels. This level of capital allocation signals a fundamental shift: AI is no longer viewed as an incremental enhancement. It is considered foundational infrastructure. But this scale of investment demands disciplined oversight. Are AI initiatives aligned with defined risk appetite statements? Is governance keeping pace with deployment velocity? Are internal teams sufficiently trained to interpret AI outputs? Is the institution prepared for heightened regulatory scrutiny around automated decisioning? The strategic sweet spot lies in controlled accelerationâmodernizing the risk stack while reinforcing control frameworks. Steps for Financial Institutions to Navigate AI Responsibly - Separate Interface From Core: Invest in AI for customer-facing tools and workflows, but maintain rigorous human oversight and governance for credit decisions, fraud detection, and systems of record. The interface layer can move fast; the core layer must move carefully. - Build Explainability Into Every Model: AI-driven decisioning must be explainable, auditable, and compliant with regulatory expectations. Governance frameworks must evolve to ensure transparency, fairness, and mitigation of bias. Data lineage and model validation processes must remain rigorous even as deployment speeds increase. - Invest in Domain Expertise, Not Just AI Tools: The controls layer requires months of domain expertise. Hire people who understand compliance, fraud, and risk management. Use AI to amplify their judgment, not replace it. The most dangerous AI systems are those deployed by teams that do not understand the domain deeply enough to catch errors. - Test Ruthlessly Before Production: AI-generated code looks good in demos. Real financial systems must survive edge cases, regulatory audits, and market shocks. Treat AI-generated code as a starting point, not a finished product. Invest in testing, code review, and validation that would make traditional software engineers proud. What This Means for the Future of Banking The institutions best positioned for sustainable growth will not simply extend capital efficiently. They will integrate advanced analytics, strengthen governance, and proactively manage emerging digital risks. The question is not whether AI will continue to transform banking. The question is whether risk organizations will lead the transformationâor react to it. The investor who vibe-coded a banking interface in an afternoon was demonstrating something real: AI coding agents have become extraordinarily powerful at generating user-facing software. But the fact that he deleted his core banking system three times and started from scratch should tell us something equally important: the systems that hold money and manage risk are not going to be disrupted by a clever prompt. They will be transformed by AI, but only when that transformation is guided by people who understand banking deeply enough to know what can go wrong.
