Opens in a new tab
needsahuman.

Will AI replace postal service clerks?

Nah.

Most of the day is counter work with documents, parcels and people, which machines can speed up but not take over. This job scores 80 out of 100 on (higher is safer). Today people do 15% of the work with AI’s help, and 85% still needs a person.

Updated 3 October 2026 43-5051 4123 2026-Q4
Office and Administrative SupportPostal Service Clerks43-5051 · 2026-Q4
0% AI does it15% AI helps85% needs a human
Your job's name, lit by the work that still needs a human.Needs a human 85%AI helps 15%AI does it 0%

AI does it: AI can do the task largely by itself. AI helps: a person still does it, faster with AI. Needs a human: AI can do little of it yet.

Why the retail counter still needs a person

Will AI replace postal service clerks? The honest picture is task erosion, not an empty counter. Clerks weigh and rate parcels, sell stamps and money orders, register and insure mail, and sort out the problems that walk through the door. Much of that looks routine from the outside. Up close, nearly every transaction carries a small decision with rules attached.

Take two examples. Checking a customs declaration against what is actually in the box is part judgment, part regulation; get it wrong and the shipment stops at a border. Verifying identity documents for a passport acceptance appointment is the same kind of work: a person compares a face to a document, spots the thing that does not fit, and signs their name to it. Software can prefill the form. The accountability stays with the clerk.

The counter is also where exceptions land. Damaged parcels, insurance claims, held mail, a money order that never cleared. Those cases arrive upset and incomplete, and they are resolved by asking the right question rather than by looking up a record. That is the part of the job machines keep handing back.

Scale matters too. The Bureau of Labor Statistics counts about 73,720 postal service clerks in the United States, with median pay of $62,130, and projects a change of -0.3% in employment between 2025 and 2035 (BLS, 2025). That is slow drift, not a cliff. It fits a job where tools absorb pieces of the work while the headcount thins at the edges.

What machines run, what they assist with, and what people keep

Some of this job already runs on machines. Postage calculation, label printing, barcode scanning and tracking lookups happen without a clerk touching them, and self-service kiosks take a share of simple stamp and parcel transactions. In our task split, 0% of task time falls into the group where software or hardware can carry the task end to end. The Can AI do it? method page explains how that time is counted, and the overall coverage figure for this job is 14 out of 100.

A second group is assisted work. Looking up rates and service options, checking delivery status, and drafting the standard reply to a routine inquiry all go faster with a system doing the retrieval. The clerk still reads the customer and makes the call. That assisted slice is 15% of task time.

The rest stays with people: 85% of task time. It covers the physical handling of odd parcels, identity and document checks, claims and complaint resolution, cash and accountability for a drawer, and talking a confused or angry customer through what happens next. The robotics panel above puts most of the physical work in the mobile robot tier, which is the harder and slower end of automation to deploy behind a retail counter.

What the evidence actually shows

The evidence grade for this job is D. A D grade means there is no direct, published test of an AI system against a trained postal clerk doing this job’s tasks, so we publish no parity number for it. Treat any outside claim that puts a precise percentage on postal clerk automation as an estimate, not a measurement.

What would settle it is narrow and testable: a timed trial on a mixed counter queue, comparing a kiosk plus assistant against a clerk on error rates for customs declarations, identity verification, insurance and claims handling, and first-contact resolution of damaged-mail cases. Until something like that is published and reviewed, the honest answer is that the gap is unmeasured. The Is it better than a person? method page sets out the grades, and the full approach is on the methodology page.

When this could shift

Most likely after 2037 (8 in 10 of our scenarios). The replacement-year method page explains exactly what that window measures and how it is built.

Two things could pull it earlier. Wider kiosk and app self-service would strip more simple transactions out of the queue, and automated parcel handling pushed further toward the retail floor would cut the lifting and staging that keeps a person on site. Hiring freezes and attrition can do quietly what no machine does openly: fewer counters, longer lines, the same work spread thinner.

Two things hold it back. First, cost and physical plant: retrofitting thousands of retail counters is capital spending on a long cycle, and the cost panel above shows why a tool bill and a staffed window are not the same decision. Second, rules and liability. Identity acceptance, customs declarations and insured mail come with legal duties that a public service has to be able to answer for, and collective agreements shape how fast roles change.

Good to know: the biggest near-term effect in jobs like this is usually fewer new hires, not current clerks being walked out.

How to stay needed behind the window

Lean into the work the kiosk hands back. First, document and identity verification appointments, where care and a signature carry weight. Second, claims, damaged parcels and lost-item cases, which need questions, records and a decision. Third, business mailing advice: helping a small shipper pick the right service, pack it properly and fill the form correctly the first time.

