Will AI replace Bookkeeping, Accounting, and Auditing Clerks?

Not as a job, but it is already doing parts of the work. Across the 28 official task statements scored for Bookkeeping, Accounting, and Auditing Clerks (United States, SOC 43-3031), 71% of the importance-weighted core work is made of tasks today's AI could already do most of. The overall exposure score is 75 out of 100 (range 70–81, band: high). That is a statement about tasks, not about headcount: this measures what AI could do, not whether any employer adopts it, whether the law allows it, or whether doing the routine parts faster creates more demand for the human parts. Figures are from release 2026-q4.1.

Which tasks in “Bookkeeping, Accounting, and Auditing Clerks” can AI already do?

The highest-scoring tasks in release 2026-q4.1 are: “Operate computers programmed with accounting software to record, store, and analyze information” (93/100, very high); “Debit, credit, and total accounts on computer spreadsheets and databases, using specialized accounting software” (93/100, very high); “Classify, record, and summarize numerical and financial data to compile and keep financial records, using journals and ledgers or computers” (93/100, very high). Each score comes from five published 0–4 ratings turned into a number by a published formula, and each carries the model's one-sentence reason on the page.

Which tasks in “Bookkeeping, Accounting, and Auditing Clerks” stay human?

About 13% of this job's task weight sits in work that scores low for AI exposure. The lowest-scoring tasks in release 2026-q4.1 are: “Receive, record, and bank cash, checks, and vouchers” (14/100, minimal); “Prepare bank deposits by compiling data from cashiers, verifying and balancing receipts, and sending cash, checks, or other forms of payment to banks” (38/100, low); “Operate 10-key calculators, typewriters, and copy machines to perform calculations and produce documents” (38/100, low). Low scores usually mean the task needs a body in a room, a legally accountable human, or trust built in real time. Those are the three things the scoring rubric treats as gates rather than obstacles.

What should someone working in “Bookkeeping, Accounting, and Auditing Clerks” do about AI?

Start from the ledger rather than the headline: 71% of this job's weighted core work is exposed, and roughly 13% is not. The practical move is to spend more of your week on the tasks that score low, the ones above, and to get fluent at directing AI through the tasks that score high, because those are the parts that change whether or not you are ready for them. This page does not predict your job, and nothing here is career advice tailored to you: the score describes the occupation, not the person.

How is the AI exposure score for Bookkeeping, Accounting, and Auditing Clerks calculated?

Each official task statement for the occupation is rated on five published 0–4 dimensions (output replicability, physical embodiment, licensed accountability, real-time human trust, and data availability) by claude-opus-5 using scoring prompt task_scoring_v1.0. The model never writes the score; a published formula turns the five ratings into a 0–100 number, so every score can be recomputed by hand. The occupation figure is the importance-weighted mean across 28 scored tasks. The prompt, the rubric, the formula and the full dataset are published at https://futureproof.collab365.com/method and https://futureproof.collab365.com/data/2026-q4.1 under CC BY 4.0.

US dataswitch to UK

Bookkeeping, Accounting, and Auditing Clerks

operating computers programmed with accounting software, reconciling records of bank transactions and compiling budget data and documents. If that's your week, this page is about your job.

The honest answer

Most tasks in this job are the kind AI has learned to do: operating computers programmed with accounting software. The tasks, though, are not you.

Your move: three real directions from here ↓

Where the work sits, by task weight

shifting to AI: 71%
changing shape: 15%
staying human: 13%

These bars are tasks changing hands, not people being counted out. The ledger below shows which.

Whole-job exposure score 75 out of 100 (70–81 allowing for uncertainty): high exposure, across 28 scored tasks. The number is the support for the sentence above it, not a headline about anyone’s future.

How we know this

What is measured: Every published task statement for bookkeeping, accounting, and auditing clerks is rated on five dimensions: can a model produce the output, does the work need a body in a room, does it need a legally accountable person, does it depend on a person being trusted in the moment, and how much data exists. A published formula turns those five ratings into the score; the model never writes the number.

How the bar is built: Each task’s share of the bar is its published importance weight, so a task you do all day counts for more than one you do twice a year.

Release: 2026-q4.1, scores computed 2026-08-04. Read the full method.

