Bookkeeping, Accounting, and Auditing Clerks
43-3031.00Median wage $50,670/yr1,373,680 employed (US)Rank #3 of 923 scored · top 1% by substitution
Compute, classify, and record numerical data to keep financial records complete. Perform any combination of routine calculating, posting, and verifying duties to obtain primary financial data for use in maintaining accounting records. May also check the accuracy of figures, calculations, and postings pertaining to business transactions recorded by other workers.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution81
Exposure84
Augmentation78
Substitution — the headline: capability discounted by cost, barriers and adoption.
Exposure — technical capability alone, regardless of whether anyone deploys it.
Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.
Tasks on the substitution scale
28 rated tasks, binned by substitution score.
Position among all scored occupations
Distribution of 923 occupation scores; the marker is this occupation.
Tasks with substitution ≥ 70
86%
Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.
Why this score
The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.
Task automatabilityw 35%84
panel mean rating 4.4/5 → substitution pressure 84/100
Technical feasibility todayw 20%83
panel mean rating 4.3/5 → substitution pressure 83/100
Cost vs. human wagew 15%88
panel mean rating 4.5/5 → substitution pressure 88/100
Adoption barriersw 20%inverted — strong barriers lower the score71
panel mean rating 2.2/5 (barrier strength) → substitution pressure 71/100
Sector adoption velocityw 10%78
panel mean rating 4.1/5 → substitution pressure 78/100
Task breakdown (28 tasks)
Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.
Transfer details from separate journals to general ledgers or data processing sheets.
100CI 100–100 · exposure 100 · augmentation 50 · importance 4.2/5 · click for rater detail
Transfer details from separate journals to general ledgers or data processing sheets.
100| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Accounting software adoption is near-universal in professional services and enterprises; automated posting is already standard practice in most digitized firms, reflecting very fast sectoral penetration. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Accounting and bookkeeping functions across finance and professional services have deeply and rapidly adopted automated ledger systems, with manual journal transfer now rare in digitized firms. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by auto-populating ledger fields, flagging discrepancies, and suggesting categorizations, improving clerk productivity. However, the core task is so automatable that augmentation is a secondary benefit rather than the primary use case. |
| Augmentation potential | claude-sonnet-5 | 3/5 | For firms still using partial manual processes, AI/software tools speed up and reduce errors in transferring entries, though the task is largely already automated rather than merely augmented. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Transferring journal entries to general ledgers is a highly structured, rule-based data transfer task with clear inputs and outputs. Modern accounting software and RPA systems can perform this end-to-end with >50% time savings at equal or superior accuracy compared to manual entry. |
| Task automatability | claude-sonnet-5 | 5/5 | This is a highly structured, rules-based data transfer task that modern accounting software and AI-driven automation already perform end-to-end with substantial time savings at equal or better accuracy. |
| Adoption barriers | claude-haiku-4-5-20251001 | 1/5 | No legal licensure requirement or human sign-off mandate exists for ledger posting itself; it is a clerical data transfer task. Organizational adoption is driven by cost and convenience, not regulation. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or legal requirement mandates a human perform this mechanical data-transfer step; it is routinely automated without regulatory obstruction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | API-driven posting costs pennies per transaction and requires minimal ongoing oversight, while manual clerk labor costs $25–50/hour for similar volume. AI cost is easily 10–100× cheaper at scale. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated posting via software costs a small fraction of a cent per transaction compared to a clerk's hourly wage for manual data transfer. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed products (QuickBooks, Xero, NetSuite, SAP) and RPA platforms routinely perform automated journal-to-ledger posting in production environments. This task is among the most mature automation domains in accounting software. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Mature accounting products (QuickBooks, Xero, SAP, NetSuite) already automate journal-to-ledger posting in production at massive scale, largely eliminating manual transcription. |
Operate 10-key calculators, typewriters, and copy machines to perform calculations and produce documents.
97CI 95–100 · exposure 100 · augmentation 63 · importance 4.5/5 · click for rater detail
Operate 10-key calculators, typewriters, and copy machines to perform calculations and produce documents.
97| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Accounting and finance sectors have been among the earliest and deepest adopters of automation, with widespread deployment of accounting software, RPA, and cloud document systems across small and large firms. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | This task is already largely obsolete in practice, having been replaced by digital tools across nearly all office and accounting environments long ago. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered accounting software and document tools substantially raise clerk productivity by automating routine calculations and formatting, while allowing humans to focus on reconciliation, validation, and judgment-based tasks. |
| Augmentation potential | claude-sonnet-5 | 3/5 | While AI/software has replaced the manual equipment described, the rationale for augmentation is limited since the task itself (physical machine operation) is largely anachronistic rather than one modern workers still perform with AI assistance. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Current AI systems can perform calculations end-to-end and can generate documents automatically via APIs to office equipment or cloud storage, easily exceeding 50% time savings at equal quality compared to manual 10-key and typewriter operation. |
| Task automatability | claude-sonnet-5 | 5/5 | Basic calculation and document production tasks are trivially handled by spreadsheet software and modern office tools, which have already replaced manual calculator/typewriter use in virtually all settings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While some organizations require human sign-off on financial records for audit purposes, the mechanical task of operating calculators and producing documents faces minimal legal barriers; oversight is typically lightweight and procedural rather than hard-coded. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There is no licensing, liability, or regulatory requirement tied to using a calculator or typewriter; nothing legally requires human operation of this equipment. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Cloud-based accounting and document automation services cost a fraction of the loaded wage for a clerk's manual calculator and typing work, easily achieving order-of-magnitude savings. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Software-based calculation and document generation costs are negligible compared to paying a human to operate physical 10-key machines or typewriters. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed products—accounting software (QuickBooks, Xero), OCR systems, RPA tools, and cloud document generation—already perform these functions reliably in production at scale across thousands of organizations. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Digital calculators, word processors, and copiers/scanners are mature, ubiquitous production tools that have essentially already automated this literal task. |
Compare computer printouts to manually maintained journals to determine if they match.
95CI 92–97 · exposure 100 · augmentation 75 · importance 4.2/5 · click for rater detail
Compare computer printouts to manually maintained journals to determine if they match.
95| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Accounting and financial services sectors are rapidly adopting automated reconciliation and RPA; many firms have already deployed systems for this exact task in production, reflecting high digital maturity and ROI incentives. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Accounting and finance functions have rapidly adopted automated reconciliation and ERP-integrated matching tools, reflecting the fast-adopting finance/professional-services pattern. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI augments clerks by flagging discrepancies, suggesting corrections, and highlighting unusual patterns, enabling human reviewers to focus on exceptions and judgment calls rather than routine verification. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Even where full automation isn't implemented, AI-assisted tools quickly highlight mismatches and anomalies, substantially speeding up a clerk's manual comparison work. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Comparing computer printouts to journals is a high-volume data reconciliation task that AI can perform end-to-end via optical character recognition, data extraction, and automated matching algorithms, easily exceeding 50% time savings at equal or better accuracy. |
| Task automatability | claude-sonnet-5 | 5/5 | Automated reconciliation of digital records against ledger entries is a straightforward data-matching task well within current AI and even simple scripting capabilities, easily exceeding 50% time savings at equal or better accuracy. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While some organizations require manual sign-off for compliance or audit trails, there are no hard legal barriers preventing AI from performing the comparison itself; most friction is organizational preference rather than regulatory mandate. |
| Adoption barriers | claude-sonnet-5 | 1/5 | This is a purely clerical verification task with no licensing requirement or legal need for human sign-off, so no meaningful regulatory or liability barrier exists. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Inference cost for document processing and comparison is negligible; integration into accounting systems is mature; oversight is minimal once validated—this is orders of magnitude cheaper than a clerk's hourly wage for equivalent reconciliation volume. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated matching software runs at near-zero marginal cost per comparison versus the loaded wage of a clerk manually cross-checking printouts line by line. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed products including accounting software integrations, RPA platforms, and specialized audit tools reliably perform automated reconciliation in production across accounting firms and enterprises today. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Reconciliation software and RPA tools are already deployed at scale in accounting departments to automatically match transaction records against journals, flagging discrepancies reliably in production. |
Calculate, prepare, and issue bills, invoices, account statements, and other financial statements according to established procedures.
95CI 95–95 · exposure 100 · augmentation 75 · importance 4.2/5 · click for rater detail
Calculate, prepare, and issue bills, invoices, account statements, and other financial statements according to established procedures.
95| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Bookkeeping and invoicing automation is deeply embedded in enterprise finance operations, with widespread deployment of accounting software, RPA, and AI-driven invoice generation across finance, professional services, and mid-market organizations. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Financial and administrative back-office functions have among the fastest, deepest software/AI adoption rates, with automated billing being standard practice across nearly all sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered accounting assistants and software augment clerk productivity by auto-categorizing transactions, flagging discrepancies, suggesting corrections, and accelerating statement review, enabling humans to focus on exception handling and analytical work. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly speed up preparation, formatting, and error-checking of invoices and statements, though clerks often still review or adjust outputs for accuracy and exceptions. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Modern accounting software and AI systems can fully automate bill and invoice generation, statement preparation, and financial statement compilation from structured transaction data with 50%+ time savings and equal or superior accuracy compared to manual entry. |
| Task automatability | claude-sonnet-5 | 5/5 | Invoice/billing generation from structured data follows well-defined rules that current accounting software and AI-augmented tools already automate end-to-end with substantial time savings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While financial record-keeping is regulated and requires audit trails, the task itself (bill/invoice calculation and generation) does not legally require human signature or authorization in most jurisdictions; internal controls and oversight are needed but do not mandate human performance. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for issuing invoices; some organizational preference for human review of financial statements before external issuance, but automation is already normalized in this process. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Cloud-based accounting software and AI invoice processing cost a fraction of a human clerk's loaded wage, with infrastructure amortized across many users; inference and integration are negligible compared to even entry-level accounting labor. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated billing software costs a fraction of a cent per invoice compared to clerk labor costs, representing well over an order-of-magnitude cost advantage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature accounting platforms (QuickBooks, NetSuite, Xero, SAP) and financial software with RPA and AI modules demonstrably perform invoice generation, statement preparation, and billing at scale in production across thousands of organizations daily. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Mature production systems (QuickBooks, NetSuite, SAP, Bill.com, AI-enhanced ERP modules) already generate invoices, statements, and billing documents reliably at scale in real organizations. |
Prepare trial balances of books.
95CI 95–95 · exposure 100 · augmentation 75 · importance 4.1/5 · click for rater detail
Prepare trial balances of books.
95| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Accounting and finance sectors have high digital maturity and fast AI adoption. Cloud-based accounting platforms with automated trial balance features are the norm in mid-market and enterprise; this task is already substantially automated in production. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Accounting and bookkeeping is a heavily digitized, software-native sector where automated trial balance generation has been standard practice for years. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assists by auto-generating trial balances for review, flagging discrepancies, and suggesting reconciliation actions, allowing clerks to focus on analysis and exception handling rather than mechanical compilation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI/software handles the compilation while clerks shift to reviewing anomalies, reconciling discrepancies, and interpreting results, meaningfully boosting productivity. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Preparing trial balances is a highly structured, rule-based task involving extracting account balances and summing debits/credits. Modern accounting software and AI systems can fully automate this end-to-end from GL data with >50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 5/5 | Preparing trial balances is a mechanical aggregation of ledger balances that accounting software already automates natively, easily meeting the 50% time-saving threshold. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No legal requirement for a human to personally prepare trial balances; internal process. Minor friction exists around audit trail documentation and GL data quality governance, but these are process matters, not legal barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement to compile a trial balance itself (unlike audit sign-off), though firms still want human review before it feeds financial statements. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Once GL data is in a system, trial balance generation is near-zero marginal cost—a few dollars of cloud compute or an included feature in software already licensed, versus 1–2 hours of clerk labor (~$30–50 all-in). |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Once ledger entries exist, generating a trial balance costs near-zero incremental compute versus the manual clerical time historically required, an order of magnitude cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Accounting software (QuickBooks, Xero, NetSuite, SAP) and ERP systems have mature, production-grade trial balance generation built-in, deployed across millions of organizations globally with reliable, auditable outputs. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Mature, widely deployed accounting systems (QuickBooks, Xero, SAP, NetSuite) generate trial balances automatically and reliably at scale in production today. |
Perform financial calculations, such as amounts due, interest charges, balances, discounts, equity, and principal.
94CI 92–95 · exposure 100 · augmentation 88 · importance 4.3/5 · click for rater detail
Perform financial calculations, such as amounts due, interest charges, balances, discounts, equity, and principal.
94| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Accounting and bookkeeping have undergone deep, rapid digital transformation. Virtually all organizations use automated accounting systems for these exact calculations. Adoption is mature and industry-wide. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Finance and accounting is a fast-adopting sector for automation, with calculation tasks already largely software-driven and AI-enhanced tools rapidly deployed. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI augments human accountants by automating routine calculations and freeing them to focus on reconciliation, anomaly detection, and higher-judgment tasks. The human remains in the verification loop, with AI handling the computational burden. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI and spreadsheet tools massively speed up and reduce errors in these calculations while accountants retain oversight and interpretation responsibilities. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Financial calculations of amounts due, interest charges, balances, discounts, equity, and principal are deterministic arithmetic operations on structured data. Current AI systems and accounting software can perform these end-to-end with formula-based automation, achieving well over 50% time savings at equal or superior accuracy. |
| Task automatability | claude-sonnet-5 | 5/5 | These are well-defined, formulaic calculations (interest, discounts, balances) that spreadsheet software and AI-augmented accounting tools already automate with high reliability and speed. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While financial calculations must ultimately be reviewed and the output signed off, there are minimal hard barriers to automating the calculation itself. No licensing requirement applies to the arithmetic step; a human typically audits results rather than performs calculations. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for basic calculations, though some oversight/audit trail expectations exist in regulated financial reporting contexts. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | The per-transaction cost of automated calculation is negligible (pennies per hundred calculations), whereas human clerks cost $25–$35/hour loaded. Automation is orders of magnitude cheaper for routine financial arithmetic. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated calculation via software costs fractions of a cent per transaction versus a clerk's hourly wage, an order-of-magnitude or greater saving. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature, production-grade accounting software (QuickBooks, SAP, NetSuite, Xero) and spreadsheet AI tools routinely perform these calculations reliably at scale in real organizations. These tasks have been automated for decades and are core to deployed financial systems. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Standard accounting software (QuickBooks, Excel, ERP systems) has performed these calculations reliably in production for decades, with AI now automating the surrounding classification and data entry. |
Calculate and prepare checks for utilities, taxes, and other payments.
91CI 87–95 · exposure 100 · augmentation 75 · importance 4.4/5 · click for rater detail
Calculate and prepare checks for utilities, taxes, and other payments.
91| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Accounting and finance are among the fastest adopters of automation. AP automation, RPA for finance, and cloud accounting platforms with integrated payment processing have achieved deep penetration in mid-market and enterprise sectors. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Accounting and finance back-office functions have seen fast, broad adoption of automated AP/bill-pay tools across firms of many sizes over the past decade. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI systems assist by auto-populating payment amounts, flagging errors, suggesting payment dates, and generating compliance reports. The human clerk's role shifts to oversight and exception handling, substantially raising throughput per hour spent. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Even where humans remain involved for oversight or approval, AI/software tools substantially speed up calculation and preparation, letting clerks focus on exceptions and review. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | This task involves deterministic calculation and document generation based on structured input data (invoice amounts, tax rates, payment schedules). Modern accounting software and AP automation tools routinely perform this end-to-end with 70%+ time savings compared to manual check preparation. |
| Task automatability | claude-sonnet-5 | 5/5 | This is a highly structured, rules-based task (calculating amounts and preparing payments) that is already well-handled by accounting software and automated payment/bill-pay systems with minimal human intervention. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | While the task itself requires no licensing, organizations often impose internal controls, multi-level approval workflows, and audit trails that add oversight friction. Most companies require human authorization before payments post, slowing full substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | Some organizational controls (approval workflows, segregation of duties) create friction, but no licensing or legal requirement mandates a human calculate/prepare these payments. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | A check-writing automation system costs pennies per transaction after initial setup, versus a clerk at loaded cost (~$50–60k annually) processing dozens of checks manually. AI/RPA solutions are an order of magnitude cheaper per payment. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated bill-pay and AP software cost a small fraction of clerk wages per transaction once set up, especially at scale, representing an order-of-magnitude cost reduction. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature products like ADP, Workday, Bill.com, and SAP are deployed at scale in production environments to automatically calculate payments and generate checks with high reliability. This is a core function of enterprise accounting systems used by millions of organizations daily. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Mature, widely deployed products (QuickBooks, Bill.com, ERP AP modules) reliably automate calculation and generation of payments for utilities, taxes, and vendors in production today. |
Reconcile records of bank transactions.
