AI underwriting shows dramatic stats
Recent social analysis claims AI‑driven underwriting is producing 78% credit approval rates and just 2% delinquency by focusing on cash‑flow patterns over legacy scores — a stat that underlines why lenders are piloting AI to speed decisions and tighten loss control. If accurate, that level of throughput supports aggressive origination without proportional headcount increases. (x.com) (x.com)
Lending pilots are already turning into headline metrics: online lender LendingClub reported a 40% reduction in delinquencies after deploying AI models across underwriting and operations. (finainews.com) Vendors selling AI decision engines advertise materially higher auto-decision rates and approval lifts — Zest AI’s product page states up to 80% auto-decisioning and average approval lifts of ~30% from its models. (zest.ai) The shift to cash‑flow‑first underwriting is increasingly mainstream: Ocrolus and Plaid publish playbooks showing banks and fintechs replacing static bureau-only screens with verified bank-transaction cash‑flow analytics to expand credit access and speed decisions. (ocrolus.com) Operational throughput gains are being practicalized by platform vendors — Solifi this month launched Document Intelligence, a capability the company and trade publications say can cut document‑verification time by up to 70%, reducing a key bottleneck in originations. (solifi.com) Equipment‑finance teams face idiosyncratic pressure from depreciation and tax shifts: recent industry reporting highlights permanent changes to bonus‑depreciation rules and recommends lenders move from static schedules to market‑based valuations while pairing that with cash‑flow underwriting. (bassets.net) Dealership and floorplan lenders are tightening around inventory and liquidity controls as new‑vehicle and used‑vehicle supply metrics rebalance; Cox Automotive and S&P report tighter inventory and normalized days‑to‑turn in early 2026, pushing captives and floorplan lenders to adopt real‑time verification and automated underwriting. (coxautoinc.com) Regulators and consulting firms warn governance gaps: NAIC surveys and Accenture research show industry concern about model explainability and that a minority of AI use cases move from development to production, underscoring the compliance and audit‑trail work lenders must complete before scaling origination throughput. (content.naic.org)