It would be a lie to soften that; receiving, recording and banking cash, checks and vouchers is what this work rebuilds around. The plan below starts there.

Your week, as this page understands it

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. The job title says “bookkeeping”, “accounting” or “auditing clerks”: officially one job, several names. The real job is the part underneath: receiving, recording and banking cash, checks and vouchers. That is the thing someone has to be right about.

The exposed part of this job is specific, and we won’t pretend it is coming back. But bookkeeping, accounting, and auditing clerks is not one task. It is 28 scored ones, and they are nowhere near equally exposed. The most durable of them, on this evidence, is receiving, recording and banking cash, checks and vouchers, and the ledger below shows exactly why.

Where the work sits, by task weight

shifting to AI: 71%
changing shape: 15%
staying human: 13%

These bars are tasks changing hands, not people being counted out. The ledger below shows which.

How we know this

How the bar is built: Each task’s share of the bar is its published importance weight, so a task you do all day counts for more than one you do twice a year.

Release: 2026-q4.1, scores computed 2026-08-04. Read the full method.

Your job, task by task

These are the official task statements for this occupation, in plain English, sorted by what the evidence says is happening to each one. The official wording sits under every line so you can check the rewrite against it.

One row is marked low confidence, so treat it as a ballpark rather than a fine measurement.
1 of this occupation's scored task statements carry a score that was measured under a different occupation's context, because the statement is shared between jobs and has only been scored once so far. Each one names the occupation it was measured under in the free routes below; none is presented as a measurement for this job.
Task clusters are not derived in this release, so the task-cluster field is empty and no Collab365 Space routing is attached to this occupation yet.

Shifting to AI

21 tasks

Tasks today’s tools can already do most of. This is the part we will not soften: where these rows are the bulk of your week, the week changes.

Operating computers programmed with accounting software
This is reading one thing and writing another: computers in, a record out. That is the shape today's tools are built for.
importance 5 · Core
Source: “Operate computers programmed with accounting software to record, store, and analyze information.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Recording, storing and analysing information in accounting software is precisely the work that software increasingly does itself.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Checking figures, postings and documents for correct entry, mathematical accuracy and proper codes
This is reading one thing and writing another: figures, postings and documents in, a record out. That is the shape today's tools are built for.
importance 5 · Core
Source: “Check figures, postings, and documents for correct entry, mathematical accuracy, and proper codes.” (O*NET task statement)
How this row was scored
Exposure score: 88 out of 100 (84–92 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Checking figures and codes against source documents is exactly the kind of accuracy work software already does well.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 4/4.
Reconciling or noting and reporting discrepancies found in records
This is reading one thing and writing another: discrepancies found in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Reconcile or note and report discrepancies found in records.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Reconciling records and flagging discrepancies is comparison work software does at least as accurately as a person.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Debiting, crediting and total accounts on computer spreadsheets and databases, using specialized accounting software
This is reading one thing and writing another: total accounts in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Debit, credit, and total accounts on computer spreadsheets and databases, using specialized accounting software.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Posting debits and credits and totalling accounts is arithmetic and data entry that software handles automatically.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.

Changing shape

4 tasks

Tasks where the machine takes the producing and a person keeps the checking, the deciding, or the answering-for-it. For most jobs this is the biggest group, and it is where "transformation, not termination" is literally visible.

Performing general office duties, such as filing, answering telephones and handling routine correspondence
The software now makes the first pass at general office duties, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 4 · Core
Source: “Perform general office duties, such as filing, answering telephones, and handling routine correspondence.” (O*NET task statement)
How this row was scored
Exposure score: 48 out of 100 (41–55 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Filing, phones and routine letters are mostly desk tasks software can handle, though paper filing needs hands.
The five ratings: output a model can produce 3/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.
Complying with federal, state and company policies, procedures and regulations
The software now makes the first pass at federal, state and company policies, procedures and regulations, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 5 · Core
Source: “Comply with federal, state, and company policies, procedures, and regulations.” (O*NET task statement)
How this row was scored
Exposure score: 50 out of 100 (38–62 allowing for uncertainty): partial exposure, low confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Following company and government rules is a responsibility carried by a person, even where software flags the requirements.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Comparing computer printouts to manually maintained journals to determine if they match
The software now makes the first pass at computer printouts, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 4 · Supplemental
Source: “Compare computer printouts to manually maintained journals to determine if they match.” (O*NET task statement)
How this row was scored
Exposure score: 56 out of 100 (49–63 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Comparing printouts with handwritten journals is checking work software can do once the pages are scanned.
The five ratings: output a model can produce 3/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.

