Office Clerks, General
In short: today's best AI models could do about 56% of this job's working time (projected from METR's April 2026 measurement and Epoch's November 2026 index) if the work were set up for them, and about the same by the end of 2028 if progress keeps its long-run pace. The conversations, in-person and hands-on work, about 44% of the time, stays with people.
Perform duties too varied and diverse to be classified in any specific office clerical occupation, requiring knowledge of office systems and procedures. Clerical duties may be assigned in accordance with the office procedures of individual establishments and may include a combination of answering telephones, bookkeeping, typing or word processing, office machine operation, and filing.
2,464,940 people do this job in the US. $45,010 a year.
Reviewed on September 30, 2026: what each task is made of, and what handing it over would ask.
A typical 40-hour week
- AI could take on21 h
- A person checks AI's work2 h
- Desk work AI can't do yet0 h
- Stays with people17 h
Office Clerks, General, now: of a 40-hour week, AI could take on 21 hours, 2 hours become checking its work, 0 hours are desk work it can't do yet, and 17 hours stay with people.
The largest pieces: Compute, record and proofread data (about 5 h); Maintain files and databases (about 3 h); Type and edit documents (about 2 h).
Estimates for a typical week of this job, if progress keeps its long-run pace, counting work the best models finish four times in five. Every task is below.
of the working time. Now is projected from METR's April 2026 measurement and Epoch's November 2026 index.
- 20%
- 36%
- 14%
- 0%
- 29%
of the work: Anthropic's measure of what people already bring to Claude at work (2025). It counts help with part of a task, so it can be higher than the share within reach.
On the chosen pace and reliability. Estimate.
Share of the working time within reach, now to 2030
Change assumptionsHide assumptionsLong-run pace, four in five
What each choice means: , , , , . Every pace starts from .
At this date and pace, models finish tasks up to 2.4 work days long four times in five.
What handing it over would take, now
The 56% within reach splits by what a model would need. The dates, pace and reliability follow the controls above. Estimates.
of the working time. The material can be handed to a model and the result handed back.
of the working time. The model has to be connected to the system the work lives in.
is conversation short enough by length. This site leaves conversation with people, so it is not in the total.
How far this has come
With GPT-4, measured in March 2023, 0% of this job's working time was within reach at . With Claude 3.7 Sonnet, measured in February 2025, 3%. With Claude Mythos Preview (early), measured in April 2026, 52%. Today's estimate: 56%.
A backcast: today's task data with the best model METR had measured by each date. It shows how fast the models have moved, not what anyone forecast at the time.
Measured use over time
Of the Claude.ai conversations Anthropic sampled and tied to a work task, 0.81% went to this job's tasks in May 2026, the 32nd largest of the 447 jobs that showed up at all; 5 of its 20 tasks showed up. Within each of Anthropic's methods, never across a change: 0.34% to 0.83% (released February and March 2025); 1.1% to 1.5% (August 2025 to February 2026); 0.77% to 0.81% (April to May 2026).
Each release
| Conversations | Share | Rank | Tasks seen |
|---|---|---|---|
| Released February 2025 | 0.34% | 61 of 708 | 9 of 20 |
| Released March 2025 | 0.83% | 24 of 703 | 9 of 20 |
| The week of Aug 4, 2025 (released September 2025) | 1.1% | 15 of 655 | 9 of 20 |
| The week of Nov 13, 2025 (released January 2026) | 1.4% | 17 of 672 | 10 of 20 |
| The week of Feb 5, 2026 (released March 2026) | 1.5% | 11 of 677 | 10 of 20 |
| April 2026 (released June 2026) | 0.77% | 35 of 434 | 4 of 20 |
| May 2026 (released June 2026) | 0.81% | 32 of 447 | 5 of 20 |
Measured, not estimated: Anthropic's samples of Claude.ai conversations, each tied by a model to an O*NET task, from the Anthropic Economic Index. It says where Claude's use goes, not how much of this job's working time it covers, so it is not on the same scale as the reach above. The releases differ in their sample, their classifier and the smallest share they report (the June 2026 file rounds each task to a hundredth of a percent, so a smaller task reads as none), which is why a change of method starts a new line. At least 96% of each older release's task-tied use matched this site's tasks, and all of the newest. Claude.ai only, so use through other tools and other companies' models is not in it. Every job ranked.
Robots and the physical work
Anthropic rates 46% of this job's working time as physical work. For 45% of the working time, a robot has been shown doing most of the task somewhere: 38% in a purpose-built cell, 7% in a structured facility. 0.52 of 3, the 303rd highest of the 923 jobs.
- Not physical 54%
- No robot shown 1%
- In a purpose-built cell 38%
- In a structured facility 7%
Anthropic's rating, not this site's estimate: Claude Opus 5, searching the web, rated each physical O*NET task by the kind of setting in which a robot it could cite (deployed, sold or demonstrated) does most of the task, weighted by its own estimate of each task's share of the time (What work can robots do?, September 30, 2026, data CC BY 4.0; every job ranked). A robot shown somewhere is not a robot in your building: cost, setup and safety decide that. The share within reach on this page still counts no body work. The robot behind each rated task is under the task below.
