Will this borrower repay? Every credit decision is a model output.
3 AI translations · Banking & Financial Services
You evaluate consumer loan applications (mortgage, auto, personal, credit card) using credit bureau data (FICO, VantageScore), DTI ratios, employment verification, collateral valuation (for secured lending), and your institution's credit policy overlays. You apply score cutoffs, policy exceptions, and adverse action requirements per Reg B/ECOA. For mortgage, you layer in Fannie/Freddie guidelines, LTV thresholds, QM/ATR requirements, and TRID disclosures. Your underwriters handle the exceptions that fall outside automated decision parameters.
You analyze commercial loan requests by evaluating financial statements (spreading and analyzing balance sheets, income statements, cash flow), industry risk, management quality, collateral (real estate appraisals, equipment valuations, A/R and inventory for ABL), guarantor strength, and debt service coverage ratios. For CRE, you evaluate property-level cash flows (NOI, cap rates, DSCR), market conditions (vacancy rates, absorption, comparable sales), environmental risk (Phase I/II), and construction risk for development loans. You prepare credit memos, present to loan committee, and manage the annual review cycle for the existing portfolio.
You manage fair lending compliance for every credit decision: HMDA data collection and reporting, Reg B adverse action notices, disparate impact testing, redlining analysis, and fair lending examination readiness (OCC, FDIC, CFPB, state regulators). For AI/ML models used in credit decisions, you manage model risk per OCC 2011-12 / SR 11-7: model validation, ongoing monitoring, documentation, and governance. The intersection of AI and fair lending is the single hottest regulatory topic in banking — regulators are scrutinizing whether ML models create or amplify disparate impact.