Staying human

3 tasks

Tasks that stay with a person, because they happen in the physical world, because the rules need someone accountable, or because the value is that a specific person does them.

Operating 10-key calculators, typewriters and copy machines to perform calculations and produce documents
This work happens in the physical world: 10-key calculators, typewriters and copy machines, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Operate 10-key calculators, typewriters, and copy machines to perform calculations and produce documents.” (O*NET task statement)
How this row was scored
Exposure score: 38 out of 100 (31–45 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Calculations are trivial for software, but feeding a copier and using office machines needs hands.
The five ratings: output a model can produce 3/4 · needs a body in a room 2/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Receiving, recording and banking cash, checks and vouchers
This work happens in the physical world: cash, checks and vouchers, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Receive, record, and bank cash, checks, and vouchers.” (O*NET task statement)
How this row was scored
Exposure score: 14 out of 100 (10–18 allowing for uncertainty): minimal exposure, high confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Receiving and banking cash, cheques and vouchers involves physically handling money as well as recording it.
The five ratings: output a model can produce 2/4 · needs a body in a room 3/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Preparing bank deposits by compiling data from cashiers
This work happens in the physical world: bank deposits, in a real place. Software cannot follow it there.
importance 5 · Supplemental
Source: “Prepare bank deposits by compiling data from cashiers, verifying and balancing receipts, and sending cash, checks, or other forms of payment to banks.” (O*NET task statement)
How this row was scored
Exposure score: 38 out of 100 (31–45 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Compiling and balancing receipts is desk work, but cash and cheques have to be handled and taken to the bank.
The five ratings: output a model can produce 3/4 · needs a body in a room 2/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.