What changes for the people in this job
From the task-by-task review, by the end of 2028 on the long-run pace at four in five. It says which parts of the work stay with people and which change, not whether anyone keeps a job: that depends on decisions no data set can see.
What stays with people
44% of the working time: the conversation, in-person and hands-on parts of the tasks.
- Deliver messages and run errands, the hands-on work, 6%
- Handle mail and answer correspondence, the hands-on work, 5%
- Answer questions and take orders, the conversation, 5%
- Operate office machines, the hands-on work, 4%
- Compile, copy and file records, the hands-on work, 4%
- Count, weigh and organize materials, the hands-on work, 3%
What becomes checking the machine
5% of the working time: desk work a model could do by then, in tasks where a mistake costs something or a licensed person signs. A person checks it. Another 51% is desk work that is small if wrong, which can be handed over with lighter checks.
- Handle money, bookkeeping and banking, costly if wrong, 4%
- Complete and mail bills and contracts, costly if wrong, 2%
What to learn
- Running AI inside the business systems. About 36% of the time is work inside business systems, which a model can do only once it is connected: knowing the systems well enough to set that up and watch it.
- The hands-on craft. About 29% of the time is hands-on work, which no model here does; robots are rated separately on this page.
- Briefing a model and checking its drafts. About 20% of the time is documents and data, the part models reach first: writing a clear brief, giving the right material, and checking what comes back.
- Reviewing work you did not do yourself. The tasks under What becomes checking the machine, where a mistake costs something, stay a person's to check: spotting what a draft gets wrong is a skill of its own.
- Dealing with people outside the organization. Customers, patients, suppliers or the public: trust with them is built by a person.
Your own version of this job: read it against your job description. The headline is free; the full personal report, every task with its timeline and what to learn, is $19 (a sample).
The tasks
Sorted by . 20 tasks from . 3 have a worked example.
| Task | ||||
|---|---|---|---|---|
| Compute, record, and proofread data and other information, such as records or reports.Calculating and checking figures, and entering them.Done daily by most who do it. See it done: one graded attempt (passes with named fixes) | 12% | 38 min13 min to 1.7 h | Documents 50%, systems 50% | 100%Half by now |
| Maintain and update filing, inventory, mailing, and database systems, either manually or using a computer.Updating records in office systems, and some paper filing.Done daily by most who do it. Robots: in a purpose-built cell, Ripcord robotic document digitization system (sold). Anthropic's rating. | 8% | 25 min8 min to 1.7 h | Systems 80%, hands-on 20% | 80%Desk part half by now |
| Compile, copy, sort, and file records of office activities, business transactions, and other activities.Part entry in a system, part paper handled by hand.Done daily by most who do it. AI observed on this task.Robots: in a purpose-built cell, UiPath Document Understanding (sold). Anthropic's rating. | 8% | 25 min8 min to 1.3 h | Systems 50%, hands-on 50% | 50%Desk part half by now |
| Open, sort, and route incoming mail, answer correspondence, and prepare outgoing mail.Mostly handling paper mail, with some letter writing.Done daily by most who do it. Robots: in a purpose-built cell, OPEX Eagle remittance processor (sold). Anthropic's rating. | 7% | 25 min8 min to 50 min | Hands-on 70%, documents 30% | 30%Desk part half by now |
| Communicate with customers, employees, and other individuals to answer questions, disseminate or explain information, take orders, and address complaints.Conversation with customers and staff, with the entry that follows.Done daily by most who do it. AI observed on this task. | 6% | 13 min4 min to 38 min | Conversation 70%, systems 30% | 30%Desk part half by now |
| Operate office machines, such as photocopiers and scanners, facsimile machines, voice mail systems, and personal computers.Working copiers and scanners by hand, and some time at a computer.Done hourly or more by most who do it. Robots: in a purpose-built cell, Ripcord robotic digitization system (sold). Anthropic's rating. | 6% | 8 min2 min to 25 min | Hands-on 70%, systems 30% | 30%Desk part half by now |
| Collect, count, and disburse money, do basic bookkeeping, and complete banking transactions.Entries in the books, and cash counted and taken to the bank.Done daily by most who do it.Robots: in a structured facility, CityBase self-service payment kiosks (deployed). Anthropic's rating. | 6% | 33 min13 min to 1.3 h | Systems 60%, hands-on 40% | 60%Desk part half by now |
| Deliver messages and run errands.Going places on foot.Done daily by most who do it. Robots: in a structured facility, Moxi (Diligent Robotics) (deployed). Anthropic's rating. | 6% | 21 min8 min to 50 min | Hands-on 100% | Stays with people |
| Type, format, proofread, and edit correspondence and other documents, from notes or dictating machines, using computers or typewriters.Pure document work.Done daily by most who do it. See it done: one graded attempt (passes) | 6% | 25 min8 min to 1.3 h | Documents 100% | 100%Half by now |
| Review files, records, and other documents to obtain information to respond to requests.Finding the record in the files and reading out what is needed.Done daily by most who do it. AI observed on this task.See it done: one graded attempt (passes with named fixes) | 5% | 17 min4 min to 50 min | Documents 50%, systems 50% | 100%Half by now |
| Complete work schedules, manage calendars, and arrange appointments.Calendar entries, with calls to confirm.Done daily by most who do it. AI observed on this task. | 5% | 25 min8 min to 1.3 h | Systems 80%, conversation 20% | 80%Desk part half by now |
| Monitor and direct the work of lower-level clerks.Day to day conversation with a desk team, and checking their work.Done daily by most who do it. | 3% | 38 min13 min to 1.7 h | Conversation 80%, documents 20% | 20%Desk part half by now |
Within reach, per task: the share of the task's time whose instances are short enough for a model to finish at the chosen date, pace and reliability. Half by: the date half of that time comes within reach. Made of: how the task's time splits among five kinds of doing. Only the documents and systems parts can count as within reach, so a task that is 30% documents can be at most 30% within reach.
