Skip to sources
Time Machine

Bookkeeping, Accounting, and Auditing Clerks

Scrub through 286years of this role's history, from when it first emerged, through every wave of technology that reshaped it, to the cited projections for where it's heading next.

2026drag to travel through time
17501775180018251850187519001925195019752000now
Country
2026
Known today as Bookkeeping, Accounting, and Auditing Clerks (BLS SOC 43-3031)
US Employment
1.37M
OEWS is a point-in-time survey snapshot, not a continuous time series; BLS advises against using it for year-over-year trend comparison.
Median Annual Wage
$50,670
≈ $49,371 in 2024 dollars
Each dot is a cited figure over time; the dotted line only links them (values between aren't measured). Hollow dots are estimates.
Tools of the era

The tools that defined the work

Select an era to see how it reshaped the work.

  • Quill, steel pen, and handwritten double-entry ledger

    The bookkeeper of the 18th and early 19th century worked entirely by hand: a quill (later a steel pen nib), a ruled ledger book, and the mental arithmetic to keep the columns balanced. Luca Pacioli's 1494 codification of double-entry bookkeeping defined the structure — debit and credit, daybook and ledger — that the counting-house clerk followed unchanged for nearly four centuries. The bottleneck was speed: a fast, careful bookkeeper might post 100 entries a day. Adding up a long column required pencil-scratch arithmetic or counting beads; an error caught late could mean re-posting weeks of entries. The quill era's bookkeeper was the organizational memory of the firm, and the firm could not function without him.

    Effect on the work

    Total US bookkeeper employment roughly doubled each generation from 1820 to 1880 as the American economy industrialized and transaction volumes grew. The limiting factor was the supply of persons who could write legibly and add accurately — not the demand for the work.

    Ledger workPaper recordkeeping
  • Burroughs adding machine — the first practical mechanical bookkeeping aid

    William Seward Burroughs filed his adding machine patent in 1885; the commercial machine was launched in 1886. Unlike earlier adding devices, Burroughs's machine printed a paper tape of each entered figure and produced a reliable running total — the first adding machine that bookkeepers could actually trust in daily work. The Burroughs Adding Machine Company sold over one million machines by 1926. The machine did not replace the bookkeeper; it replaced the pencil-scratch arithmetic that bookkeepers did between ledger entries. A bookkeeper who had spent an hour totaling a 200-line column by hand could now do it in minutes. Per-transaction labor fell; business volume grew; the net effect on bookkeeper headcount was roughly neutral through 1910 as demand expanded to absorb the productivity gain.

    Effect on the work

    Burroughs sold over 1 million machines by 1926; each machine served roughly one bookkeeping department. Individual bookkeeper productivity for arithmetic tasks increased 5-10x, but the expansion of corporate record-keeping absorbed the gain.

    Mechanical calculationTen-key speed
  • NCR Class 2000 and Burroughs Sensimatic — mechanical bookkeeping machines

    By the 1920s a more sophisticated machine had appeared: the mechanical bookkeeping machine, which could post a figure, update a balance, and print an account statement in a single keystroke operation. National Cash Register's Class 2000 series and Burroughs's Sensimatic were the flagship products of this era. A Burroughs Sensimatic operator could post accounts payable faster and more accurately than any hand-posting bookkeeper. Large banks, insurance companies, and department stores installed rooms full of these machines — the bookkeeping machine pools of the 1930s and 1940s were the direct predecessors of the mainframe data centers of the 1960s. The operators were overwhelmingly women, paid less than the male bookkeepers they had partly displaced, and trained at manufacturer-run schools much as Comptometer operators had been.

    Effect on the work

    Mechanical bookkeeping machines concentrated bookkeeping work in large firms with machine pools while reducing headcount in mid-sized firms where one machine could replace several hand-posting clerks. The net occupation headcount continued growing through this era as the corporate economy expanded.

    Mechanical calculationTen-key speed
  • IBM 402 Accounting Machine — electromechanical tabulating for payroll and ledger

    The IBM 402 Alphabetical Accounting Machine, introduced in 1948, was the first mass-market electromechanical accounting machine capable of producing formatted accounting reports — balance sheets, payroll registers, accounts-payable aging reports — directly from IBM punch cards. Unlike the purely mechanical bookkeeping machines of the 1920s-1940s, the 402 separated data input (card punching) from output (printed accounting reports), enabling the first true batch processing of accounting records. A payroll run for a 500-person company that had required two bookkeepers and a Burroughs machine for a week could now be run overnight on the 402. The Social Security Administration's payroll processing requirements had already created the infrastructure for punch-card accounting at scale; the 402 brought that infrastructure to mid-market corporations.