Show the other 18 tasks

Classifying, recording
shifting to AI
This is reading one thing and writing another: numerical in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Classify, record, and summarize numerical and financial data to compile and keep financial records, using journals and ledgers or computers.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Classifying, recording and summarising financial data into ledgers is structured work accounting software already does.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Calculating, preparing and issuing bills, invoices, account statements and other financial statements according to established procedures
shifting to AI
This is reading one thing and writing another: bills, invoices, account statements and other financial statements in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “Calculate, prepare, and issue bills, invoices, account statements, and other financial statements according to established procedures.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Producing bills, invoices and statements from set procedures is standard automated output in billing systems.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Compiling statistical, financial, accounting or auditing reports and tables pertaining to such matters as cash receipts, expenditures
shifting to AI
This is reading one thing and writing another: this work in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “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.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Compiling financial and statistical reports and tables from existing records is routine work software produces on demand.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Coding documents according to company procedures
shifting to AI
This is reading one thing and writing another: documents in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Code documents according to company procedures.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Coding documents to company rules is repetitive classification that software does quickly and consistently.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Performing financial calculations, such as amounts due, interest charges, balances, discounts, equity and principal
shifting to AI
This is reading one thing and writing another: financial calculations in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Perform financial calculations, such as amounts due, interest charges, balances, discounts, equity, and principal.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Working out amounts due, interest, discounts and balances is standard calculation software performs reliably.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Reconciling records of bank transactions
shifting to AI
This is reading one thing and writing another: records of bank transactions in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “Reconcile records of bank transactions.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Bank reconciliation is a matching exercise that accounting software already performs accurately and automatically.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Monitoring status of loans and accounts to ensure that payments are up to date
shifting to AI
This is reading one thing and writing another: status of loans in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “Monitor status of loans and accounts to ensure that payments are up to date.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Tracking whether loan and account payments are up to date is monitoring that systems do continuously.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Transferring details from separate journals to general ledgers or data processing sheets
shifting to AI
This is reading one thing and writing another: details in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “Transfer details from separate journals to general ledgers or data processing sheets.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Copying details from journals into the general ledger is data transfer that software does automatically.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Compiling budget data and documents
shifting to AI
This is reading one thing and writing another: budget data in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “Compile budget data and documents, based on estimated revenues and expenses and previous budgets.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (86–100 allowing for uncertainty): very high exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Pulling budget figures together from past budgets and estimates is compilation work software does quickly.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Matching order forms with invoices
shifting to AI
This is reading one thing and writing another: order forms in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Match order forms with invoices, and record the necessary information.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Matching orders to invoices and recording the details is standard automated processing in purchase systems.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Maintaining inventory records
shifting to AI
This is reading one thing and writing another: inventory records in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “Maintain inventory records.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (86–100 allowing for uncertainty): very high exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Keeping inventory records up to date from receipts and issues is structured record-keeping software handles.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Preparing purchase orders and expense reports
shifting to AI
This is reading one thing and writing another: purchase orders in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “Prepare purchase orders and expense reports.” (O*NET task statement)
How this row was scored
Exposure score: 93 out of 100 (89–97 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Producing purchase orders and expense reports from existing data is routine automated output.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Preparing trial balances of books
shifting to AI
This is reading one thing and writing another: trial balances of books in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “Prepare trial balances of books.” (O*NET task statement)
How this row was scored
Exposure score: 81 out of 100 (77–85 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Any ledger system produces a trial balance at the press of a button.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Preparing and processing payroll information
shifting to AI
This is reading one thing and writing another: payroll information in, a record out. That is the shape today's tools are built for.
importance 5 · Supplemental
Source: “Prepare and process payroll information.” (O*NET task statement)
How this row was scored
Exposure score: 81 out of 100 (77–85 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Payroll runs almost entirely on software, though the figures still need a responsible person's approval before payment.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Accessing computerized financial information to answer general questions as well as those related to specific accounts
shifting to AI
This is reading one thing and writing another: computerized financial information in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Access computerized financial information to answer general questions as well as those related to specific accounts.” (O*NET task statement)
How this row was scored
Exposure score: 79 out of 100 (72–86 allowing for uncertainty): high exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Looking up account information to answer questions is a search task software does instantly, though the asker is a person.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.
Calculating costs of materials, overhead and other expenses, based on estimates, quotations and price lists
shifting to AI
This is reading one thing and writing another: costs of materials, overhead and other expenses in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “Calculate costs of materials, overhead, and other expenses, based on estimates, quotations and price lists.” (O*NET task statement)
How this row was scored
Exposure score: 75 out of 100 (68–82 allowing for uncertainty): high exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Costing materials and overheads from price lists and quotes is calculation work software does, though some judgement is involved.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Calculating and preparing checks
shifting to AI
This is reading one thing and writing another: checks in, a record out. That is the shape today's tools are built for.
importance 4 · Supplemental
Source: “Calculate and prepare checks for utilities, taxes, and other payments.” (O*NET task statement)
How this row was scored
Exposure score: 61 out of 100 (54–68 allowing for uncertainty): high exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Working out and preparing payments is automated in most systems, though someone authorised has to approve them.
The five ratings: output a model can produce 4/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Completing and submitting tax forms and returns
changing shape
The software now makes the first pass at tax forms, but someone has to be answerable for the result, and it cannot be the software. So the job becomes checking and deciding rather than producing.
importance 4 · Supplemental
Source: “Complete and submit tax forms and returns, workers' compensation forms, pension contribution forms, and other government documents.” (O*NET task statement)
How this row was scored
Exposure score: 56 out of 100 (49–63 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; someone qualified has to answer for it.
The rating behind it: Tax and government forms can be filled from the books automatically, but someone qualified must stand behind what is submitted.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 2/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.

What this job pays, and how many people do it

Median pay: $50,670a year, the middle of the range, so half earn more and half earn less.bls-oews, 2025 · May 2025 estimates (national_M2025_dl.xlsx)
How we know this
Source: bls-oews
Reference period: May 2025 estimates (national_M2025_dl.xlsx)
Rounding: Shown as published.
People doing this job: 1,373,680in the US, 2025.bls-oews · May 2025 estimates (national_M2025_dl.xlsx)

What is deliberately not here: a forecast of how many of these jobs exist in ten years. Where an official projection exists for a market we publish it with its vintage; where it does not, we leave the space empty rather than borrow the other country’s number.