91CI 87–95 · exposure 95 · augmentation 88 · importance 4.4/5 · click for rater detail
Reconcile records of bank transactions.
91| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Accounting is highly digitized and early-adopting sector; automated bank reconciliation is now table stakes in finance and accounting operations, with widespread production deployment across small and large firms. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Accounting and finance functions have adopted automated reconciliation tools broadly and quickly, as this is one of the most mature use cases in finance software adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI reconciliation tools assist humans by pre-flagging exceptions, proposing matches, and highlighting suspicious patterns, allowing clerks to focus on exceptions and complex cases rather than routine matching—substantially raising productivity even when humans remain in review. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools flag mismatches, suggest matches, and surface anomalies, substantially speeding up the clerk's remaining review and exception-handling work. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Reconciling bank transactions is highly structured and rule-based: matching transactions, identifying discrepancies, and flagging anomalies. Current AI systems (accounting software, RPA, and LLM-agents with database access) can process bank statements, match ledger entries, and flag exceptions end-to-end with >50% time savings at equal or better quality. |
| Task automatability | claude-sonnet-5 | 5/5 | Bank reconciliation is highly structured, rule-based matching of transactions against ledgers, which off-the-shelf accounting software and AI-enhanced tools already automate end-to-end for most standard cases. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No legal requirement mandates human sign-off on transaction reconciliation itself; regulations require accurate records but not manual reconciliation. Adoption is primarily organizational inertia and legacy system integration, not hard barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates a human perform reconciliation itself, though some internal control frameworks require human sign-off on discrepancies and final approval, creating mild friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated reconciliation via APIs or RPA costs pennies per transaction and requires minimal oversight, whereas a human clerk performing the same task at median wage ($40–50k/year) costs dollars per transaction when loaded costs are included. The cost advantage is an order of magnitude or more. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated reconciliation software costs a small fraction of clerk hours per transaction volume, especially at scale, making AI dramatically cheaper than manual reconciliation. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed accounting platforms (QuickBooks, Xero, NetSuite) and bank APIs offer automated reconciliation as standard features. Production systems routinely reconcile millions of transactions daily with high accuracy; manual intervention is typically <5% of transactions. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Products like QuickBooks, Xero, and dedicated reconciliation platforms (BlackLine, Trintech) reliably auto-match the vast majority of transactions in production today, though exceptions and unusual entries still require human review. |
Match order forms with invoices, and record the necessary information.
90CI 87–92 · exposure 95 · augmentation 75 · importance 4.1/5 · click for rater detail
Match order forms with invoices, and record the necessary information.
90| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Accounting and finance are fast-adopting sectors for AP automation and invoice processing; major enterprises and mid-market firms already deploy such solutions, indicating strong and accelerating adoption momentum. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Accounts payable automation is widely adopted across finance and professional services, a sector with fast, deep AI/software adoption patterns. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assistants can significantly speed up the human's ability to identify discrepancies, flag exceptions, and complete reconciliation by pre-matching documents and highlighting flagged items, raising productivity while the accountant remains in control. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly speeds up matching and flags discrepancies, letting clerks focus on exceptions and judgment calls rather than routine matching. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Matching order forms with invoices and recording information is a highly structured, rule-based task involving document comparison and data extraction—core strengths of current AI. OCR combined with rule-based matching and database entry can achieve >50% time savings at equal or better quality with existing tools. |
| Task automatability | claude-sonnet-5 | 5/5 | Matching order forms to invoices and recording data is a structured, rules-based data-matching task that off-the-shelf RPA/AI document-processing tools already handle end-to-end with substantial time savings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While internal controls and audit trails are required, there are few legal barriers preventing automation of matching and recording. Oversight and reconciliation steps remain but do not require a licensed human to perform the core matching task itself. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this clerical task; main friction is internal controls and exception handling, not legal or regulatory barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | The cost of AI-driven document matching and data entry (inference + integration) is orders of magnitude cheaper than human labor for this repetitive, volume-based task, especially when amortized across many transactions. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated matching software costs a small fraction of a clerk's hourly wage per transaction processed at scale, especially for high-volume invoice matching. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature products (invoice processing platforms, RPA tools, accounting software with AI-powered matching) reliably perform this task in production across finance departments today. Systems like Coupa, SAP Ariba, and others handle high volumes with measurable accuracy rates. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Mature products (e.g., AP automation platforms with OCR/ML matching like Bill.com, Tipalti, SAP Concur) perform three-way matching reliably in production, though exceptions still require human review. |
Code documents according to company procedures.
86CI 79–92 · exposure 87 · augmentation 75 · importance 4.3/5 · click for rater detail
Code documents according to company procedures.
86| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Accounting and finance sectors show rapid automation adoption; cloud-based accounting platforms and RPA are widely deployed in mid-market and enterprise firms. Small firms lag, but the trend is clearly toward automated coding in digitalized workflows. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Finance and accounting functions across many industries have rapidly adopted automated bookkeeping/document coding tools as part of broader fintech and ERP digitization trends. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI classification systems assist human reviewers by pre-coding high-confidence documents and flagging uncertain ones for manual review, significantly reducing human effort while maintaining control over edge cases and procedure changes. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Even where full automation isn't used, AI tools significantly speed up clerks' coding work by suggesting categories and flagging anomalies, keeping a human in the loop for exceptions. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Document coding against company procedures is highly rule-based and repetitive, making it well-suited to automation. Current OCR and classification systems can extract document content and apply coded categories with substantial time savings, though edge cases and ambiguous documents may require human review. |
| Task automatability | claude-sonnet-5 | 5/5 | Coding documents (invoices, receipts, transactions) into standardized categories/GL codes is a rules-based classification task that current AI/OCR+ML systems handle end-to-end with substantial time savings at equal or better accuracy. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No licensing requirement or legal mandate exists for human review of document coding. The main friction is organizational inertia, audit trail transparency, and the need for occasional override when procedures change or edge cases arise. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal sign-off is required for document coding; the main friction is internal quality control and occasional exception handling, but no hard regulatory barrier prevents automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Once integrated, per-document classification costs pennies via cloud APIs and RPA tools, versus the loaded wage of a clerk to manually code documents. The cost advantage is substantial and grows with volume. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated document coding via OCR/ML costs a small fraction of a cent to a few cents per document versus a clerk's loaded hourly wage for manual data entry, an order-of-magnitude or greater saving. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Multiple products (tax software, accounting platforms like Xero and QuickBooks, document classification APIs) perform this task in production with reliable accuracy on standard documents. Some manual oversight remains common, but the core functionality is deployable at scale. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Mature commercial products (e.g., QuickBooks, Xero, Bill.com, SAP Concur with AI-based auto-coding) already perform document coding reliably in production at scale for many organizations. |
Debit, credit, and total accounts on computer spreadsheets and databases, using specialized accounting software.
85CI 75–95 · exposure 87 · augmentation 75 · importance 4.2/5 · click for rater detail
Debit, credit, and total accounts on computer spreadsheets and databases, using specialized accounting software.
85| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Accounting and finance are among the most digitized, automation-forward sectors. Adoption of automated entry and reconciliation tools is deep and accelerating, with many firms already in production deployment rather than pilot phase. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Accounting and bookkeeping software with automated transaction categorization is already broadly adopted in small and midsize businesses, a sector with generally fast digitization of routine financial workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists human accountants by pre-populating entries, flagging anomalies, and automating routine posting, allowing them to focus on exception handling and analysis. This augmentative capability is well-established and widely deployed in modern accounting software. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-enabled accounting software significantly speeds up data entry, categorization, and reconciliation, letting bookkeepers focus on review and exceptions rather than manual entry. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | This task involves structured data entry and formula-based calculations on spreadsheets and databases—exactly the kind of rule-driven, high-volume repetitive work that current AI and RPA tools can automate end-to-end with significant time savings. Modern accounting software already integrates with AI-powered entry and reconciliation tools that meet the ≥50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 4/5 | Entering transactions, categorizing debits/credits, and totaling accounts is highly structured and rule-based, and modern accounting software (QuickBooks, Xero) already automates most of this via bank feeds, rules, and AI categorization, though edge cases and reconciliation still need review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While audit trail and compliance requirements create some oversight friction, there are no hard legal mandates requiring a human to sign off on the mechanical act of debiting and crediting. Most jurisdictions allow fully automated posting as long as controls and reconciliation are in place. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing is required to enter and total transactions (unlike auditing sign-off), though firms may want human review for error-proofing and internal controls, creating mild friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | The inference and integration cost of accounting automation is a fraction of a clerk's loaded wage, and the per-task cost is orders of magnitude lower once deployed. A single automated workflow can handle volumes that would require multiple FTE. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated bookkeeping software subscriptions cost a small fraction of a clerk's hourly wage for high-volume routine transaction entry, though oversight and exception handling retain some human cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature products (e.g., BlackLine, Intacct, Workiva, RPA platforms like UiPath) demonstrably perform automated debits, credits, and account totaling in production at scale across thousands of organizations. These systems reliably execute this task with minimal error rates in real financial operations. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed products like QuickBooks Online, Xero, and Bill.com already auto-categorize transactions and populate ledgers reliably for standard cases, though exceptions and complex entries still require human correction. |
Maintain inventory records.
