Banking

US commercial banking generates over $700 billion in annual revenue across roughly 4,500 commercial banks and 4,800 credit unions. JPMorgan Chase, Bank of America, Wells Fargo, and Citigroup hold over $10 trillion in combined assets. Core activities include deposit-taking, commercial and consumer lending, treasury services, and increasingly fee-based services. Banks have been heavy AI investors for over a decade in fraud detection and risk modeling, but generative AI is now reshaping customer-facing interactions, knowledge work, and software engineering inside the bank.

JPMorgan's LLM Suite (deployed firmwide to 200K+ employees) and IndexGPT for research lead among in-house deployments. Goldman Sachs and Morgan Stanley have parallel firmwide AI platforms. Customer-service AI (Bank of America's Erica, Capital One's Eno) handles billions of routine interactions annually. Fraud detection (FICO Falcon, Featurespace, Bridgei2i) processes every transaction in real time. Credit underwriting AI (Zest AI, Upstart) extends credit to thin-file borrowers. Software-engineering AI (GitHub Copilot, Cursor) is widely deployed in bank engineering teams.

Customer-service representative roles face the steepest displacement; banks have absorbed thousands of role eliminations through attrition. Fraud-analyst work shifts from manual review to AI-output investigation. Credit underwriters work from AI-generated scorecards rather than building models from scratch. Investment research analysts use AI for first-draft synthesis. Software engineering productivity has improved meaningfully — JPMorgan reports 10-20% engineering time savings. New roles: AI risk officer, model governance specialist, conversational-AI product manager.