COOKED
The core loop — pulling charges from a system, generating invoices and statements, posting payments, reconciling remittance advice, and flagging discrepancies — is structured screen work that RPA and now LLM-driven document extraction handle end to end. Medical billing variants add payer-rule complexity, but coding edits, claim scrubbing, and denial routing are exactly what clearinghouse automation targets. What lingers is the messy exception queue: a stubborn denial, a patient on the phone confused about a balance, a contract term the system misread.
Part 2020 shock, part continued decline in the years since.
Median pay $38,740 → $48,500 +0.2% 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
-0.4% 429,800 → 427,900 on the projections basis
Exposed, and shrinking
Both signals point the same way: the tasks are largely automatable and the BLS projects -0.4% 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.
~42,200 openings a year on average, including replacing people who leave.
RaterBillerPricerFee ClerkTax ClerkRate ClerkReconcilerRate MarkerRate SetterBill CheckerCheck WriterPrice ListerRating ClerkReport ClerkAccount ClerkBilling ClerkCheck TotalerComptometristCost RecorderDeposit ClerkForeign ClerkInvoice ClerkPosting ClerkPricing Clerk
Holding it up: trust premium . Weakest point: liability shield .
Core tasks are already automatable Compiling charge data, keying invoices, posting cash receipts against open A/R, and matching 835 remittance files to claims are all deterministic lookups against structured fields — the only genuinely non-automatable slice is chasing a payer by phone for a denial the clearinghouse rule set never anticipated, which is why this sits at 4 rather than in the 7-13 band.
Fully desk- and screen-based Beyond stuffing statement envelopes, running the printer, and pulling a paper superbill or packing slip off someone's desk, the entire job happens in a billing module and a spreadsheet — the physical residue is real but incidental, hence 2 rather than 0.
No licence, no signature requirement No state licence gates this work; CPC or CPB certification helps you get hired in a physician practice and nothing more, and when a claim goes out wrong it is the provider's NPI and the practice that answer to the payer or to False Claims Act exposure, not the clerk who keyed it.
Executes defined procedures on defined inputs Deciding whether a $40 variance is a short-pay to appeal or a write-off, or which of three CPT modifiers the encounter note supports, is real discretion bounded hard by the fee schedule, the payer contract, and a supervisor's approval threshold — recurring judgment inside a rulebook, which is a 5 rather than a 10.
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 (4/20 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 (1/20) is whether the law requires a licensed human to sign. Trust premium (5/20) is whether buyers specifically pay for a person. Judgment and accountability (5/20) 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 11 of this occupation's 17 points (65%).
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
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 34/100 — EXPOSED.
Genuine two-tier structure: once clearinghouse auto-adjudication clears clean claims, the residual role is denial appeals, payer contract variance analysis, and underpayment recovery — work requiring reading payer policy bulletins, assembling medical-necessity documentation, and negotiating with a human rep. If the job title survives the shrink and reconstitutes as 'denial management / AR recovery specialist', the surviving version scores materially higher on task resistance than the posting clerk it replaced.
A CMS or OIG requirement that claims generated or edited by AI carry a named human attestation — an extension of the existing provider certification on the CMS-1500/UB-04 that submission is accurate. False Claims Act exposure for AI-generated upcoding is already being discussed by DOJ health care fraud units and payer compliance counsel; a rule naming an accountable human reviewer for algorithmically produced claim lines would create a signature that a clerk-level role could hold.
No Surprises Act independent dispute resolution and state balance-billing statutes push consequential, ambiguous calls — is this balance patient responsibility, does this qualify for IDR, should this go to collections — onto a named billing staffer. If health systems formalize a 'patient financial advocate' function that owns write-off and charity-care determinations under 501(r), the role owns real decisions with audit consequences.
State insurance department rules mirroring California SB 1120 (2024), which bars AI alone from making medical necessity determinations — if the mirror-image obligation lands on the provider side, requiring a human to review AI-flagged claim denials before writing off or rebilling a patient balance, that inserts a required human step in the billing loop.
Weak route only: hospital price-transparency enforcement and patient-experience scoring create some demand for a reachable human who explains a bill. This is a call-center trust premium, not a professional one, and buyers are hospitals managing complaint volume rather than patients paying extra.
The limit. Even with every lever, this tops out in the 30s. The levers protect a much smaller headcount doing denial and dispute work; they do nothing for charge entry, statement generation, and payment posting, which is where most of the 404,000 sit. Volume decline is the dominant risk, not per-worker task resistance.
| New York-Newark-Jersey City, NY-NJ | 32,930 | $56,160 +16% |
| Los Angeles-Long Beach-Anaheim, CA | 16,670 | $53,070 +9% |
| Chicago-Naperville-Elgin, IL-IN | 13,610 | $49,960 +3% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 8,680 | $46,610 -4% |
| Dallas-Fort Worth-Arlington, TX | 7,960 | $48,290 +0% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 7,650 | $50,190 +3% |
| Houston-Pasadena-The Woodlands, TX | 6,820 | $47,360 -2% |
| Atlanta-Sandy Springs-Roswell, GA | 6,500 | $49,320 +2% |
| San Jose-Sunnyvale-Santa Clara, CA | 2,350 | $70,280 +45% |
| Rochester, MN | 500 | $67,390 +39% |
| San Francisco-Oakland-Fremont, CA | 4,690 | $66,390 +37% |
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 17. 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.