Two skills travel well from here. One is regulated record handling, which means knowing the rules, documenting cleanly and being auditable. The other is plain explanation under pressure, because most counter complaints are resolved by someone saying clearly what went wrong and what happens next.

If you want to look sideways, the closest work is Postal Service Mail Carriers, Postal Service Mail Sorters, Processors, and Processing Machine Operators, and Mail Clerks and Mail Machine Operators, Except Postal Service. Each has its own task split and its own window. You can put two of them side by side with the job comparison tool.

For the wider picture, this role sits in the material recording and distributing job family and in our government sector pages. If the BLS projection is what concerns you, the list of jobs expected to shrink is the one to read next.

Frequently asked questions

Will AI replace postal workers at the counter?

Not as a single event. Self-service kiosks, apps and automated postage already take simple transactions out of the queue, and that trend continues. What stays is identity and document checks, claims, damaged parcels, cash accountability and difficult conversations. The task list above shows which parts sit with machines, which are assisted and which still need a person, so you can see where the pressure actually falls.

Are self-service kiosks replacing postal clerks?

Kiosks replace transactions, not roles. They handle stamps, simple postage and some parcel drop-offs well. They do not verify a passport applicant’s documents, judge whether a box will survive shipping, or settle an insurance claim. In practice kiosks shorten the routine part of the queue and leave clerks with the harder cases, which is task erosion rather than a full swap.

What is the job outlook for postal service clerks?

The Bureau of Labor Statistics counts about 73,720 postal service clerks in the United States, with median pay of $62,130, and projects employment to change by -0.3% between 2025 and 2035 (BLS, 2025). That is close to flat with a slight decline. Slow declines usually show up first as fewer openings for new entrants rather than as cuts to experienced staff.

What jobs will be gone by 2030 because of AI?

Very few whole jobs disappear on a fixed date. Work is a bundle of tasks, and AI takes tasks before it takes titles. The more useful question is which tasks in your job are already automated, which are assisted, and what is left. Our rankings page lets you check any occupation, and each job page shows its own task split and dated range.

What is the difference between a postal clerk and a mail carrier?

Clerks work mostly at the retail counter and in the back office: selling postage, rating parcels, registering and insuring mail, handling claims and verifying documents. Carriers work on a route, sorting for delivery and handling mail and packages outdoors. The task mixes are different, so the automation pressure is different too. Both have their own page on this site.

Each ridge is a slice of the job's task time.Needs a human 85%AI helps 15%AI does it 0%
The job’s mark

No two jobs leave the same print

Every job gets its own fingerprint, drawn from its code. The amber ridges are the share of task time that still needs a person. Below them, the same ridges are written out in ones and zeros: slate for the work AI helps with, white for the work AI can do.

Postal Service Clerks, O*NET-SOC 43-5051. 85% of the job’s task time still needs a human, so 85 of every 100 ridges are amber; slate is what AI helps with, white what AI can do.

What AI can and cannot do

The tasks that make up the job, from , and where AI stands on each today: , (a person does it, with AI speeding it up) or . 85% of the still needs a human.

Each block is one task; its height is its share of working time.Needs a human 85%AI helps 15%AI does it 0%
The job's task list: the parts AI can do are blacked out.Needs a human 85%AI helps 15%AI does it 0%
Weigh letters and parcels, compute mailing costs based on type, weight, and destination, and affix correct postage.Needs a human
Check mail to ensure correct postage and that packages and letters are in proper condition for mailing.Needs a human
Sort incoming and outgoing mail, according to type and destination, by hand or by operating electronic mail-sorting and scanning devices.Needs a human
Obtain signatures from recipients of registered or special delivery mail.Needs a human
Answer questions regarding mail regulations and procedures, postage rates, and post office boxes.AI helps
Transport mail from one work station to another.Needs a human
Sell and collect payment for products such as stamps, prepaid mail envelopes, and money orders.Needs a human
Keep money drawers in order, and record and balance daily transactions.Needs a human
Register, certify, and insure letters and parcels.Needs a human
Complete forms regarding changes of address, or theft or loss of mail, or for special services such as registered or priority mail.AI helps
Receive letters and parcels, and place mail into bags.Needs a human
Put undelivered parcels away, retrieve them when customers come to claim them, and complete any related documentation.Needs a human
Respond to complaints regarding mail theft, delivery problems, and lost or damaged mail, filling out forms and making appropriate referrals for investigation.Needs a human
Provide assistance to the public in complying with federal regulations of Postal Service and other federal agencies.Needs a human
Rent post office boxes to customers.Needs a human
Provide customers with assistance in filing claims for mail theft, or lost or damaged mail.Needs a human
Feed mail into postage canceling devices or hand stamp mail to cancel postage.Needs a human
Cash money orders.Needs a human
Order retail items and other supplies for office use.AI helps

Is it better than a person? The evidence

No direct test against people in this job yet. Every study is , and vendor studies are labelled as such.