    Effect on the work

    The IBM 402 and its successors (the 407, 1949) began the first genuine reduction in per-account bookkeeping labor at large firms. The bookkeeping machine operators whose jobs disappeared were replaced by card-punch operators — a different form of clerical labor, typically paying less.

    Ledger workPaper recordkeeping
  • IBM System/360 — mainframe AR/AP/payroll in batch

    The IBM System/360, announced April 7, 1964, was the first general-purpose mainframe family that made large-scale, integrated business data processing economically accessible to mid-market corporations. COBOL (1960) running on System/360 hardware became the standard platform for payroll, accounts payable, accounts receivable, and general ledger batch processing through the late 1960s and 1970s. For the first time, the complete accounting cycle — from transaction capture to financial statement production — could run automatically overnight. Bookkeeping clerks in large organizations shifted from posting ledger entries by hand to reviewing exception reports, handling discrepancies, and managing the source documents that fed the batch system. The work became supervisory of a process rather than performative of a task.

    Effect on the work

    Mainframe batch accounting was the first technology wave to produce demonstrable net reductions in bookkeeping clerk headcount at large employers. Aggregate employment in bookkeeping continued growing through this era because mid-market and small firms remained unaffected — the technology was too expensive for firms below a certain scale.

    Mainframe processingComputerized records
  • Peachtree (1978) and Quicken (1983) — PC accounting arrives

    Peachtree Software launched in 1978 — before the IBM PC, before the commercial spreadsheet, running on the CP/M operating system — as the first commercially viable PC-based accounting software for small businesses. When the IBM PC arrived in 1981, Peachtree was there with an MS-DOS version. Intuit launched Quicken in 1983, targeting personal finance; Intuit founder Scott Cook has described watching his wife manually pay bills as the inspiration. By the mid-1980s, a small-business owner with an IBM PC and Peachtree could manage accounts payable, accounts receivable, payroll, and general ledger without a dedicated bookkeeping clerk. The PC accounting era began the slow erosion of bookkeeping employment in small businesses — the firms that had never been able to afford mainframe systems but now had a $2,000 alternative.

    Effect on the work

    PC accounting software reduced the bookkeeping-clerk headcount needed per small-business transaction by an estimated 30-50% compared to purely manual bookkeeping. The effect was offset at the aggregate level by the overall expansion of small businesses during the 1980s Reagan economy.

    Work toolChanging equipment
  • QuickBooks (1992) — dominant SMB accounting software

    Intuit launched QuickBooks in 1992 at approximately $140 — an order of magnitude cheaper than Peachtree's professional editions and explicitly designed for business owners with no accounting background. Within a decade, QuickBooks claimed approximately 85% of the US small-business accounting software market. Inc. Magazine named it the best-selling business software of 1994. QuickBooks did what Peachtree had begun: it moved the routine bookkeeping tasks of millions of small businesses — invoicing, bill payment, bank reconciliation, payroll — into software that required no dedicated bookkeeping clerk to operate. Businesses that had employed a part-time bookkeeper now had software. Businesses that had employed a full-time bookkeeper now employed a part-time one.

    Effect on the work

    QuickBooks is widely cited in BLS OOH historical analysis as a primary driver of bookkeeping employment erosion in small businesses during the 1990s. The occupation reached its modern peak headcount around 2002, just as cloud accounting was beginning to arrive — suggesting that QuickBooks alone did not kill the role, but it set the ceiling.

    Accounting softwareIntegrated ledgers
  • Xero (2006) and QuickBooks Online — cloud accounting + bank feeds

    Xero, incorporated July 6, 2006 in Wellington, New Zealand by Rod Drury and Hamish Edwards, was the first cloud-native accounting platform built around automated bank feeds — the feature that imports and auto-categorizes bank transactions without manual data entry. Bank feeds changed the daily work of the bookkeeping clerk from entering transactions to reviewing and approving auto-categorized ones. QuickBooks Online had launched in 2004 and by the early 2010s was being aggressively marketed alongside bank-feed partners. Receipt Bank (founded 2010, later rebranded Dext) brought OCR-based receipt capture to the workflow: a photo of a receipt became an auto-categorized expense entry. The cloud shift meant that a bookkeeper in 2015 spent their day reviewing a pre-populated ledger rather than building it from scratch.

