■ CRITICAL RISK ■ Finance
For the underwriting part, AI already has the job — algorithms make most consumer credit decisions today. What's really being replaced is the loan officer as salesperson and hand-holder, and that's happening now as digital lenders prove borrowers will take a mortgage from an app.
“An algorithm already decides. You just deliver the bad news.”
Our AI replacement risk score — how we score jobs
Strip the title down and a loan officer does three things: finds borrowers (sales and referral relationships with realtors, dealers, and past clients), packages them (collecting documents, structuring the loan, shepherding the file to underwriting), and advises them (explaining products, rates, and what that debt-to-income ratio means for their dreams). In consumer lending the actual credit decision left human hands years ago — automated underwriting systems like the ones behind conforming mortgages decide, and the officer works around the verdict.
Digital-first lenders attacked the packaging and advising layers. Online mortgage platforms pull income, asset, and employment data directly from source systems, so 'collect the documents' becomes 'grant permissions.' Chat and voice AI answer product questions at any hour without commission pressure. Personal loans and auto financing went nearly end-to-end digital first; mortgages, the most paper-encrusted product, are following. Meanwhile AI tools draft the officer's follow-up emails, pre-fill applications, and score leads — which sounds like help, until you notice each officer now handles triple the pipeline and the branch needs a third of the officers. That's what a 96 looks like in practice: not a firing, a slow shrink through productivity.
The defensible ground is complexity and trust. First-time homebuyers who are scared, self-employed borrowers whose finances defy standard forms, commercial lending where deals are genuinely negotiated, private banking where relationships move money — these keep humans because the borrower wants an accountable person, not because software can't compute the numbers. The commission structure is the real question: as easy loans go self-serve, the surviving officers keep only the hard, lower-volume ones, and the economics of the job change even where the job survives.
Automatability: our editorial assessment of current and near-term AI capability
Automated underwriting took the decision years ago; this decade takes the rest of the workflow. Digital lenders already close personal and auto loans with no officer involved, and mortgage origination is steadily going self-serve for straightforward borrowers. Expect officer headcount to fall through the late 2020s as each survivor handles more volume with AI tooling, with relationship-heavy commercial and complex-borrower niches holding out longest.
Mostly no, in consumer lending. Automated underwriting systems make or heavily constrain the decision for standard mortgages, auto loans, and personal loans; the officer packages the file and communicates the outcome. Human judgment still genuinely decides in commercial lending, private banking, and exception cases — which is exactly where the durable careers are.
As a long-term bet on the traditional retail role, yes — our risk score of 96 reflects a job whose decision-making is automated and whose sales-and-paperwork layer is going self-serve. If you enter anyway, aim immediately for commercial lending, complex-borrower niches, or relationship banking, and treat standard consumer origination as a training ground, not a destination.
Because a mortgage is the largest, scariest transaction most people ever make, and many borrowers — especially first-timers and the self-employed — want an accountable human guide. Regulation and paperwork complexity slow full automation too. But watch the trend: each cohort of borrowers is more comfortable closing digitally, so this protection erodes generationally rather than holding.
Pick a moat: complex-income and non-QM lending, commercial deals, or a genuinely owned referral network. Simultaneously master the digital origination tools, because the surviving officers will be the high-volume, tech-leveraged ones. If you enjoy the advisory part most, financial advising or private banking uses the same trust-building skills with better automation resistance.