Share with your CIO
McKinsey’s latest report on AI in banking risk functions puts a sharp number on what’s coming: 60-70% of manual operational tasks, including data gathering, compliance checks, and monitoring, could be automated away. Early wholesale lending deployments are already showing 30-50% productivity gains with 95% output quality. Leading banks are now targeting productivity improvements well above 30% across risk functions within two to three years, against the 5-10% annual gains that passed for ambition under traditional approaches. The catch is that capturing those gains requires shared platforms, not fragmented point solutions.
What this means for your business
Banks sitting on a patchwork of AI pilots are the ones most exposed here. The McKinsey argument, shaped by an advisory firm that sells the transformation it’s describing, still holds on the structural logic: productivity gains that reach 30-60% at the workflow level don’t materialize from one-off deployments. They require a common data layer, shared governance, and engineers who understand both risk and AI. The institutions that have been running cautious, siloed experiments are about to discover that caution has a compounding cost.
The workforce math deserves more attention than it usually gets. McKinsey proposes splitting the freed capacity roughly in half: one half toward deeper analytical and strategic risk work, the other toward AI engineering and data governance roles. That’s not a natural org-chart outcome. It requires deliberate reclassification of what a risk function actually is, and most banks haven’t started that conversation at the workforce planning level. The skills gap between today’s risk analyst and the hybrid risk-plus-AI profile McKinsey envisions is wide enough that retraining timelines will stretch past the two-to-three year productivity horizon the report sets.
The real competitive separator won’t be which bank adopts AI in risk first. It will be which bank builds the shared platform infrastructure that makes automation replicable across credit, compliance, and financial crime simultaneously. A bank that automates credit assessment in isolation still has three other siloed functions consuming manual labor. The leading indicator to watch is whether CIOs are funding a unified risk data platform or continuing to approve function-by-function vendor contracts. The former compounds; the latter just shifts the bottleneck.
Concept deep-dive: Risk operating model
A risk operating model is the full architecture of how a bank identifies, measures, and responds to financial and regulatory risk, covering people, processes, data flows, and technology together. Traditional models relied on manual review cycles that were slow and labor-intensive by design. An AI-enabled version replaces those cycles with continuous, automated monitoring, think of it as moving from monthly audits to real-time dashboards that flag problems as they emerge. The business consequence is faster credit decisions and earlier detection of borrower stress.
Based on reporting from AI Transformation: Banks Could Automate 70% of Risk Tasks, Says McKinsey, ETHRWorld, originally published 2026-11-10 02:36:00.

