Commercial Banking

Commercial banking covers the largest segment of US financial services — depository institutions that take deposits, make loans, and provide payment services to consumers and businesses. The big four (JPMorgan Chase, Bank of America, Wells Fargo, Citi) hold roughly $10 trillion in combined assets and dominate the high-margin segments (commercial lending, treasury services, wealth management). Mid-tier regionals (US Bank, Truist, PNC, Capital One) serve more localized markets, and roughly 4,000 community banks fill out the long tail. Revenue comes from net interest margin (the spread between loan rates and deposit costs), service fees, card-interchange revenue, and wealth-management fees. The industry has been heavily regulated since the 2008 financial crisis (Dodd-Frank, stress testing, capital requirements) — making AI deployment slower and more cautious than in less-regulated sectors.

JPMorgan's LLM Suite is the most ambitious enterprise-AI deployment in financial services to date — 200,000+ employees have access, with use cases spanning research, code generation, customer-service drafting, and analyst-grade financial analysis. Bank of America's Erica handles roughly 2 billion customer interactions annually as the bank's AI virtual assistant. Wells Fargo's Fargo handles similar volume. NICE Actimize is the dominant AML and fraud-detection platform across mid-tier and regional banks — the workhorse most banks rely on for transaction monitoring and suspicious-activity reporting. Bloomberg AI and AlphaSense get used heavily for market research and earnings analysis at the wealth-management and investment-banking ends. Hebbia Matrix is gaining traction with research teams that need answers across thousands of internal documents. The horizontal models (ChatGPT, Claude, Microsoft Copilot) handle general-purpose drafting, code, and research.

The economics are bifurcated. Routine commercial-banking work — branch teller transactions, loan processing, account-opening, basic customer service — is being automated aggressively, with branch-employee headcount falling at most major banks every quarter. Meanwhile, mid-tier analyst roles (credit analyst, risk analyst, compliance analyst) are being augmented rather than replaced — AI handles the document-review heavy lifting, but final decisions still require licensed bank staff under regulatory frameworks. The regulatory layer is the real moat: even with strong AI, banks need humans of record on most credit decisions, AML reporting, and customer disputes. New roles emerging: AI compliance officer (validates AI decisions against fair-lending and disparate-impact rules), AI prompt-engineering specialist (builds workflows on JPM LLM Suite or equivalent), AI risk-and-control validator. The community-bank tier faces the steepest challenge — they lack scale to build proprietary AI but face competitive pressure from AI-augmented megabanks.