Why this is shifting

The reason is boringly specific. Most of what is shifting here is reading one thing and writing another: figures, postings and documents in, a record out. The rows above are exactly that shape: operating computers programmed with accounting software and checking figures, postings and documents for correct entry, mathematical accuracy and proper codes. What it cannot do is be there in the room, and that is still where cash, checks and vouchers get done. Which is why this page talks about your tasks changing, not your job ending.

Your move

Over a pint: what I’d tell you if you were my friend

The exposed part of your job is the biggest part, and I am not going to dress that up: operating computers programmed with accounting software is work today's tools do quickly and cheaply, and that is not coming back.

So, given all that: 71% of this job's task weight sits in rows the software is already learning, 15% in rows that change shape rather than disappear, and 13% in rows it is nowhere near. That is the position, measured across 28 scored tasks. It is not a forecast about you.

What you have that the software does not is receiving, recording and banking cash, checks and vouchers, plus the years of knowing when something looks wrong before you can say why. That is the raw material for everything below.

This week: one thing

Sit on the machine's side of the desk. Pick one real piece of figures, postings and documents you would normally do yourself, let whatever software you already have take the first pass at it, and then go through what it produced and write down every single thing it got wrong. One evening this week. Do not fix anything yet. Just catch it.

What you end up holding: a written list of the machine’s mistakes, in your handwriting
How long it takes: an evening, or an hour if you pick one job rather than one client

If you can’t run software on your employer’s or your clients’ data: Do the same hunt on paper. Take one printed piece of figures, postings and documents, work through it the way you always do, and mark every point where you made a call rather than followed a rule. Same evening, same list, nothing to log into and nobody to ask permission from. That list is the same artifact: it is the judgment written down.

Over the next 90 days

Change one sentence about what you do. Not on a CV. Out loud, to whoever pays you. From “my job is operating computers programmed with accounting software” to “I check what the software does and tell you what it means.” Your error list from this week is what makes that sentence true instead of a claim, so use it: show it once, to one person, inside the next ninety days. Same skills, priced as judgment rather than as typing.

Over the next 12 months

Walk toward the end of this job that answers for things, and get it recognised. Pick the one part of receiving, recording and banking cash, checks and vouchers you are already best at, and spend the year making it formal: a qualification, a named responsibility, a specialism people ask for by name. Price it honestly: that is evenings, it is months rather than weeks, and the seats are competitive because everyone in your position is looking at the same door. Before you pay for anything, use CareerOneStop - Find local training. It is free, it is the Labor Department's own service, and it is listed below with the rest of the free routes.

The roads out of here

Now the roads out, with the honest bill on each one. One route survived the checks. The rest are below with the reason they did not.

Bookkeeping, Accounting, and Auditing ClerksFinancial and Investment Analysts
a year or morematched on shared tasks
The reporting end of your job - compiling the figures, building the tables, answering what a number means - is the entry hall of financial analysis work.
You already do the compiling-and-explaining half of an analyst's job; the missing half is the degree and the modelling.
The work the two jobs share
- You already do
  “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.” (importance 74 in your job)
  They do
  “Present oral or written reports on general economic trends, individual corporations, and entire industries.”
- You already do
  “Access computerized financial information to answer general questions as well as those related to specific accounts.” (importance 80 in your job)
  They do
  “Inform investment decisions by analyzing financial information to forecast business, industry, or economic conditions.”
- You already do
  “Compile budget data and documents, based on estimated revenues and expenses and previous budgets.” (importance 83 in your job)
  They do
  “Employ financial models to develop solutions to financial problems or to assess the financial or capital impact of transactions.”
- You already do
  “Debit, credit, and total accounts on computer spreadsheets and databases, using specialized accounting software.” (importance 79 in your job)
  They do
  “Draw charts and graphs, using computer spreadsheets, to illustrate technical reports.”
What you would not already have: The analyst's client and markets work - 'Develop and maintain client relationships', 'Perform securities valuation or pricing' - is not work a bookkeeper already does. Coverage is moderate, not high.
The honest bill
Pay: $102,740 against your $50,670 (OEWS May 2025 (both)). A gap that size is the market telling you the barrier is real, which is why this sits in the year-or-more slot, not the near ones.
- The entry ticket: Typical entry-level education moves from 'Some college, no degree' to 'Bachelor's degree' (BLS EP, both). A degree gate, so the route is banded year-plus and the gate is named on the card.
- What the pay gap is telling you: The 2.03x pay ratio is above the 2.0 threshold: the market is saying the barrier is real. This route may only appear in the 12-month / year-plus slot, never as a near move.
- What you live on meanwhile: This is a study-while-earning route: evenings and weekends for a year or more, not a career break. Free careers advice before spending anything is in the free routes below.
- Is the target job itself holding up: Passes on growth: financial and investment analysts +5.7% 2024-2034 (BLS EP). The exposure leg re-runs when per-task exposure lands; a high band would demote this route.
How long: A year or more of part-time study alongside work, driven by the degree gate rather than by any single course.
See what AI is doing to financial and investment analysts
That page has a nearby route → Microsoft 365 Report Builders.