85CI 75–95 · exposure 87 · augmentation 88 · importance 3.6/5 · click for rater detail
Maintain inventory records.
85| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Inventory automation adoption is deep and fast across retail, manufacturing, healthcare, and e-commerce. Most medium and large organizations have already deployed automated inventory systems; even small businesses increasingly use cloud-based tools. This is among the most widely automated accounting functions. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Inventory and accounting software with automation features is broadly adopted across retail, manufacturing, and wholesale sectors, though full end-to-end automation still varies by firm size. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assistants enhance human inventory clerks by flagging anomalies, predicting stockouts, suggesting reorder points, and automating routine reconciliation—allowing clerks to focus on exception handling and analysis. The human remains productive but with substantial AI-driven support. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI-enabled inventory systems significantly boost clerk productivity by automating data entry, flagging discrepancies, and forecasting needs, while humans remain involved in verification and exception resolution. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Inventory record maintenance is highly structured, rule-based data entry and tracking that modern accounting software and AI systems handle end-to-end today. Current ERP systems, inventory management tools, and AI agents can automatically log stock movements, reconcile records, and generate reports with minimal human intervention, easily achieving 50%+ time savings. |
| Task automatability | claude-sonnet-5 | 4/5 | Inventory tracking is largely structured data entry and reconciliation, which modern ERP/inventory systems with AI-assisted automation (barcode/RFID scanning, automated reconciliation, anomaly flagging) can handle with substantial time savings, though exception handling and physical counts still need human input. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No licensing requirement mandates a human maintain inventory records; the task is fully delegable to automated systems. Minor friction exists around data quality oversight and system configuration, but these are manageable organizational choices rather than legal barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for maintaining inventory records; main friction is internal controls, audit trail requirements, and integration with existing accounting systems rather than legal barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated inventory systems cost pennies per transaction and scale across unlimited records, while a clerk's loaded wage for manual entry runs $25–45/hour. The per-unit cost of AI-driven inventory maintenance is orders of magnitude cheaper than human labor. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Software subscription and automation costs are far lower than dedicating clerk hours to manual record-keeping, though initial integration and periodic human oversight add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed inventory management systems (NetSuite, SAP, QuickBooks, Zoho Inventory) and warehouse management software reliably maintain inventory records at scale across thousands of organizations. These are mature, production-grade systems with proven error handling and integration with point-of-sale and purchasing systems. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Mature inventory management software (SAP, NetSuite, QuickBooks, Fishbowl) with automated tracking, reorder alerts, and reconciliation is deployed at scale in production across industries today. |
Operate computers programmed with accounting software to record, store, and analyze information.
84CI 75–92 · exposure 87 · augmentation 88 · importance 4.8/5 · click for rater detail
Operate computers programmed with accounting software to record, store, and analyze information.
84| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Accounting and finance sectors are among the fastest adopters of AI and automation, with widespread migration to cloud platforms and automated workflows already underway in production across small and large firms alike. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Accounting/finance is a fast-adopting sector for AI tools, with widespread use of automated categorization and reconciliation features in mainstream small-business and enterprise software. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered accounting software significantly augments clerk productivity by automating data entry, flagging anomalies, suggesting categorizations, and generating routine reports, allowing humans to focus on exception handling and analysis. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI substantially speeds up data entry, categorization, and anomaly flagging in accounting workflows while clerks remain in the loop for review, exceptions, and judgment calls. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Modern accounting software and AI-powered tools can fully automate transaction recording, categorization, and basic analysis. End-to-end automation of data entry, ledger posting, and standard report generation easily achieves >50% time savings at equal or superior quality compared to manual entry. |
| Task automatability | claude-sonnet-5 | 4/5 | Core operations of entering, categorizing, and reconciling transactions in accounting software are largely automatable today via AI-enabled bookkeeping tools and OCR/ML categorization, though edge cases and judgment calls still need human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal regulatory barriers exist for automating data entry and recording in accounting software; no license requirement to operate the tools. Modest friction comes from audit trail oversight and the need for human sign-off on financial statements, but core automation faces few obstacles. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for basic bookkeeping entry, though some oversight expectations exist for accuracy in financial reporting and audit trails, creating mild organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Cloud-based accounting automation and RPA solutions cost significantly less per transaction processed than loaded clerk wages, often by an order of magnitude when amortized over high-volume processing. Integration costs are modest relative to the wage replacement. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated software subscriptions plus minimal oversight cost far less per transaction processed than a human clerk's loaded wage for routine data entry and reconciliation. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature, widely deployed products (QuickBooks, Xero, SAP with automation plugins, RPA platforms) reliably perform transaction recording and analysis in production across countless organizations at scale. Cloud-based accounting platforms routinely automate these tasks with high accuracy. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Products like QuickBooks Online, Xero, and various AI bookkeeping add-ons (e.g., automated bank feed categorization, receipt scanning) are deployed at scale in production, though they still require periodic human correction and oversight. |
Classify, record, and summarize numerical and financial data to compile and keep financial records, using journals and ledgers or computers.
84CI 75–92 · exposure 87 · augmentation 88 · importance 4.2/5 · click for rater detail
Classify, record, and summarize numerical and financial data to compile and keep financial records, using journals and ledgers or computers.
84| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Adoption is rapid and widespread in the information and professional services sectors; SMBs increasingly use cloud accounting software, and larger firms deploy RPA and ML-based accounts-payable/accounts-receivable automation. Displacement of routine clerk work is measurable and accelerating. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Accounting and financial services are among the faster-digitizing sectors, with widespread production use of automated transaction categorization and reconciliation tools already embedded in SMB and enterprise workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assistants significantly boost clerk productivity by automating routine classification, flagging exceptions, and suggesting corrections, allowing humans to focus on reconciliation and complex entries. This remains one of the highest-impact augmentation use cases in the back office. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI substantially speeds up data entry, categorization suggestions, and summarization while a human bookkeeper still reviews and finalizes records, making this a strong augmentation case. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | This task—classifying, recording, and summarizing financial data into journals and ledgers—is highly codified and repetitive, matching exactly what modern accounting software and AI systems do end-to-end today. Current systems (QuickBooks, Xero, automation tools with OCR and ML) achieve well over 50% time savings at equal or better quality for routine transaction classification and ledger posting. |
| Task automatability | claude-sonnet-5 | 4/5 | Modern accounting software with AI-based categorization (e.g., QuickBooks, Xero) can automatically classify transactions and post to ledgers, meeting the 50% time-saving bar for routine entries, though exceptions still need review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Barriers are low: no license is legally required to automate bookkeeping itself (though auditors/CPAs may be required for sign-off). Organizational inertia and internal audit policy can require human review, but do not prevent substitution of the core classification and recording work. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for basic bookkeeping entry, though some oversight/reconciliation and audit trail requirements create minor friction before fully autonomous deployment. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI/software inference and integration for transaction processing costs pennies per transaction, while a clerk's loaded wage (salary + overhead) is typically $20–35/hour; the cost ratio heavily favors automation by an order of magnitude for routine volume work. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated bookkeeping software with AI classification costs a small monthly subscription versus an hourly clerk wage, yielding roughly an order-of-magnitude cost advantage for high transaction volumes. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature, production-deployed products reliably perform transaction classification, automated journal entry, and ledger reconciliation at scale across thousands of organizations. Bank feeds, OCR-to-ledger pipelines, and rule-based + ML classification are standard in modern accounting platforms. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed products (bank feed auto-categorization, OCR receipt capture, ML-based coding rules) are widely used in production accounting systems, though accuracy on edge cases still requires human correction. |
Prepare purchase orders and expense reports.
84CI 75–92 · exposure 87 · augmentation 75 · importance 4.1/5 · click for rater detail
Prepare purchase orders and expense reports.
84| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Finance and accounting functions are digitization leaders with rapid adoption of automation platforms; large firms have already deployed AI-driven expense and procurement systems, and mid-market adoption is accelerating as tools become more accessible and user-friendly. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Finance and accounting functions are among the faster-adopting back-office areas, with expense management and procurement automation tools already widely deployed across firms of many sizes. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assistants meaningfully enhance human productivity by auto-filling templates, pre-categorizing expenses from receipt images, and flagging anomalies, allowing clerks to focus on review and exception handling rather than data entry. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly speed up the drafting, categorization, and error-checking of purchase orders and expense reports, letting clerks focus on exceptions and approvals rather than manual entry. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Preparing purchase orders and expense reports involves structured data entry, rule-based validation, and document generation—all tasks current AI systems handle end-to-end with 50%+ time savings. Systems can extract information from receipts, categorize expenses, auto-populate forms, and flag compliance issues without human intervention. |
| Task automatability | claude-sonnet-5 | 4/5 | Preparing purchase orders and expense reports is largely templated, rule-based data entry that current AI systems combined with existing accounting software (OCR, receipt scanning, expensify-style automation) can complete with substantial time savings at comparable accuracy. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal legal barriers exist—no licensed professional signature is required to generate these documents. Some organizational friction remains (internal audit sign-off, change management), but companies have broad freedom to automate these clerical tasks. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human create these documents; the main friction is internal approval policies and integration with existing ERP/accounting systems rather than regulatory or liability barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI-powered document automation and data entry costs (per-transaction inference, cloud processing, oversight) are orders of magnitude cheaper than the $20–35/hour loaded cost of a clerk manually entering and validating these routine documents. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated expense/PO software costs a small fraction of clerk hourly wages per transaction once integrated, though setup and occasional oversight add some cost, keeping it just short of order-of-magnitude savings in all contexts. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature products (Workday, NetSuite, SAP, and specialized accounting automation tools) demonstrably perform purchase order and expense report generation in production at scale across organizations, with APIs and plug-ins widely deployed in enterprise accounting workflows. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Mature deployed products (Expensify, SAP Concur, QuickBooks automation, Bill.com) already automate receipt capture, expense categorization, and PO generation in production at scale, though occasional exceptions still require human review. |
Check figures, postings, and documents for correct entry, mathematical accuracy, and proper codes.
83CI 74–92 · exposure 87 · augmentation 100 · importance 4.7/5 · click for rater detail
Check figures, postings, and documents for correct entry, mathematical accuracy, and proper codes.
83| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Financial services, accounting firms, and large enterprises with digitized ledgers have rapidly adopted AI audit tools and automated reconciliation. Adoption is pronounced in digitized, regulated sectors, though smaller firms lag behind. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Finance and accounting functions have been early and deep adopters of automation and AI-driven reconciliation tools across many company sizes. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI dramatically amplifies a clerk's productivity by flagging anomalies, pre-checking thousands of entries, and highlighting suspicious patterns before manual review. The human remains in the loop but processes work far faster and with higher accuracy. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools flag discrepancies, anomalies, and miscoded entries for human bookkeepers to review, substantially speeding up verification work while keeping a human in the loop for judgment calls. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Current AI systems (OCR, document parsing, formula verification, anomaly detection) can check figures, verify mathematical accuracy, and validate coding against rules end-to-end with >50% time savings. This is structured, rule-based work with clear correctness criteria—precisely what modern automation handles well. |
| Task automatability | claude-sonnet-5 | 4/5 | Verifying entries, math accuracy, and coding is a rule-based, pattern-matching task well suited to AI/automation with OCR and validation logic, though edge cases and ambiguous codes still need human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | While the checking itself can be automated, regulatory frameworks (SOX, GAAP, audit standards) and organizational policy often require human review and sign-off on high-stakes entries. AI typically works as a pre-check tool rather than sole decision-maker, creating material oversight friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this specific verification task, though some organizational oversight and internal control/audit trail requirements create mild friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI inference for document scanning, rule-checking, and flagging discrepancies costs pennies per transaction, while a clerk's fully-loaded wage is $25–40/hour. The cost ratio favors automation by one to two orders of magnitude. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated validation and reconciliation software processes thousands of transactions at a fraction of the cost of manual clerical checking. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature products exist in production: automated accounting compliance platforms, invoice/document verification systems, and AI-powered audit tools deployed by major accounting firms and enterprises reliably perform figure checking and code validation at scale today. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Accounting software (QuickBooks, Xero, SAP) and AI-based audit tools already perform automated reconciliation, anomaly detection, and coding checks in production at scale, though not fully error-free for complex or ambiguous entries. |
Access computerized financial information to answer general questions as well as those related to specific accounts.
82CI 72–92 · exposure 87 · augmentation 100 · importance 4.2/5 · click for rater detail
Access computerized financial information to answer general questions as well as those related to specific accounts.
82| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Financial and accounting sectors are among the fastest adopters of automation technology; major accounting firms and companies with digitized systems are actively deploying AI for data retrieval and query handling. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Finance/accounting software vendors are rapidly embedding AI assistants, but many smaller firms and clerks still rely on manual lookup, making adoption moderate and uneven. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | Even where humans remain in the loop for final verification, AI dramatically accelerates the information retrieval process, allowing clerks to answer questions and resolve account inquiries far faster than manual database navigation. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI assistants substantially speed up how clerks find and communicate financial information, serving as a strong augmentation tool even when clerks remain responsible for final answers. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | AI systems can reliably query databases, retrieve financial records, and answer questions about accounts and balances with high accuracy and significant time savings—this is a core strength of LLM-based systems combined with database access and search tools. |
| Task automatability | claude-sonnet-5 | 4/5 | Retrieving and summarizing account data to answer questions is well within current AI capability via database/API integration and LLM-based query interfaces, meeting the time-saving threshold for most routine queries. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While some organizations require human verification of financial queries for internal controls and auditing purposes, there are no legal licensing barriers preventing AI from accessing and reporting on computerized financial data directly. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement to answer account questions, but data security, access control, and accuracy concerns around financial records create moderate organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI inference cost for querying and responding to account queries is orders of magnitude cheaper than the hourly wage of a bookkeeping clerk, especially when integrated into existing systems. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Once integrated with accounting systems, AI query tools cost a fraction of a clerk's hourly wage per query handled, though initial integration adds some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed products (financial data APIs, AI assistants integrated with accounting software like QuickBooks, Xero, and SAP) already perform this at scale in production environments with high reliability and minimal error rates. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed products (ERP chatbots, finance copilots like those in QuickBooks, SAP, Oracle) already answer account queries in production, though edge cases and complex reconciliations still require human clarification. |
Calculate costs of materials, overhead, and other expenses, based on estimates, quotations and price lists.
82CI 72–92 · exposure 87 · augmentation 88 · importance 4.1/5 · click for rater detail
Calculate costs of materials, overhead, and other expenses, based on estimates, quotations and price lists.
82| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Accounting and finance sectors are among the fastest adopters of automation; automated cost allocation and calculation modules are now commonplace in mid-to-large organizations and increasingly standard in SMB accounting software suites, reflecting rapid, deep production adoption. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Accounting/bookkeeping is a digitized, software-heavy field with growing AI tool adoption, but many smaller firms still rely on manual clerk processes for this specific task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assists significantly by auto-populating calculations from structured data, flagging cost anomalies, and generating cost reports, allowing human clerks to focus on exception review, judgment calls on allocation methodology, and compliance verification rather than rote calculation. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools substantially speed up cost calculations by auto-extracting data from quotes/price lists and performing computations, letting clerks focus on verification and exceptions. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | This task is highly structured data entry and arithmetic involving cost calculations from standardized inputs (estimates, quotations, price lists). Current AI systems can reliably extract data, perform calculations, and generate cost summaries at significantly greater speed than manual processes, meeting the ≥50% time-saving threshold at equal quality. |
| Task automatability | claude-sonnet-5 | 4/5 | This is a structured, rules-based calculation task using estimates, quotations, and price lists—well suited to spreadsheet automation and AI tools that can parse documents and compute costs, though sourcing/interpreting varied quotation formats may need setup. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal legal or regulatory barriers prevent automation; organizations often retain human oversight for audit trails and exception handling, but these are operational preferences rather than hard legal requirements. Some firms may require internal sign-off, creating modest friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this specific calculation task, but some organizational caution around cost accuracy for financial reporting creates moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI-driven cost calculation via APIs or integrated accounting platforms costs pennies per transaction, while a human clerk performing the same calculation (sourcing data, cross-referencing, computing) costs $15–30+ in loaded wages; the cost differential is easily an order of magnitude in AI's favor. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated calculation and data extraction from price lists/quotations is far cheaper than manual clerical computation once integrated, though initial setup and periodic oversight add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature accounting software with AI cost-calculation modules, ERP systems with integrated pricing engines, and intelligent document processing tools are in production use across thousands of organizations today, reliably performing cost estimation and overhead allocation with minimal error. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | ERP/accounting software and AI-enhanced tools (e.g., automated cost estimation modules, OCR+LLM pipelines for quotes) already perform this reliably in many organizations, though edge cases with unusual pricing structures still need human review. |
Prepare and process payroll information.
