COOKED
The core of this job — spreading financial statements, computing ratio and covenant metrics, pulling bureau and industry data, and drafting credit memos with a recommended risk rating — is structured text-and-numbers work that models and existing credit-scoring engines already handle at usable quality. What persists is the ambiguous middle-market file: judging management quality, unpicking related-party transactions or aggressive revenue recognition, and structuring covenants that a committee will actually approve. No license protects the role, and the borrower relationship is usually owned by the relationship manager, not the analyst.
Part 2020 shock, part continued decline in the years since.
Median pay $73,650 → $83,510 -9.3% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
-4.4% 67,800 → 64,800 on the projections basis
Exposed, and shrinking
Both signals point the same way: the tasks are largely automatable and the BLS projects -4.4% by 2034. This is the case where the score and the forecast agree, and it is the one worth taking seriously.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~3,700 openings a year on average, including replacing people who leave.
FactorerCredit AnalystCredit ManagerCredit OfficerCredit NegotiatorCredit SpecialistCredit CoordinatorCredit Risk AnalystCredit Risk ManagerLoan Review AnalystCredit AdministratorCredit RepresentativeEscrow RepresentativeCredit Risk SpecialistCredit Assistant ManagerCredit Portfolio ManagerCommercial Credit AnalystCommercial Credit ManagerCredit Assessment AnalystCredit and Collections AnalystMunicipal Fixed Income Analyst
Holding it up: judgment & accountability . Weakest point: liability shield .
Core tasks are already automatable Spreading a borrower's tax returns and audited statements into a standard template, recalculating DSCR, leverage and fixed-charge coverage, running the bureau pull and populating a memo template already happens inside Moody's/nCino/OCR pipelines at production banks — the 5 rather than a 2 reflects the site visits and management interviews on middle-market files that still need a person to sit in the room.
Fully desk- and screen-based Everything you touch is a core banking system, a spreadsheet and a PDF of a K-1; the occasional plant tour or borrower visit before an annual review is the only reason this isn't a flat 0.
No licence, no signature requirement There is no license to lose — CFA or CRC is a resume item, not a regulatory gate — and when a credit sours the exam finding lands on the approving officer's signature and the committee minutes, not the analyst who spread the numbers.
Meaningful discretion You choose which addbacks to allow, whether a related-party rent is arm's-length, and what covenant package to propose, but the risk rating you recommend is scrutinized against the bank's rating grid and overridden at committee, so the 9 marks genuine discretion inside a policy that names the limits.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (5/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (2/20) is whether the law requires a licensed human to sign. Trust premium (6/20) is whether buyers specifically pay for a person. Judgment and accountability (9/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 17 of this occupation's 24 points (71%).
Embodiment (2/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
Accountants and Auditors EXPOSED
Loan Officers EXPOSED
Economists EXPOSED
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 39/100 — EXPOSED.
Formal delegation of lending authority downward: if banks respond to automated spreading by making analysts voting members of credit committee or holders of individual approval limits (a documented sign-off on structure and covenant package, with that name tied to the file in post-mortem and regulatory loan review), the role owns the consequential call instead of feeding someone else's. Watch for job titles shifting from 'credit analyst' to 'credit officer' with stated hold limits.
Pure task-mix shift, no law needed: this occupation has a genuine two-tier structure — spreading/ratio/bureau-pull tier and the ambiguous-file tier (management quality, related-party transactions, revenue-recognition aggressiveness, covenant structuring a committee will accept). If headcount falls and the surviving roles are staffed only on non-conforming middle-market, asset-based, and workout files, the residual day is mostly the tier models handle badly. Recognisable by shrinking analyst-to-loan ratios alongside rising average deal complexity.
Bank supervisors extending model-risk rules to generative credit tools: if OCC/Fed SR 11-7 guidance (or a successor exam manual update) is read to require a named, qualified human credit officer to independently validate and attest to each AI-generated risk rating before it enters the loan file — and examiners cite institutions where memos are machine-drafted without an attributed reviewer — the analyst becomes the required signature rather than an optional one. CFPB's 2023 circular on adverse-action notices under Reg B (reasons must be specific and accurate, no 'black box' checklist) is the same mechanism operating on the consumer side.
Narrow route only, in private credit and non-bank direct lending: if LPs' side letters or fund diligence questionnaires begin requiring disclosure of whether underwriting memos were human-authored, and a named human underwriter of record is treated as a diligence checkbox, a thin premium attaches. This is speculative and would not extend to bank commercial lending, where the borrower never meets the analyst.
The limit. No licensure exists or is being proposed for credit analysis, so liability_shield cannot reach the levels of appraisal, audit, or actuarial work — the ceiling is exam-driven attestation duty, which attaches to the institution first and the individual only derivatively. Trust premium is structurally capped because the buyer-facing relationship belongs to the relationship manager. Realistic combined ceiling is roughly the low-to-mid 40s, and only if the judgment-tier files stay outside model competence.
| New York-Newark-Jersey City, NY-NJ | 7,530 | $136,500 +63% |
| Dallas-Fort Worth-Arlington, TX | 3,210 | $80,530 -4% |
| Chicago-Naperville-Elgin, IL-IN | 2,560 | $92,900 +11% |
| Charlotte-Concord-Gastonia, NC-SC | 2,190 | $100,640 +21% |
| Los Angeles-Long Beach-Anaheim, CA | 2,100 | $95,640 +15% |
| Phoenix-Mesa-Chandler, AZ | 1,990 | $77,020 -8% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 1,430 | $92,460 +11% |
| Atlanta-Sandy Springs-Roswell, GA | 1,300 | $80,750 -3% |
| New York-Newark-Jersey City, NY-NJ | 7,530 | $136,500 +63% |
| San Jose-Sunnyvale-Santa Clara, CA | 330 | $129,240 +55% |
| San Francisco-Oakland-Fremont, CA | 850 | $122,590 +47% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 24. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.