3 moves I checked and rejected

These are the obvious-looking jumps. They are here with their reasons rather than quietly dropped, because the ones that fail are worth knowing about. It is one less thing to turn over at night.

Paralegals and Legal Assistants
Why it looked obvious: Both jobs involve compliance-driven document work: 'Comply with federal, state, and company policies, procedures, and regulations' and 'Complete and submit tax forms and returns... and other government documents'.
Why I am not recommending it: Fails the coverage gate. 'Handles regulated paperwork' is a category, not shared detailed work activity - the paralegal's durable core is legal research, drafting and case preparation, none of which a bookkeeper does. The surviving 43-3031 route (financial analysis) has genuinely shared reporting work; this one does not.
Look at that job’s page anyway →
Tellers
Why it looked obvious: It came up as a near neighbour on the work AI is not taking: you already receive, record, and bank cash, checks, and vouchers, and their equivalent is to receive and count daily inventories of cash, drafts, and travelers' checks. Across both published task lists that is about 8% of the durable work in that job.
Why I am not recommending it: Almost none of it is work you already do: about 8% of the durable side of that job. That is a different job, not a next step. It is a pay cut, in those words: $43,030 against your $50,670, 15.1% less (OEWS May 2025 (both)). Retraining to earn less is a decision, not advice. And it is a narrow door: about 329,480 of those jobs against 1,373,680 of yours (OEWS May 2025), 24% as many seats.
Look at that job’s page anyway →
Cashiers
Why it looked obvious: It came up as a near neighbour on the work AI is not taking: you already receive, record, and bank cash, checks, and vouchers, and their equivalent is to receive payment by cash, check, credit cards, vouchers, or automatic debits. Across both published task lists that is about 6% of the durable work in that job.
Why I am not recommending it: Almost none of it is work you already do: about 6% of the durable side of that job. That is a different job, not a next step. It is a pay cut, in those words: $32,880 against your $50,670, 35.1% less (OEWS May 2025 (both)). Retraining to earn less is a decision, not advice.
Look at that job’s page anyway →

What I’d stop worrying about

A friend tells you what not to spend fear on. This is that list.

The headline number you read somewhere
The big “X% of jobs” figures are about the whole economy, not about you. The number that describes your job is on this page: 71% of its task weight, across 28 scored tasks. Every row behind it is printed above with the source. Check ours; ignore theirs.
“It’s too late for me to become something else”
You are not starting from zero, and the page shows why: receiving, recording and banking cash, checks and vouchers is work the software cannot do and you already do it. The move above is a repricing of what you know, not a new career. Nobody who has just left college has that.
“I should learn to code”
Almost certainly not. The value in your job is moving toward checking, deciding and answering for the output, not toward writing the software. A weekend of Python will not change your position; the error list from this week will.
The “obvious” next job everyone suggests
I checked the obvious moves and most of them did not survive. The reasons are printed with the routes above, including the pay and the gate. A move that fails on the numbers is worth knowing about so you can stop turning it over at night.

You are reading the United States figures

The United Kingdom splits this work across more than one official group, of which Financial and accounting technicians is the closest. The pay and employment figures are not directly comparable, and we do not average them together.