    Effect on the work

    Cloud accounting with bank feeds reduced per-client monthly bookkeeping time by an estimated 50-70% compared to manual QuickBooks desktop entry, per accounting-firm benchmarking data cited in the 2013-2016 period. Employment fell from the 2002 peak through this period without recovery — structural, not cyclical.

    Accounting softwareIntegrated ledgers
  • AI auto-categorization — QuickBooks Live, Pilot.com, Bench, Vic.ai

    By 2017, the combination of machine learning and cloud accounting data had reached the point where AI could auto-categorize the majority of transactions without human review — not just pattern-match against prior periods, but learn a business's chart-of-accounts and make categorization decisions that previously required a bookkeeper's judgment. Pilot.com (founded 2017 by Waseem Daher, Jeff Arnold, and Jessica McKellar) used a human-plus-AI hybrid model: software handled routine transactions, human bookkeepers handled exceptions and client communication. Bench Accounting, which had raised $113 million including from Shopify and Bain Capital, used a similar hybrid model for 35,000+ US customers. QuickBooks Live Bookkeeping (Intuit, 2019) added a human bookkeeper layer on top of QuickBooks AI. The AI layer in all these models was doing what a junior bookkeeping clerk had previously done; the human layer was doing what a senior bookkeeping clerk or junior accountant had done.

    Effect on the work

    AI-assisted bookkeeping platforms reduced the human-hours of bookkeeping required per small business by an additional 40-60% beyond what cloud + bank feeds had already achieved. Pilot.com, Bench, and similar services were explicitly substituting AI-plus-one-human for the traditional human bookkeeping workforce.

    AI audit toolsPattern detection
  • LLM-based bookkeeping — QuickBooks Intuit Assist, Xero JAX, BILL AI, Stampli

    The large-language-model wave beginning in 2023 brought generative AI into the bookkeeping workflow at every major platform simultaneously. QuickBooks Intuit Assist auto-categorizes transactions, generates invoice descriptions, and answers questions about financial records in natural language. Xero's JAX (Just Automatic Xero) automatic bank reconciliation, launched in beta November 2025, targets automating 80-90% of bank statement lines without human input. BILL.com's Invoice Coding Agent reads invoices, codes them to GL accounts by historical pattern, and increased touchless (fully automated) invoice processing by 80%+ in FY2025-26 per SEC filings. Stampli's Billy the Bot achieves 95%+ GL-code accuracy with 3-way PO matching. The work that defined the bookkeeping clerk for 250 years — reading a source document and coding it to the right account — is now largely performed by software. The human role shifts to exception review, vendor escalation, and client relationship management: the things the software cannot do.

    Effect on the work

    BLS OOH 2024 explicitly cites software automation as the primary driver of projected -6% employment decline for 43-3031 through 2033-34. The WEF Future of Jobs 2025 names accounting, bookkeeping, and payroll clerks the fastest-declining occupational cluster globally through 2030. BILL.com SEC filings document 8 million fraud attempts blocked by AI — tasks previously requiring human review.

    AI audit toolsPattern detection
Projection cone · present → 2034

What credible sources project

Scrub the slider past now to anchor each scenario on the scrubber. The spread is the range of futures credible sources project for this role.