When could it be replaced?

When AI could largely do this job: no sooner than 2037

Most likely after 2037 (8 in 10 of our scenarios). A range from our of how fast AI improves, how fast employers take it up and what holds it back, not a forecast that the job ends. “” has a strict meaning here. Today’s answer is at the top of the page; this is how it could change.

The sand is the human working years left, measured in the same 40-year glass for every job, so a safe trade starts nearly full and an exposed job with a thin layer.

The sand is the human working years left, in the same 40-year glass for every job.Years still needing a humanYears run out

How this job could shift, year by year

Where the job could sit on our scale each year to 2060, across the ten behind its .

Today
Will AI replace this job?
Nah.
By 2045
60%
of our scenarios have AI largely doing this job by 2045 (Largely.)
10% still have it mostly needing a person (A little. or Nah.)
By 2060
90%
of our scenarios have AI largely doing this job by 2060 (Largely.)
10% still have it mostly needing a person (A little. or Nah.)

We run this job as ten scenarios spread across its replacement range. In each, the score moves towards the bottom band (Largely: AI could largely do the job) by the year that scenario reaches it, slowly at first and faster later, as adoption usually goes. Each bar splits the ten by the band they put the job in. The model stops at 2060. How the timeline works

Share of this job's scenarios in each verdict band, today to 20600%25%50%75%100%2026: 100.0% of scenarios: this job mostly needs a person (Nah.)100%Today2030: 10.0% of scenarios: this job mostly needs a person (Nah.)10%2030: 80.0% of scenarios: AI could do a little of this job (A little.)80%2030: 10.0% of scenarios: AI could partly do this job (Partly.)10%20302035: 10.0% of scenarios: this job mostly needs a person (Nah.)10%2035: 30.0% of scenarios: AI could do a little of this job (A little.)30%2035: 30.0% of scenarios: AI could partly do this job (Partly.)30%2035: 20.0% of scenarios: AI could mostly do this job (Mostly.)20%2035: 10.0% of scenarios: AI could largely do this job (Largely.)10%20352040: 10.0% of scenarios: this job mostly needs a person (Nah.)10%2040: 30.0% of scenarios: AI could partly do this job (Partly.)30%2040: 20.0% of scenarios: AI could mostly do this job (Mostly.)20%2040: 40.0% of scenarios: AI could largely do this job (Largely.)40%20402045: 10.0% of scenarios: this job mostly needs a person (Nah.)10%2045: 10.0% of scenarios: AI could partly do this job (Partly.)10%2045: 20.0% of scenarios: AI could mostly do this job (Mostly.)20%2045: 60.0% of scenarios: AI could largely do this job (Largely.)60%20452050: 10.0% of scenarios: this job mostly needs a person (Nah.)10%2050: 20.0% of scenarios: AI could mostly do this job (Mostly.)20%2050: 70.0% of scenarios: AI could largely do this job (Largely.)70%20502055: 10.0% of scenarios: this job mostly needs a person (Nah.)10%2055: 90.0% of scenarios: AI could largely do this job (Largely.)90%20552060: 10.0% of scenarios: this job mostly needs a person (Nah.)10%2060: 90.0% of scenarios: AI could largely do this job (Largely.)90%2060
Will AI replace the job?Largely.Mostly.Partly.A little.Nah.
Share of this job's scenarios in each band, year by year. Updated with every release.
Show the data
YearLargelyMostlyPartlyA littleNah
Today (2026)0.0%0.0%0.0%0.0%100.0%
20300.0%0.0%10.0%80.0%10.0%
203510.0%20.0%30.0%30.0%10.0%
204040.0%20.0%30.0%0.0%10.0%
204560.0%20.0%10.0%0.0%10.0%
205070.0%20.0%0.0%0.0%10.0%
205590.0%0.0%0.0%0.0%10.0%
206090.0%0.0%0.0%0.0%10.0%

What’s stopping AI taking over?

The things that keep this work with people, strongest first. Each is scored 0 to 100 from work context, licensing and the evidence we have.

Clients want a personFace-to-face contact is rated 3.7 and physical closeness 4.3 out of 5; caring for or serving people is 2.5 out of 5 in importance.
Evidence gapNo study yet compares AI with people doing this job, so employers have no proof it is good enough.
LiabilityMistakes are rated 2.3 out of 5 for consequence and decisions 3.6 out of 5 for impact; someone has to answer for them.
Physical work68% of the task time is physical; robots have been shown on 100% of that time.
RegulationWorkers rate responsibility for others' health and safety 3.0 out of 5.
LicensingUsual entry requirement (BLS): no formal educational credential, then short-term on-the-job training; 3 task statements mention a licence or certification.