81CI 75–86 · exposure 80 · augmentation 75 · importance 4.7/5 · click for rater detail
Prepare and process payroll information.
81| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Payroll automation has been widespread for 15+ years; nearly all mid-to-large organizations and increasingly small businesses use automated payroll systems rather than manual processing, representing one of the earliest and deepest AI/software adoptions in accounting. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Payroll automation is one of the most mature and widely adopted back-office software categories across nearly all firm sizes and sectors, with high penetration of automated systems already in production. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | Even in hybrid setups where humans oversee payroll, AI systems assist by pre-validating data, flagging anomalies, auto-populating fields from HR systems, and generating compliance reports, significantly raising human productivity in review and exception-handling roles. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven payroll tools significantly reduce manual entry, flag anomalies, and auto-calculate taxes/deductions, letting clerks focus on exceptions and compliance review rather than rote calculation. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Payroll processing is highly structured, rule-based work involving data entry, calculation, and tax deduction application. Modern payroll software and AI systems can handle 70–90% of routine payroll end-to-end (data ingestion, gross-to-net calculations, tax withholding, direct deposit setup) with minimal manual intervention, meeting the ≥50% time-saving bar for standard cases. |
| Task automatability | claude-sonnet-5 | 4/5 | Payroll processing is highly structured, rules-based, and already handled largely by software (ADP, Gusto, QuickBooks Payroll) with AI-enhanced automation for time capture, tax calculation, and deduction application; most steps meet the 50% time-saving bar though exception handling still needs review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While payroll involves tax/regulatory compliance and audit trails, no law requires a human to execute payroll processing; organizations retain responsibility but can delegate execution to software. Minimal licensing barriers exist for the task itself, though audit oversight remains required. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement to process payroll, but errors carry real liability (tax penalties, wage law violations) creating moderate caution and typically requiring human sign-off or audit trail review. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Cloud-based payroll processing costs $2–10 per employee per month, far below the fully-loaded wage of a payroll clerk ($25–35/hour); even accounting for integration and oversight, AI payroll solutions are 10–50× cheaper per transaction than human manual processing. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated payroll software processes thousands of employee records for a fraction of the cost of manual clerk hours, though ongoing subscription fees and oversight staff prevent a full order-of-magnitude gap in all cases. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature, deployed payroll automation products (Gusto, ADP, Paychex, Rippling) reliably process millions of paychecks monthly in production at scale, with error rates well below human levels for routine transactions and full compliance with tax and regulatory requirements. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Mature payroll platforms with automated calculation, tax filing, and direct deposit are deployed at massive scale today, though clerks still verify inputs and handle exceptions, edge cases, and compliance nuances. |
Monitor status of loans and accounts to ensure that payments are up to date.
81CI 75–86 · exposure 80 · augmentation 75 · importance 4.4/5 · click for rater detail
Monitor status of loans and accounts to ensure that payments are up to date.
81| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Financial services, accounting firms, and large enterprises have broadly adopted automated payment monitoring and status alerts over the past decade. This is a core feature in widespread use across information and finance sectors; displacement in routine clerk roles is already measurable. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Finance and accounting functions have been early and consistent adopters of automated transaction and payment monitoring tools, reflecting the sector's high digitization. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-driven dashboards and alerts meaningfully assist human clerks by surfacing exceptions and aggregate trends, freeing them for investigation and customer communication. The human remains in the loop for judgment calls, but productivity gains from automation-powered visibility are substantial. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Dashboards and automated flags significantly boost a clerk's ability to track many accounts simultaneously while the clerk retains oversight for exceptions and follow-up decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can reliably query account databases, flag overdue payments, and generate status reports with minimal setup. While human judgment on edge cases (hardship requests, pending arrangements) remains valuable, 50%+ time savings on core monitoring and alert generation is readily achievable today with standard accounting software and rule-based automation. |
| Task automatability | claude-sonnet-5 | 4/5 | Monitoring payment status against expected schedules is a structured, rules-based data task well-suited to automated reconciliation systems and alerts, though some edge-case judgment on delinquency handling remains. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No license requirement exists for automating payment monitoring itself. Light organizational friction exists (system setup, employee resistance to reduction in manual work) and regulatory oversight is generic (fraud prevention, data protection) rather than task-specific. Nothing legally mandates a human perform this monitoring. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this monitoring task itself, though some organizations retain human review for compliance or customer relations reasons before escalating delinquent accounts. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated monitoring via deployed accounting software costs pennies per account per month; human clerk labor loaded wage is $30–50/hour. The cost ratio heavily favors automation by an order of magnitude for routine status checks and payment flagging. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated monitoring software costs a small fraction of a clerk's time-equivalent wage once implemented, though initial integration and occasional exception review add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed accounting platforms (QuickBooks, SAP, NetSuite) and loan servicing systems routinely perform automated status monitoring and payment tracking in production at scale across thousands of organizations. This is a mature, standard feature of enterprise financial software. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Accounting software (QuickBooks, NetSuite, ERP modules) and loan servicing platforms already automate aging reports, payment tracking, and overdue alerts in production at scale today. |
Reconcile or note and report discrepancies found in records.
78CI 75–81 · exposure 75 · augmentation 88 · importance 4.3/5 · click for rater detail
Reconcile or note and report discrepancies found in records.
78| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Finance, accounting, and professional services sectors have rapidly and deeply adopted reconciliation automation; it is now standard practice in mid-to-large enterprises and increasingly common in smaller firms using cloud accounting platforms. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Accounting and finance functions have been early and consistent adopters of automation tools for reconciliation, with widespread deployment of software-driven matching in production environments. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI reconciliation tools greatly enhance human productivity by pre-filtering and automatically matching records, allowing clerks to focus on investigating discrepancies and exceptions rather than manual data entry and routine comparison; the human remains in the loop for judgment calls. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI-powered tools significantly speed up discrepancy detection and reporting, allowing clerks to focus on investigating and resolving flagged exceptions rather than manual matching. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI can automatically identify, flag, and document numerical discrepancies in structured financial records with high accuracy, meeting the ≥50% time-saving threshold; however, human judgment is often needed to resolve root causes and decide on corrective actions, preventing full end-to-end automation. |
| Task automatability | claude-sonnet-5 | 4/5 | Reconciliation of records against source documents is highly structured and rule-based, and current AI/automation tools can match transactions, flag discrepancies, and generate exception reports with substantial time savings, though edge cases still need human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While internal controls and audit sign-off may require human review, there are no strict licensing barriers or legal requirements that a human must personally perform the reconciliation itself; organizational friction is low in digitized accounting functions. |
| Adoption barriers | claude-sonnet-5 | 2/5 | There's no licensing requirement to perform reconciliation itself, though final sign-off on financial statements may involve accountant review; general bookkeeping tasks face minimal regulatory barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Reconciliation automation (once configured) runs at near-zero marginal cost per transaction compared to a clerk's hourly wage; even accounting for integration and oversight, the cost ratio is typically one to two orders of magnitude in AI's favor. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated reconciliation software operates at a fraction of the cost of a human clerk performing line-by-line matching, especially at scale, though initial setup and periodic oversight add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature accounting software and AI-powered reconciliation tools (e.g., invoice-matching, bank reconciliation engines) are deployed in production by thousands of organizations and reliably detect discrepancies; minor edge cases and complex manual adjustments remain, but the core function is proven and widely used. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Mature accounting software (e.g., QuickBooks, Xero, BlackLine) and RPA tools already perform automated matching and discrepancy flagging in production at many organizations, though some manual investigation of flagged items remains common. |
Compile statistical, financial, accounting, or auditing reports and tables pertaining to such matters as cash receipts, expenditures, accounts payable and receivable, and profits and losses.
