Will AI replace correspondence clerks?

Some tasks are already handled by AI or machines; the rest still needs a person. About 3h 57m of an 8-hour shift goes to tasks that AI or robots largely take over. The other 4h 3m (51%) stays with people. Separately, BLS expects about 5.6% fewer of these jobs by 2035 (all jobs: +3.5%).

AI shows up regularly in 4 of 17 tasks, about 3h 57m of a typical shift. The biggest ones:

• Compose letters in reply to correspondence concerning such items as requests for merchandise, damage claims, credit information requests, delinquent accounts, incorrect billing, or unsatisfactory service. (1h 30m a shift, AI mostly does it alone)
• Read incoming correspondence to ascertain nature of writers' concerns and to determine disposition of correspondence. (1h 4m a shift, AI mostly does it alone)
• Gather records pertinent to specific problems, review them for completeness and accuracy, and attach records to correspondence as necessary. (51m a shift, AI mostly does it alone)
• Complete form letters in response to requests or problems identified by correspondence. (32m a shift, AI mostly does it alone)

Not in workplaces like this one, based on a 2026 review of robot deployments. 2 more tasks have only been done in factory setups built around the machine, so they stay with people here.

4h 3m of a typical shift, or 51%, goes to tasks with no AI or robot evidence yet. No AI usage or robot deployment has shown up for these so far. The biggest:

• Maintain files and control records to show correspondence activities. (45m a shift)
• Prepare documents and correspondence, such as damage claims, credit and billing inquiries, invoices, and service complaints. (45m a shift)
• Review correspondence for format and typographical accuracy, assemble the information into a prescribed form with the correct number of copies, and submit it to an authorized official for signature. (26m a shift)
• Compile data from records to prepare periodic reports. (19m a shift)
• Route correspondence to other departments for reply. (19m a shift)

Part of the answer is the setting, not just the tasks. Work that is in person, hands on or high stakes is where AI and robots have made the least headway so far. In the official O*NET survey of people in this job:

• 87% talk face to face with others every day
• 79% are in constant contact with other people
• 65% say dealing with the public is very or extremely important

For the 4 tasks in this job where AI shows up and timing data exists, a typical request takes about 1h 13m for a person alone and about 21m with AI. These are Claude's own estimates from real conversations in April and May 2026 (Anthropic Economic Index). They describe single requests, not whole workdays, and time saved on a task doesn't mean jobs are cut. The biggest gaps:

• Gather records pertinent to specific problems, review them for completeness and accuracy, and attach records to correspondence as necessary (about 4h 44m alone, 43m with AI)
• Compose letters in reply to correspondence concerning such items as requests for merchandise, damage claims, credit information requests, delinquent accounts, incorrect billing, or unsatisfactory service (about 1h 28m alone, 26m with AI)
• Complete form letters in response to requests or problems identified by correspondence (about 58m alone, 16m with AI)

On Indeed, job postings in the Customer Service sector are at 87% of their February 2020 level (all U.S. postings: 104%), down 1.5% from a year earlier, as of Oct 9, 2026. BLS separately expects about 500 openings a year in this job through 2035. Across all U.S. postings, 7.3% now mention AI skills. Sector postings cover more than this one job and move for many reasons besides AI.

The median pay is about $46,800 a year ($22.50 an hour), 8% below the median for all U.S. jobs ($50,980). Most people in this job earn between $33,300 and $64,510 a year; that range covers the middle 80%. Pay figures are from BLS, May 2025, and vary a lot by state and employer. Pick a state to compare:

BLS projects a decline of 5.6% from 2025 to 2035, compared with 3.5% growth for all jobs. That's about 4,500 jobs in 2035. Even so, it expects about 500 openings a year, mostly from people retiring or changing careers. These projections don't single out AI as a cause.

About 4,290 people in the U.S., roughly 0.00% of all workers (BLS, May 2025). That makes it the 738th largest of the 796 jobs on this site. BLS expects about 4,500 in 2035.

The most common requirement is a high school diploma: about 43% of people in the job said so in the official O*NET survey. O*NET rates the job at level 2 of 5 for preparation, which usually means: high school diploma, some training. Requirements vary by state and employer, especially for licensed work.

These come from the job's official skill ratings and from the tasks that AI and robots touch least today. Building them keeps you strong in the parts of the work that still need a person:

• Skills that matter more here than in most jobs: Writing, Reading Comprehension, Active Listening, Service Orientation
• Get strong at the parts of the job that stay with people (listed above), which take 4h 3m of a typical shift
• Tools employers list: Microsoft Word, Microsoft Outlook, Microsoft Office software, Microsoft Excel, Epic Systems
• Learn to use AI well on the tasks where it already shows up in this job (listed above)

These use similar skills, with less of a typical shift touched by AI or robots. Matches come from 109 official ratings of skills, knowledge and work activities, and each needs the same or less preparation unless the official list relates the jobs directly:

• Bookkeeping, Accounting, and Auditing Clerks (1h 58m touched, $51k median pay)
• Court, Municipal, and License Clerks (1h 28m touched, $49k median pay)
• Bill and Account Collectors (2h 38m touched, $47k median pay)
• Insurance Claims and Policy Processing Clerks (1h 20m touched, $49k median pay)
• Legal Secretaries and Administrative Assistants (42m touched, $56k median pay)

Each task carries its own evidence label. For this job: 4 observed in real AI use, 13 with no AI or robot evidence found. The minutes per task are estimates, not measured hours, and "no evidence" doesn't prove a task is safe.