What would it cost to hand the work to AI?

The share of the year AI could handle (287 of 2,080 hours a year), priced two ways. Both are ranges, not quotes.

AI model usage, a year
$30–$2,870
A person’s wage for the same hours
$5,880–$10,350

AI cost covers model usage only: no integration, licences, oversight or the human time still needed to review the work. Human cost is the wage for the same hours, without benefits or overheads. As of 2026-10.

Robots and humanoids

AI software can only take the work at a screen. The rest needs a robot that can do it.

68%
of the task time is physical work
Mobile robots
the kind of robot the physical work would need
Commercial in warehouses, hospitals and some outdoor sites; hands are still limited.

Source: Anthropic Economic Index, 'What work can robots do?' (30 September 2026); O*NET 31.0 task weights.

Which AI skills does this job lean on?

The job’s task time split by what an AI model would need to be good at, and where models stand today.

Each star is a task, grouped by the AI skill it leans on.Needs a human 85%AI helps 15%AI does it 0%
Writing · 5.8% of time
Strong
Drafts, edits and translates most routine documents at professional quality.
Analysis · 6% of time
Strong
Reliable on structured data and rules; uneven on judgement calls with thin information.
Coding · 0% of time
Strong
Agents complete many routine software tasks end to end; larger systems still need people.
Vision and design · 0% of time
Good
Reads documents, images and layouts well; specialist imaging needs dedicated, approved tools.
Speech · 5.6% of time
Good
Voice agents handle routine calls and live interpreting; complex or sensitive calls still go to people.
Planning and agents · 7.7% of time
Emerging
Multi-step agents work in narrow, well-tooled workflows; open-ended coordination is unreliable.
Physical manipulation · 61.7% of time
Early
Robots handle structured, repetitive handling; general dexterity outside fixed settings is not commercial.
Care and persuasion · 13.3% of time
Limited
Can script, coach and advise; trust, presence and accountability still need a person.
Amber matter holds its orbit, slate circles the inner disc, white falls in.Needs a human 85%AI helps 15%AI does it 0%
How exposed is it?

Still needs a human: 80/100↑ safer

The amber matter on the outside holds its orbit: that is the task time that still needs a person. The slate matter circling the inner disc is the work AI helps with. The white matter is the share AI can do; it spirals in and is gone over the edge.

The matter follows the job’s task split: 85% needs a human, 15% AI helps, 0% AI does it. Still needs a human: 80/100 ↑ safer. Will AI replace them? Nah.

What the AIs say

We asked four AI assistants the same question: will AI replace this job in the next 10 years? One word (Yes, Partly or No), then one sentence. Our own answer, from the data: Still needs a human: 80/100 ↑ safer. Will AI replace them? Nah.

ChatGPTPartly

AI and automation will take over many routine postal tasks, but human clerks will still be needed for complex transactions, exceptions, and in-person customer service.

gpt-5.5-2026-04-23 · asked 2026-10-03
ClaudeNo

Postal clerk roles involve physical sorting, customer service, and handling varied in-person transactions that remain difficult to fully automate within a decade, though some tasks will likely be augmented by AI and automation.

claude-sonnet-5 · asked 2026-10-03
GeminiPartly

While AI and automation will increasingly handle routine tasks like sorting, package intake, and basic customer inquiries, human clerks will still be needed for complex problem-solving, identity verification, and personalized customer service.

gemini-3.8-flash · asked 2026-10-03
PerplexityPartly

AI and self-service systems will reduce routine postal-clerk work, but human clerks will likely remain for complex transactions, exceptions, and customer assistance.

sonar · asked 2026-10-03

Assistants answer from what they learned in training (Perplexity also searches the web), so they can be confidently wrong, and the same question can get a different answer tomorrow. Our score is built from task data and graded evidence. Answers collected through DataForSEO.

Cite this page

NeedsAHuman.com (2026). Will AI replace Postal Service Clerks? Nah. Still needs a human: 80/100, higher is safer; release 2026-Q4. https://needsahuman.com/jobs/postal-service-clerks/ (accessed 4 October 2026).

Scores change with each , so cite the release. The data is open under : credit NeedsAHuman.com with a link. Open data · Press

Put the badge on your site

The badge updates itself with each release and links back to this page.

Sources

  • Tasks and work context: 31.0, ().
  • Jobs, pay and projections: US , and 2025–35.
  • How AI is used today: ; Microsoft Research, .
  • What AI can do: our task ratings ( r1) and the quality evidence register.
  • UK names and employment: coding index and .

How each score is built: methodology. Every figure on this page: open data. Release 2026-Q4.