Employment outlook
Projected change in the number of people doing this work.
BLS Occupational Outlook Handbook 2023-33
2033
-6%
BLS Employment Projections 2023-33 cycle. Published employment change for SOC 43-3031: -6% (approximately -98,000 jobs), from a base of approximately 1,642,600 (2023). BLS OOH explicitly cites "software innovations" and accounting automation as the primary driver. This is the most conservative mainstream projection — the BLS tends to lag real-world adoption curves, and its methodology does not model AI-adoption acceleration beyond current productivity trends. Annual average job openings: approximately 183,900 (dominated by replacement need from retirements and occupational exits, not net growth).
BLS National Employment Matrix 2024-34
2034
-6%
BLS Employment Projections 2024-34 cycle (most current). Projects continued -6% decline consistent with 2023-33 cycle, from a baseline of approximately 1,600,000 (2024). The BLS National Employment Matrix for 43-3031 is one of the few major occupations where BLS projects outright employment decline rather than growth or flat. The consistency across two projection cycles suggests BLS treats this as a structural trend, not a temporary disruption.
WEF Future of Jobs Report 2025
2030
-26%
WEF survey of 1,000+ employers across 22 industry clusters in 55 economies (January 2025). Names "Accounting, Bookkeeping, and Payroll Clerks" the fastest-declining occupational cluster globally through 2030, projecting approximately 26% net employment decline — among the steepest in the entire WEF dataset. The WEF methodology is employer-stated-intent rather than econometric; the result reflects what actual hiring managers say they plan to hire fewer of, which is arguably more predictive than model-based projections for near-term horizons. This is the pessimistic tail of the uncertainty cone.
AI task exposure
Share of the role’s tasks that researchers estimate AI can do. This is a measure of task exposure, not a forecast of jobs lost.
Frey & Osborne (2013)
2033
98%
of tasks
Gaussian-process classifier on O*NET task features. Frey & Osborne rated Bookkeeping, Accounting, and Auditing Clerks at approximately 0.98 probability of computerization — one of the highest of all 702 occupations studied. The occupation scored near the ceiling on every bottleneck criterion: high numerical facility requirement, low creative intelligence requirement, low social intelligence requirement, and high rule-following structure. Tax Preparers (13-2082) at 0.99 were the only related occupation rated higher. Reported here as -98% to represent the F&O ceiling finding, not a precise employment-loss forecast. In practice, employment has declined by ~25% in the 12 years since the paper; F&O appears to have correctly identified direction and degree of risk while underestimating adoption speed on the one hand and overestimating how completely the occupation would vanish on the other.
Eloundou et al. — "GPTs are GPTs" (2023)
2028
20%
of tasks
GPT-4 task-by-task LLM exposure labeling on O*NET tasks for 43-3031. Bookkeeping, Accounting, and Auditing Clerks score very high on LLM exposure because the core tasks — transaction coding, ledger reconciliation, report generation, invoice processing — are information-structured, rule-governed tasks that LLMs handle well. Secondary-source convention places the β (E1 + 0.5×E2) exposure metric for this occupation among the top decile of all US occupations. The -20% estimate represents the near-term displacement component (5-year horizon) from AI adoption at the pace observed in 2023-2025, not the full theoretical exposure ceiling.
Today, in this role

What's shifting in the work right now

The historical view above shows how this role has moved. This is the present-day detail: which AI tools are picking up which tasks, where the edge still is, and the natural directions this work can grow.

What's changing in your day

Three parts of your work where AI is already doing real lifting, and what stays yours.

AI is taking this onProcess accounts payable invoices by overseeing AI autonomous coding and 3-way PO matching in BILL or Stampli (Billy the Bot)

Process accounts payable invoices by overseeing AI autonomous coding and 3-way PO matching in BILL or Stampli (Billy the Bot); review the exception queue for invoices AI could not auto-approve; escalate disputed amounts or unfamiliar vendors to the appropriate manager.[9],[10],[11]

Where your edge is

Shift from processing invoices to owning the exception queue and vendor escalation workflow — the 10-20% of invoices that AI flags are where human judgment is irreplaceable. Build skills in reading AI confidence scores and audit trails so you can override or retrain the model with well-documented reasoning.

AI is taking this onCapture and categorize employee receipts and corporate card transactions using AI expense platforms (Ramp, Expensify SmartScan)

Capture and categorize employee receipts and corporate card transactions using AI expense platforms (Ramp, Expensify SmartScan); review flagged policy violations and out-of-policy spend; approve or reject AI-categorized expenses before ERP sync.[12],[13],[14]

Where your edge is

Ramp and Expensify now auto-categorize and auto-code 85-90% of transactions including VAT/GST extraction. Develop skills in configuring spend policies within these platforms and in conducting exception-based audits — the value-add is building controls, not processing receipts.