Jobs that share skills with this one
From O*NET's related occupations. Within reach now and by the end of 2028, long-run pace, four in five.
- Secretaries and Administrative Assistants, Except Legal, Medical, and Executive59% now, 68% by 2028
- File Clerks50% now, 53% by 2028
- Receptionists and Information Clerks60% now, 60% by 2028
- Administrative Services Managers52% now, 70% by 2028
- Medical Secretaries and Administrative Assistants84% now, 84% by 2028
- Correspondence Clerks85% now, 85% by 2028
Where this job works
The industry groups that employ the most of it (BLS, May 2025).
Where the most of these jobs are
The metro areas with the most people in it (BLS, May 2025).
Questions about this job
- How much of the work of Office Clerks, General is within reach of AI now?
- An estimated 56% of the working time is within reach of AI models now and 56% by the end of 2028, on METR's long-run pace at four in five. Within reach means a model could finish work that long at that reliability. It is not a forecast of jobs lost or of savings.
- Which tasks of Office Clerks, General come within reach first?
- Weighted by the time each takes, the tasks with the most within reach now (projected from METR's April 2026 measurement and Epoch's November 2026 index) are: Compute, record and proofread data (100% of the task's time now); Maintain files and databases (80% of the task's time now); and Type and edit documents (100% of the task's time now).
- Which tasks of Office Clerks, General stay with people?
- Tasks that are people work or body work are not counted as within reach at any date. The largest: Deliver messages and run errands; and Count, weigh and organize materials. Together such tasks are about 9% of the working time, and the people and body parts of the other tasks stay with people too.
- What should Office Clerks, General learn as AI takes on more of the work?
- By the job's own task-by-task review, the first things to learn are running AI inside the business systems; the hands-on craft; and briefing a model and checking its drafts. By the end of 2028 on the long-run pace at four in five, about 44% of the working time stays with people, 5% becomes checking a model's work and 51% could be handed over with lighter checks. It says which parts of the work change, never whether anyone keeps a job.
- Do people already use AI for the work of Office Clerks, General?
- In the Claude.ai conversations Anthropic sampled for May 2026 and tied to a work task, 0.81% went to this job's tasks, the 32nd largest share of the 447 jobs that showed up; 5 of its 20 tasks appeared. Anthropic's labor market measure puts the share of the job's tasks seen in real Claude use at work at 45%. These are measures of use, not of how much of the work AI could finish, and they cover Claude only.
- Can robots do the physical work of Office Clerks, General?
- Anthropic's robot exposure rating (September 2026) finds 46% of the working time is physical work, and that a robot has been shown doing most of the task, somewhere, for 45% of the working time, 7% in a structured facility. Its index is 0.52 of 3. One example: CityBase self-service payment kiosks for "handle money, bookkeeping and banking". A robot shown somewhere is not one in every workplace; this site's own estimates still count no physical work as within reach.
- How is this estimated?
- From the 20 tasks O*NET lists for the occupation, each with an estimated length checked against GDPval and a share of the working time, reviewed one task at a time on September 30, 2026 for what each is made of (documents, business systems, conversation, in person, hands-on), and METR's measurements of how long a task the best AI models can finish, carried forward to the date from the latest, Claude Mythos Preview (early) in April 2026. Data checked November 10, 2026. The Work Bench has 3 worked examples of its tasks: a brief, a rubric written beforehand, one attempt by an AI model and its grade.
This job in your state or metro area
Embed these figures on your site
A small card with the share of working time within reach now and at the end of 2027, 2028 and 2030, the date and a link back. Free under CC BY 4.0; the card carries the credit.
Your own version of this job
The page above is the occupation. Your job is your title and what you actually do: give both, and they are read against these tasks in your browser, nothing sent anywhere. The headline is free: how much of your working time is within reach now and by the end of 2028, and how many of the tasks your description names.
The full personal report is $19, to print or save as a PDF: every task with when it comes within reach, which ones your description names, the parts of your description the task list does not cover, what stays with you, what becomes checking the machine and what to learn.