78CI 75–81 · exposure 75 · augmentation 88 · importance 4.0/5 · click for rater detail
Compile statistical, financial, accounting, or auditing reports and tables pertaining to such matters as cash receipts, expenditures, accounts payable and receivable, and profits and losses.
78| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Accounting and finance sectors have rapidly and deeply adopted automated reporting tools; most mid-to-large organizations now generate routine financial reports via software rather than manual clerk work, with continued migration to AI-assisted analytics platforms. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Finance and accounting functions across most sectors have rapidly adopted automated bookkeeping and reporting software, with AI-enhanced tools now common in production rather than pilot stage. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments clerks by automating data aggregation and template population, allowing humans to focus on exception handling, reconciliation, and quality assurance rather than repetitive table construction and figure entry. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI substantially speeds up data compilation, categorization, and report generation while clerks retain review and judgment roles, making this a strong augmentation case regardless of full automation. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can compile financial reports and tables from structured data with high accuracy, automatically pulling figures from ledgers, categorizing transactions, and generating standard report formats. While some judgment about non-standard exceptions may remain, the core 50% time-saving threshold is readily met for routine statistical and financial compilations. |
| Task automatability | claude-sonnet-5 | 4/5 | Compiling standardized financial reports and tables from structured data is largely automatable today using accounting software, ERP systems, and AI-driven reconciliation/reporting tools, though some judgment calls on anomalies or categorization still require human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal or regulatory barriers prevent automation of report compilation itself; auditors and accountants typically review outputs rather than mandate manual creation. Organizational friction and data integration challenges exist but are not hard barriers, and no licensing requirement mandates human task performance. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates a human specifically for compiling these reports, though internal controls, audit trail requirements, and organizational preference for verified accuracy create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI-driven report automation costs are negligible per report once systems are configured, whereas a clerk's loaded hourly wage for manual compilation is substantial. The cost differential easily exceeds an order of magnitude for high-volume standardized reporting. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated reporting tools cost a small fraction of clerk labor once integrated, though initial setup, data cleaning, and periodic human oversight add nontrivial ongoing cost, keeping it just below the top tier. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature accounting software and business intelligence platforms (QuickBooks, SAP, Tableau, Power BI) routinely automate report generation from transactional data in production environments. These systems reliably produce cash flow, expense, accounts payable/receivable, and P&L reports at scale, though configuration and data validation still require human oversight. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Mature products (QuickBooks, Xero, SAP, NetSuite with AI features) reliably generate these reports in production at scale for standard cases, though edge cases and non-standard data still cause errors requiring human correction. |
Complete and submit tax forms and returns, workers' compensation forms, pension contribution forms, and other government documents.
72CI 70–74 · exposure 75 · augmentation 88 · importance 4.2/5 · click for rater detail
Complete and submit tax forms and returns, workers' compensation forms, pension contribution forms, and other government documents.
72| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Accounting and tax preparation sectors are information-intensive and highly digitized; major firms and mid-market practices have deployed automated form-filing solutions widely, though smaller firms lag behind. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Payroll, tax, and accounting software adoption is already widespread and mature in finance and professional services, with automated filing tools in common production use. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assists clerks by pre-populating forms, flagging missing or inconsistent data, and suggesting corrections before submission, substantially raising accuracy and speed while humans retain control over final approval. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI-driven accounting tools substantially speed up form preparation, data population, and error-checking, letting clerks review and submit rather than manually complete each form. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can extract, validate, and populate most standard tax and government forms with high accuracy using OCR and structured data from accounting systems; however, final verification and submission often require human judgment on edge cases, making full end-to-end autonomy without human review fall slightly short of the 50% time-saving bar at equal quality. |
| Task automatability | claude-sonnet-5 | 4/5 | Tax and government form preparation is highly structured, rules-based data entry that current AI-integrated accounting software (e.g., automated tax prep, e-filing tools) can largely complete, though final submission and edge-case handling often still involve human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Forms must typically be reviewed and digitally signed by an authorized human or CPA before submission; regulatory and liability frameworks require human accountability, creating meaningful friction despite technical feasibility. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Some government filings require certified/authorized signatures or licensed preparers for certain forms, and errors carry regulatory penalties, creating moderate compliance friction even though most steps are automatable. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI form-completion systems cost pennies to dollars per form after setup, orders of magnitude below the loaded hourly wage of a bookkeeping clerk, especially for high-volume submission scenarios. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated filing software costs a small fraction of clerical labor hours per form processed, though licensing, integration, and occasional human review add some cost relative to a fully unattended process. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed accounting software (e.g., TurboTax, ADP, Workiva) and AI-powered document automation platforms perform form population and validation reliably in production; minor integration friction and occasional form-specification updates prevent a perfect 5. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Products like TurboTax, QuickBooks, ADP, and payroll/tax filing platforms already automate large portions of form completion and e-filing in production at scale, though complex or non-standard cases still require human correction. |
Perform general office duties, such as filing, answering telephones, and handling routine correspondence.
71CI 67–75 · exposure 70 · augmentation 63 · importance 4.2/5 · click for rater detail
Perform general office duties, such as filing, answering telephones, and handling routine correspondence.
71| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Office automation and email/document management systems are nearly ubiquitous in information and professional services sectors; large organizations have already shifted much of this work to systems or outsourced processes. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Office administrative functions are adopting AI tools like chatbots and email automation at a moderate pace, with pilots common but full deployment still uneven across small and mid-size offices. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by auto-drafting routine responses, suggesting file categories, and flagging priority emails, meaningfully raising clerk productivity on correspondence handling even when humans retain final authority. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially speeds up drafting routine correspondence and organizing/searching files, letting clerks handle higher volumes with less manual effort while still overseeing accuracy. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Email and correspondence handling, filing organization, and phone call routing can be substantially automated via AI agents and tools; modern systems can handle 60–80% of routine incoming/outgoing correspondence, document classification, and appointment scheduling with minimal human oversight. |
| Task automatability | claude-sonnet-5 | 4/5 | Filing, phone triage, and routine correspondence are highly structured, repetitive tasks well within reach of current AI tools like document management systems, AI phone/voice agents, and email drafting assistants, though some physical filing and live call handling still require humans. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | These are routine, low-liability tasks with no licensing requirement or mandatory human contact; the main friction is organizational inertia and preference to retain staff for unpredictable work, not legal or regulatory barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human for filing or answering phones, though some organizational and customer-preference friction exists for phone interactions. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated email triage, filing systems, and phone routing cost pennies per interaction versus $25–40/hour loaded wage for clerk labor; the cost advantage is substantial once setup is amortized. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Cloud-based document management, AI email drafting, and virtual receptionist services cost a small fraction of clerical wages per unit of routine correspondence handled. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (email filters, chatbots, document management systems, voicemail-to-text) reliably perform subcomponents of these tasks in production; end-to-end office task automation remains partially integrated but core functions are mature and widely used. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like AI receptionists, chatbots, and email autoresponders are deployed in real offices, but reliability for open-ended phone calls and correspondence nuance still requires human backup in many settings. |
Compile budget data and documents, based on estimated revenues and expenses and previous budgets.
69CI 62–75 · exposure 70 · augmentation 88 · importance 4.3/5 · click for rater detail
Compile budget data and documents, based on estimated revenues and expenses and previous budgets.
69| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Accounting and finance sectors have rapidly adopted automation tools for data processing and budget compilation. Enterprise accounting departments, outsourced bookkeeping firms, and financial service providers routinely deploy these systems in production, with visible industry-wide momentum. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Finance and accounting functions are adopting AI tools steadily for reporting and forecasting, but full automation of budget compilation workflows remains in pilot/rollout phases at many firms. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-assisted budget compilation tools directly augment clerk productivity by automating data gathering and formatting, freeing staff to focus on validation, variance analysis, and interpretation. The human remains in the loop for quality assurance and judgment while AI handles routine aggregation. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools strongly augment this task by auto-populating spreadsheets, flagging anomalies versus prior budgets, and generating draft projections for clerks to review and adjust. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI systems can extract, organize, and compile budget data from documents and spreadsheets with high accuracy. Modern tools can ingest historical budgets, expense categories, and revenue projections, then aggregate and format them into standard structures, achieving substantial time savings with minimal human intervention. |
| Task automatability | claude-sonnet-5 | 4/5 | Compiling budget data from historical figures and estimates is a structured, spreadsheet-based task that AI tools with data connectors can largely automate, including pulling prior budgets and projecting line items. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some organizational friction exists around data validation, audit trails, and internal controls requiring human sign-off; however, no hard legal barrier prevents AI compilation. Many organizations maintain human review steps for compliance, but the compilation itself is not legally gated. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for compiling budget data, though organizations often want a human to validate figures before submission to management, creating light oversight friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automation of data compilation via software is significantly cheaper than manual data entry and organization by human clerks, especially at scale. The per-document cost of AI-driven extraction and consolidation is typically a small fraction of the loaded hourly wage for this clerical work. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated data compilation and formula-driven projections cost a small fraction of clerk hours once the pipeline is set up, though initial integration with company financial systems adds some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature accounting software and AI-powered data extraction tools (OCR, document parsing, spreadsheet automation) are deployed in production across accounting firms and enterprises. These systems reliably compile structured budget data from diverse sources with acceptable error rates for routine compilation tasks. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AI-powered finance/ERP add-ons (e.g., in NetSuite, SAP, or AI copilots in Excel) can compile and organize budget data today, but human review is still standard practice for accuracy and context before finalizing documents. |
Prepare bank deposits by compiling data from cashiers, verifying and balancing receipts, and sending cash, checks, or other forms of payment to banks.