AI is taking this onPerform bank reconciliation by reviewing transactions auto-matched by Xero JAX or QuickBooks Intuit Assist

Perform bank reconciliation by reviewing transactions auto-matched by Xero JAX or QuickBooks Intuit Assist; investigate the 10-20% of statement lines flagged for human review; investigate timing differences, uncleared checks, and undeposited funds that the AI model marks as ambiguous.[7],[15]

Where your edge is

Xero's JAX targets 90% of bank reconciliation automated without manual input by end of 2026. Redirect your time to reviewing the AI's reconciliation audit trail, catching edge-case mismatches (refunds split across periods, foreign-currency variances), and documenting exception rationale for the client's records.

Where this role is heading

Natural next steps for someone with your foundation: not exits, evolutions.

A direction you could grow

Accountants and Auditors

The classic bookkeeper-to-accountant path is the most defensible career move as AI absorbs the clerical execution layer. Accountants hold the professional judgment, client advisory, and regulatory sign-off roles that automation leaves intact. The CPA exam is the gate; the EA (Enrolled Agent) is a faster intermediate credential for those on a tax path. Robert Half data shows accountants and auditors are in the highest demand of any finance occupation in 2026, with starting salaries 30-50% above full-charge bookkeeper rates.

What you'd add
  • · Enrolled Agent (EA) exam — fastest credential path, IRS-issued, no degree required
  • · GAAP foundational concepts: accrual accounting, revenue recognition, adjusting entries
  • · Financial statement analysis: reading and interpreting balance sheets, income statements, and cash flow
  • · QuickBooks ProAdvisor or Xero Advisor certification — valued by small-business accountants
  • · Two-year accounting degree or equivalent course work toward CPA eligibility
What it takesSome new skills to pick up
Share this year
Drops anyone you send it to straight into 2026.
Preview card
Part of Admin & Office · see all 39roles →
Different role?

See the same long-arc view for your own profession.

Browse the directory by industry, or search by title or SOC code. New roles ship every few weeks. Every profile cites every claim.

Browse all roles

The data behind this timeline

On record since1750
Latest tracked employment1,373,680 (US, 2025)
Latest median pay$50,670 (2025)
Outlook-6% by 2033 (BLS Occupational Outlook Handbook 2023-33)
View all 34 cited data points
YearUS employmentMedian annual paySource
187038,000n/aCENSUS-IPUMS
1900254,000n/aCENSUS-DECENNIAL
1910n/a$750BLS-HISTORICAL-BULLETIN
1930750,000n/aCENSUS-DECENNIAL
19601,200,000$4,200ESTIMATE
19801,700,000n/aESTIMATE
1982n/a$13,000BLS-CPS
1984n/a$13,500BLS-CPS
1988n/a$16,000BLS-CPS
1999n/a$25,865CENSUS
20022,093,000$28,900BLS-OEWS
20031,750,680$27,760BLS-OEWS
20041,770,860$28,570BLS-OEWS
20051,815,340$29,490BLS-OEWS
20061,856,890$30,560BLS-OEWS
20071,858,500$31,560BLS-OEWS
20081,855,010$32,510BLS-OEWS
20091,757,870$33,450BLS-OEWS
20101,898,000$34,030BLS-OEWS
20111,643,470$34,740BLS-OEWS
20121,606,260$35,170BLS-OEWS
20131,586,380$35,730BLS-OEWS
20141,575,060$36,430BLS-OEWS
20151,760,000$37,250BLS-OEWS
20161,566,960$38,390BLS-OEWS
20171,532,340$39,240BLS-OEWS
20181,530,430$40,240BLS-OEWS
20191,512,660$41,230BLS-OEWS
20201,443,940$42,410BLS-OEWS
20211,509,370$45,560BLS-OEWS
20221,550,750$45,860BLS-OEWS
20231,642,600$47,440BLS-OEWS
20241,600,000$48,050BLS-OEWS
20251,373,680$50,670BLS-OEWS
Embed this timeline on your site

Free for any site. Paste this where the timeline should appear; it stays interactive, every datapoint stays cited, and it sets no cookies on your page. How embedding works

<iframe src="https://futurehistory.earth/embed/43-3031"
  width="100%" height="430" style="border:0"
  title="Bookkeeping, Accounting, and Auditing Clerks, a Future History timeline"
  loading="lazy"></iframe>

See all roles in Admin & Office