COOKED
The core of this job — coding transactions to the general ledger, reconciling bank and credit card accounts, matching invoices to purchase orders, chasing AR, running payroll entries, and producing monthly trial balances — is exactly what bank-feed rules, OCR receipt capture, and now LLM-driven categorization in QuickBooks/Xero/Bill.com already do at usable quality. There is no license requirement, no signature liability (the CPA signs, not the clerk), and clients rarely pay for the relationship with the bookkeeper specifically. What persists is the messy edge: small-business owners with chaotic records, exception investigation, sales-tax and multi-state quirks, and being the person who notices the numbers are wrong before the accountant does.
Part 2020 shock, part continued decline in the years since.
Median pay $41,230 → $50,670 -1.7% 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
-5.8%
Percentage only. The projection counts a different population from the 1,373,680 above — it includes self-employed workers, which for this occupation is most of them, so the two headcounts are not comparable.
Exposed, and shrinking
Both signals point the same way: the tasks are largely automatable and the BLS projects -5.8% 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.
~170,000 openings a year on average, including replacing people who leave.
ClipmanBond ClerkBookkeeperClassifierAudit ClerkChart ClerkCheck ClerkField ClerkPrice ClerkAdvice ClerkBudget ClerkChart PickerChart ReaderEscrow AgentExport ClerkFigure ClerkLedger ClerkMargin ClerkAccount ClerkBalance ClerkBooking ClerkChart ChangerClerk AnalystEscrow Closer
Holding it up: trust premium . Weakest point: liability shield .
Core tasks are already automatable A 5 reflects that bank-feed rules plus receipt OCR already handle the transaction coding, three-way match, and reconciliation clearing that fill most of a clerk's week; it isn't a 0 because chasing a client for the missing vendor invoice, untangling a prior year's misposted retainer, and fixing a payroll journal that broke when someone changed a pay period still require a human to go find facts that exist nowhere in the system.
Fully desk- and screen-based A 3 covers the residual paper — opening the mail, scanning the shoebox of receipts, walking a deposit to the bank, pulling a physical file from the cabinet — none of which requires you to leave the office or handle anything more demanding than a stapler, which is why it sits above pure-screen work but nowhere near a 5.
No licence, no signature requirement At 2 there is no credential gate at all: you can be hired with no certificate, the CPA or EA signs the return and the audit opinion, the owner signs the payroll tax filings, and even a voluntary QuickBooks ProAdvisor or AIPB certification is a hiring signal rather than a legal barrier anyone can be sued through.
Executes defined procedures on defined inputs A 5 fits because the decisions are bounded — which expense account a charge belongs in, whether to accrue or defer a small item, when a variance is big enough to flag upward — all inside a chart of accounts and closing checklist someone else designed, with the material calls on revenue recognition, reserves, and tax positions escalated to the controller or CPA.
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 (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 13 of this occupation's 21 points (62%).
Embodiment (3/20) 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 39/100 — EXPOSED.
Genuine two-tier structure: if bank-feed rules and LLM categorization absorb the coding/reconciliation tier entirely, the residual role becomes exception investigation, chart-of-account redesign at system migration, multi-state sales-tax nexus judgment (post-Wayfair registrations), and untangling owner-commingled records. Watch for job postings retitled 'close analyst' or 'AP exception specialist' with headcount falling but per-head scope rising — the score rises for whoever remains, not for the occupation's size.
Fidelity-bond and cyber-insurance underwriters increasingly require documented segregation of duties and a named human approver for outbound payments above a threshold. If insurers explicitly exclude losses from AI-auto-approved vendor payments (a live concern after business-email-compromise claims), a named clerk owning the release-payment decision becomes contractually required.
Payroll and sales-tax remittance is the only place personal exposure exists: IRS trust fund recovery penalty (IRC 6672) already reaches 'responsible persons' who control payroll tax deposits, and states like California and New York apply similar personal liability for sales tax. If enforcement guidance or state rules name the person who authorizes the filing rather than only the owner/CPA, a signature role attaches to the clerk.
Fraud detection is the one thing clients pay a human for: small-business embezzlement is overwhelmingly committed by the person with ledger access, and an outsourced human bookkeeper functions as a control. If insurers or lenders require an independent (non-employee) human reconciler as a condition of a line of credit or bond, that is a paid-for human, not a paid-for relationship.
State-level bookkeeper registration on the model of the tax-preparer regimes (California CTEC, Oregon Board of Tax Practitioners, Maryland) extended from tax prep to bookkeeping/payroll services — a bill creating a registered-bookkeeper credential with bonding requirements. No such bill is currently moving; the industry certifications (AIPB CB, NACPB) are voluntary and carry no statutory weight.
The limit. Even with every lever, this stays a low-scoring occupation. The levers protect a shrinking senior fraction — the exception handler, the payroll-tax signer — while the 1.37M headcount is driven by the routine tier that is already automated. Nothing here raises the floor; it only raises the ceiling for a minority. BLS already projects decline, and the liability routes attach to a handful of people per firm, not to the role.
| New York-Newark-Jersey City, NY-NJ | 77,010 | $60,640 +20% |
| Los Angeles-Long Beach-Anaheim, CA | 53,910 | $58,590 +16% |
| Chicago-Naperville-Elgin, IL-IN | 35,660 | $51,440 +2% |
| Dallas-Fort Worth-Arlington, TX | 34,930 | $51,720 +2% |
| Houston-Pasadena-The Woodlands, TX | 26,090 | $49,460 -2% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 25,910 | $50,300 -1% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 22,660 | $52,710 +4% |
| Atlanta-Sandy Springs-Roswell, GA | 22,480 | $51,170 +1% |
| San Jose-Sunnyvale-Santa Clara, CA | 7,590 | $66,130 +31% |
| San Francisco-Oakland-Fremont, CA | 18,100 | $65,530 +29% |
| New Haven, CT | 2,900 | $62,890 +24% |
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.