Switch to the United Kingdom page →partial match

The other groups this work is counted across:

In UK official statistics this job is counted as Financial and accounting technicians, Book-keepers, payroll managers and wages clerks and Financial accounts managers. Pay is shown separately for each of those groups (medians cannot be averaged together), while the task list and the scores on this page are for this group only.

Your route through this

Two honest options, and no deadline on either

Free, and complete

The moves above cost nothing. These are the real services that go with them: public, government-funded, and free at the point of use. Nothing on this page is behind an email address or a payment.

Anywhere in the US:
CareerOneStop - Find local training
Search what's running near you, from the Labor Department's own database, before anyone sells you a course.
Free to search; individual programs vary, and some are funded
Anywhere in the US:
American Job Center locator
An American Job Center will sit down with you for free. Find yours by ZIP code.
Free
- CareerOneStop local help hub
Anywhere in the US:
CareerOneStop - Licensed occupations finder
Check what your state actually requires before you pay for anything.
Free
Anywhere in the US:
CareerOneStop
Free

A nearby route

There's no Space built for bookkeepers yet.

Collab365 Spaces is built by the same people as this site. We find the problems that AI and automation are creating inside one kind of work, then solve them as short courses, briefings and Blueprints. Each Space is the community too, so the research and the people doing your job are in the same place.

What a Space actually is, in full

The closest match is Microsoft 365 Report Builders - a community for people who build business reports in Excel, Power Query and Power BI without a data team behind them. That overlaps with the part of your job that's growing: compiling the reports and checks, rather than keying in the entries. If that overlap isn't you, the free route below covers the same ground.

Try Microsoft 365 Report Builders free →

7 days free, no card needed. Explore up to 2 Spaces before you choose a plan: you pick a plan later, not now.

After the trial it is a paid community, and you get identical data either way. If the overlap above is not your job, the moves above cost nothing and stand on their own.

Nearby moves

The jobs above, as pages you can read the same way as this one. Your job shares its core work with these. That is what the match is, and it is all it is.

Financial and Investment Analysts →You already do the compiling-and-explaining half of an analyst's job; the missing half is the degree and the modelling.has a nearby route → Microsoft 365 Report Builders

Noted, and thank you. We’ll email you if a Space for bookkeeper launches. Nothing else.

That did not look like an email address, so nothing was saved. Have another go below.

We could not save that. The fault is ours, not yours, and your address was not stored. Please try again later.

No deadline on any of this. The page will still be here, and the data is refreshed on a published schedule rather than when someone wants a headline.

Questions people ask about this job

Will AI replace Bookkeeping, Accounting, and Auditing Clerks?: Not as a job, but it is already doing parts of the work. Across the 28 official task statements scored for Bookkeeping, Accounting, and Auditing Clerks (United States, SOC 43-3031), 71% of the importance-weighted core work is made of tasks today's AI could already do most of. The overall exposure score is 75 out of 100 (range 70–81, band: high). That is a statement about tasks, not about headcount: this measures what AI could do, not whether any employer adopts it, whether the law allows it, or whether doing the routine parts faster creates more demand for the human parts. Figures are from release 2026-q4.1.
Which tasks in “Bookkeeping, Accounting, and Auditing Clerks” can AI already do?: The highest-scoring tasks in release 2026-q4.1 are: “Operate computers programmed with accounting software to record, store, and analyze information” (93/100, very high); “Debit, credit, and total accounts on computer spreadsheets and databases, using specialized accounting software” (93/100, very high); “Classify, record, and summarize numerical and financial data to compile and keep financial records, using journals and ledgers or computers” (93/100, very high). Each score comes from five published 0–4 ratings turned into a number by a published formula, and each carries the model's one-sentence reason on the page.
Which tasks in “Bookkeeping, Accounting, and Auditing Clerks” stay human?: About 13% of this job's task weight sits in work that scores low for AI exposure. The lowest-scoring tasks in release 2026-q4.1 are: “Receive, record, and bank cash, checks, and vouchers” (14/100, minimal); “Prepare bank deposits by compiling data from cashiers, verifying and balancing receipts, and sending cash, checks, or other forms of payment to banks” (38/100, low); “Operate 10-key calculators, typewriters, and copy machines to perform calculations and produce documents” (38/100, low). Low scores usually mean the task needs a body in a room, a legally accountable human, or trust built in real time. Those are the three things the scoring rubric treats as gates rather than obstacles.
What should someone working in “Bookkeeping, Accounting, and Auditing Clerks” do about AI?: Start from the ledger rather than the headline: 71% of this job's weighted core work is exposed, and roughly 13% is not. The practical move is to spend more of your week on the tasks that score low, the ones above, and to get fluent at directing AI through the tasks that score high, because those are the parts that change whether or not you are ready for them. This page does not predict your job, and nothing here is career advice tailored to you: the score describes the occupation, not the person.
How is the AI exposure score for Bookkeeping, Accounting, and Auditing Clerks calculated?: Each official task statement for the occupation is rated on five published 0–4 dimensions (output replicability, physical embodiment, licensed accountability, real-time human trust, and data availability) by claude-opus-5 using scoring prompt task_scoring_v1.0. The model never writes the score; a published formula turns the five ratings into a 0–100 number, so every score can be recomputed by hand. The occupation figure is the importance-weighted mean across 28 scored tasks. The prompt, the rubric, the formula and the full dataset are published at https://futureproof.collab365.com/method and https://futureproof.collab365.com/data/2026-q4.1 under CC BY 4.0.