64CI 50–79 · exposure 62 · augmentation 63 · importance 4.6/5 · click for rater detail
Prepare bank deposits by compiling data from cashiers, verifying and balancing receipts, and sending cash, checks, or other forms of payment to banks.
64| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Finance and accounting sectors show rapid, deep AI adoption; many mid-market and enterprise organizations already deploy RPA and accounting automation for exactly this workflow. Digitization is high and vendor solutions are mature. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Accounting and bookkeeping functions have moderate AI/software adoption with common use of automated reconciliation tools, though full deposit workflows including physical cash handling remain traditionally executed, placing this in middling adoption territory. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI assists by pre-populating deposit summaries and flagging discrepancies for human review, reducing manual verification time. However, the task does not require deep human judgment once automated, so augmentation is less transformative than replacement potential. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-powered accounting software significantly speeds up compiling cashier data, verifying receipts, and flagging discrepancies, meaningfully boosting clerk productivity even though a human still executes the physical deposit. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI and RPA systems can reliably extract data from cashier reports, verify receipt totals, balance entries, and generate deposit instructions with minimal human oversight. The task is highly structured, rule-based, and requires no subjective judgment, meeting or approaching the 50% time-saving threshold in most implementations. |
| Task automatability | claude-sonnet-5 | 3/5 | The data compilation, verification, and balancing portions can be automated via accounting software and RPA integration with POS/cashier systems, but physical handling and transport of cash/checks to banks remains a manual, physical step that AI cannot perform. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory barriers exist; the task is not legally restricted to human performers. Main friction is organizational (resistance to change, legacy system integration, internal audit sign-off requirements), but these are surmountable and common in mid-to-large firms already using modern platforms. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement exists for this task specifically, but banks require authorized human representatives for deposits, and there's inherent friction from physical cash handling, custody, and fraud-prevention controls that resist full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Once integrated, automated deposit preparation costs pennies per transaction compared to a clerk's hourly wage (typically $18–28/hour loaded). Volume discounts and cloud-based processing push the cost ratio well below 0.1x. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Software subscriptions and automated reconciliation tools are cheap relative to clerk wages for the data-matching portion, but the physical transport and deposit step still requires paid human labor, keeping overall cost comparable rather than dramatically lower. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature accounting software and RPA platforms (SAP, Workato, UiPath) routinely handle receipt compilation, reconciliation, and deposit preparation in production environments. Error rates are low for well-formatted input; the main limitation is integration with legacy systems or non-standard data formats. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Mature accounting software (QuickBooks, bank reconciliation tools, RPA bots) reliably automates the digital verification and balancing steps in production, but the physical deposit-making and cash handling still requires a human, limiting full-task feasibility. |
Receive, record, and bank cash, checks, and vouchers.
62CI 50–74 · exposure 62 · augmentation 63 · importance 4.5/5 · click for rater detail
Receive, record, and bank cash, checks, and vouchers.
62| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Financial and accounting services have aggressively adopted automated check processing, reconciliation, and deposit systems for decades; current AI refinements are accelerating an already-established trend. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Accounting/bookkeeping functions have seen steady but uneven automation adoption; digital recordkeeping is common while physical cash/check handling automation lags, giving a middling pace overall. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI assists humans by pre-populating records, flagging discrepancies, and suggesting corrections, but the human typically remains responsible for final verification and exception handling rather than being transformed by the tool. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-enabled accounting software significantly speeds up recording, categorization, and reconciliation of transactions, letting clerks focus on physical handling and exception management. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI-powered systems can automate receipt, recording, and banking of cash, checks, and vouchers through OCR, bank APIs, and automated ledger entry, achieving substantial time savings. However, physical cash handling and some verification steps still require human oversight, preventing a full 5. |
| Task automatability | claude-sonnet-5 | 3/5 | AI/automation software can handle recording and reconciling transactions and digital deposits, but physical receipt and banking of cash/checks still requires human or hardware (e.g., check scanners, physical cash handling) intervention. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some regulatory oversight and audit trail requirements apply, plus internal controls and authorization rules; however, no licensing requirement mandates a human perform this task, and automation is already common in practice. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Handling cash and negotiable instruments involves internal control requirements, segregation of duties, and audit trail rules that create moderate procedural and fraud-prevention barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated check scanning, OCR, and bank integration cost a fraction of a full-time clerk's loaded wage, easily achieving an order-of-magnitude savings per transaction batch. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Software subscriptions are cheap relative to clerk wages for the recording portion, but the physical handling and verification components still require paid human labor, keeping blended cost roughly comparable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature products (bank reconciliation software, accounting platforms with check imaging and automated posting) reliably perform these tasks in production at scale. Minor friction remains around edge cases and exceptions, but core functionality is widely deployed. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Accounting software (QuickBooks, Xero) and bank deposit automation (remote deposit capture) are widely deployed for recording and digital banking, but physical cash handling and check intake remain manual in most small/mid organizations. |
Comply with federal, state, and company policies, procedures, and regulations.
26CI 25–28 · exposure 25 · augmentation 63 · importance 4.6/5 · click for rater detail
Comply with federal, state, and company policies, procedures, and regulations.
26| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While finance and accounting are early AI adopters, compliance automation remains cautious and slow in practice because regulatory bodies and audit firms require clear human sign-off, and organizations fear liability more than they chase efficiency gains on compliance tasks. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Finance and accounting functions are adopting AI compliance and audit tools at a moderate pace, with pilots more common than full production deployment for this specific task.' |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by flagging policy changes, generating compliance checklists, and cross-referencing regulations against transaction data, improving a clerk's productivity and reducing missed items, but the human must still review and take responsibility for compliance decisions. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools like compliance checklists, automated flagging of rule violations, and regulatory update summarization meaningfully help clerks stay compliant while they remain accountable.' |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can retrieve and summarize policy documents, actual compliance requires interpreting context-specific regulations, making judgment calls on gray-zone scenarios, and integrating multiple overlapping rules—tasks that require human expertise and legal accountability today. Current systems cannot reliably handle novel compliance scenarios or take responsibility for regulatory violations. |
| Task automatability | claude-sonnet-5 | 2/5 | Compliance requires ongoing interpretation, judgment, and accountability across shifting regulations; AI can check rules but cannot independently 'comply' end-to-end with equal quality assurance.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Compliance with federal and state regulations carries legal and fiduciary responsibility; regulatory bodies and auditors typically require documented human accountability, and errors can result in substantial fines or penalties, creating strong barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Regulatory and organizational accountability structures typically require a human employee to be responsible for compliance, creating a meaningful liability barrier to full automation.' |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Compliance automation requires domain expertise, legal review, and continuous policy updates; the integration and oversight costs remain high relative to a clerk's wage, especially given liability exposure if automation misses a requirement. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Because human sign-off and accountability remain necessary, AI tools add to cost as an oversight layer rather than fully substituting, so savings are modest relative to wages.' |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs end-to-end compliance monitoring for accounting clerks across federal, state, and company rules. Tools exist for specific compliance checks (e.g., tax code lookups), but they require substantial human oversight and are narrow in scope; they do not independently ensure organizational compliance. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | There are compliance-checking and rules-engine tools, but no deployed product autonomously ensures full regulatory/company policy compliance for a clerk's role reliably in production.' |
Related occupations — Office & Administrative Support
How to read this
A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.
What would change this score
New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.