noworkleft.com

Home › Jobs › Couriers and Messengers

Office and Administrative Support · SOC 43-5021

Couriers and Messengers

Pick up and deliver messages, documents, packages, and other items between offices or departments within an establishment or directly to other business concerns, traveling by foot, bicycle, motorcycle, automobile, or public conveyance.

Will AI replace this job?

The data cannot say, and anyone who gives you a percentage for “replaced” is guessing. What it does show is a gap between three things, for the 16 tasks O*NET lists for this occupation:

Could an LLM do it?

0%–25%

of tasks: an LLM alone could do the task or cut its time by half. Human raters say 0%, GPT-4 says 25%.

Counting tasks that need extra software around the model: 25%–38%.

Is it happening?

1 of 16 tasks

show measurable use in Claude conversations (Anthropic Economic Index). Anthropic’s occupation score, which counts only tasks that were also rated feasible, is 4%. Observed is a subset of “could” by construction.

Claude only, mostly US users, sampled periods in 2025. Not all AI use, and not employment.

What is left?

10 of 16 tasks

were rated “no LLM speed-up” by every rater (so, by Anthropic’s method, they cannot show up as observed use either). That is the part of the job these studies see no sign of AI touching.

It does not mean those tasks are safe. It means these methods found nothing there.

Task by task

Sorted with the tasks seen in use first. The three coloured squares are the three raters, in this order: human raters, GPT-4, GPT-4 with an alternative rubric. “Usage score” is Anthropic’s published task_penetration value; 0 means none seen.

LLM aloneneeds extra softwareno speed-up
Task (O*NET 31.0)Could an LLM?Claude usage score
Plan and follow the most efficient routes for delivering goods. all raters 0.92
Record information, such as items received and delivered and recipients' responses to messages. all raters none seen
Check with home offices after completed deliveries to confirm deliveries and collections and to receive instructions for other deliveries. (supplemental) all raters none seen
Use telephone to deliver verbal messages. (supplemental) all raters none seen
Receive messages or materials to be delivered, and information on recipients, such as names, addresses, telephone numbers, and delivery instructions, communicated via telephone, two-way radio, or in person. 2 of 3 raters none seen
Sort items to be delivered according to the delivery route. 1 of 3 raters none seen
Deliver and pick up medical records, lab specimens, and medications to and from hospitals and other medical facilities. no rater none seen
Walk, ride bicycles, drive vehicles, or use public conveyances to reach destinations to deliver messages or materials. no rater none seen
Load vehicles with listed goods, ensuring goods are loaded correctly and taking precautions with hazardous goods. no rater none seen
Deliver messages and items, such as newspapers, documents, and packages, between establishment departments and to other establishments and private homes. no rater none seen
Unload and sort items collected along delivery routes. no rater none seen
Perform routine maintenance on delivery vehicles, such as monitoring fluid levels and replenishing fuel. no rater none seen
Obtain signatures and payments, or arrange for recipients to make payments. (supplemental) no rater none seen
Unload goods from large trucks, and load them onto smaller delivery vehicles. (supplemental) no rater none seen
Open, sort, and distribute incoming mail. (supplemental) no rater none seen
Perform general office or clerical work, such as filing materials, operating duplicating machines, or running errands. (supplemental) no rater none seen
How to read this. These are averages over task lists written for the US labour market, rated by people and by GPT-4 using early-2023 model capabilities, next to a count of what one company’s chatbot users actually ask for. A task being “exposed” says nothing about whether employers will cut jobs, shift work, or hire more of you. See how to read an exposure score.

Try it with your own task list

Sources: task list from O*NET® Database (U.S. Department of Labor, CC BY 4.0; modified by noworkleft.com); “could” labels from Eloundou et al., GPTs are GPTs (OpenAI/University of Pennsylvania, MIT licence, labels from early-2023 models); “seen in use” from the Anthropic Economic Index (Claude usage only, CC BY). Dates, versions and limits.

Other jobs in Office and Administrative Support

Data built 2026-11-06. O*NET 31.0. Task match rates: 99.4% with the OpenAI/Penn labels, 99.2% with Anthropic usage.