Where these numbers come from

About the data on this page

One row is marked low confidence, so treat it as a ballpark rather than a fine measurement.
The match between this job and its counterpart in the other country is partial, so the two pages count slightly different groups of people.
1 of this occupation's scored task statements carry a score that was measured under a different occupation's context, because the statement is shared between jobs and has only been scored once so far. Each one names the occupation it was measured under in the free routes below; none is presented as a measurement for this job.
Task clusters are not derived in this release, so the task-cluster field is empty and no Collab365 Space routing is attached to this occupation yet.

Task statements: onet-dbProcessing: catalogue-bridge → onet-im-rt-weighting → task-scoring → score-aggregation
Task weights: onet-db (im-rt)
Scores: Rubric task_scoring_v1.0, model claude-opus-5, computed 2026-08-05.
Pay and employment: bls-oews (May 2025 estimates (national_M2025_dl.xlsx))bls-oews (May 2025 estimates (national_M2025_dl.xlsx))

Figures on this page come from release 2026-q4.1, published 2026-08-05. Every release keeps its own permanent address, so a figure you cite in March is still there, unchanged, in November.

The plain-English wording on this page is assembled directly from the task statements and the published ratings, not written by hand for this occupation. That is why it is specific, and it is also why we say so.

The routes and free resources further up are today’s, not the release’s (last reviewed 2026-08-05). A route is an offer, not a historical fact, so it moves on its own clock.

How we score a job Download this release Look up another job

Using these figures?

Cite this

Everything on this site is published under CC BY 4.0. Quote it, chart it, sell something built on it. Just say where it came from, and cite the dated release rather than the site, so the figure you quote stays checkable.

Plain text

Collab365 (2026). Collab365 Futureproof: task-level AI exposure for US and UK occupations, release 2026-q4.1 (methodVersion 2.0.0, promptVersion task_scoring_v1.0). https://futureproof.collab365.com/data/2026-q4.1. Licensed CC BY 4.0. Built with O*NET data (USDOL/ETA, CC BY 4.0); ONS data (Open Government Licence v3.0); GAISI task framework (arXiv:2507.22748, MIT); BLS data (public domain).

BibTeX

@misc{collab365futureproof2026q41,
  title        = {Collab365 Futureproof: task-level AI exposure for US and UK occupations, release 2026-q4.1},
  author       = {{Collab365}},
  year         = {2026},
  url          = {https://futureproof.collab365.com/data/2026-q4.1},
  note         = {Release 2026-q4.1, methodVersion 2.0.0, promptVersion task_scoring_v1.0, CC BY 4.0}
}

Permalink

https://futureproof.collab365.com/data/2026-q4.1

Data as of release 2026-q4.1, published 2026-08-05. Releases never change after publication; when the figures move, a new dated release is published beside this one and this one stays exactly where it is.