Order clerks

Three parts of your work where AI is already doing real lifting, and what stays yours.

• Handwritten order ledger1872–1920

  Handwritten order ledger and postal correspondence (mail-order catalog era)

  The order clerk of the late 19th century worked with pen, ledger, and postal scale. A customer's letter arrived in the morning mail; the clerk compared the requested items against a catalog price list, checked a stock ledger by hand, wrote out a picking ticket for the warehouse floor, calculated the total charge, and composed a postal confirmation. The entire transaction was recorded in a bound order book. Montgomery Ward required clerks to write in a specified hand so that order books were legible to supervisors and auditors. At Sears, the 1906 West Side plant introduced a mechanical conveyor system to move order papers from the correspondence department to fulfillment, but the core task of reading a customer's handwritten order and translating it into a warehouse instruction remained manual.

• Telephone order-taking1920–1965

  Telephone order-taking and typewritten order forms

  By the 1920s, the telephone had become a practical order-taking instrument for commercial wholesale accounts. Industrial distributors, food service suppliers, and regional catalog houses began staffing telephone order desks alongside their mail-order departments. The clerk now spoke with a buyer, typed the order onto a multi-carbon form, and retained copies for shipping, billing, and records. The typewriter shifted the work from a writing-intensive to a keyboarding-intensive task. Telephone order-taking expanded dramatically in the 1940s-1950s as postwar industrialization created large B2B order flows: industrial suppliers, food distributors, and auto parts wholesalers all built telephone order departments to serve regional customers.

• Punch card1965–1985

  Punch card and mainframe order entry (IBM System/360 era)

  IBM's System/360, launched in April 1964, gave large manufacturers and distributors a computing platform capable of handling order management at scale. By the late 1960s, large wholesale houses and catalog retailers began converting paper order forms to punch-card entry systems: a clerk received the customer's order (by phone or mail), then manually keypunched the data onto cards fed into a batch-processing queue. The mainframe produced a picking list, an invoice, and a ledger entry overnight. The clerk's judgment still governed every transaction, but the record-keeping burden shifted from handwritten ledgers to machine-readable media. Order turnaround times fell from days to hours as batch runs replaced manual posting.

  Effect on the work

  Mainframe order systems increased the volume of orders a single clerk could process per day but did not displace the role. By standardizing the order transaction, they also made it easier to hire and train new clerks, which fueled headcount growth as order volumes rose through the 1960s and 1970s.

• EDI1979–2000

  EDI (ANSI X12 standard, 1979) and early screen-based order entry terminals

  The ANSI X12 EDI standard, ratified in 1979, established a machine-readable format for purchase orders, invoices, and ship notices. Early adopters were large retailers and their key suppliers: by 1988, Walmart had mandated EDI compliance for all major suppliers, meaning that a Walmart purchase order arrived in the supplier's order management system electronically rather than on paper. For order clerks, EDI had an ambiguous effect: it automated high-volume, large-account order flows (which required little clerk judgment anyway) while leaving the complex, exception-heavy, and smaller-account orders in human hands. Screen-based order entry terminals, replacing the punch-card keypunch station, gave clerks real-time access to inventory and pricing databases for the first time, dramatically reducing the lookup labor that had previously consumed much of their shift.

  Effect on the work

  EDI adoption through the 1980s-1990s is estimated to have eliminated a sizeable share of order entry labor for large wholesale and retail trade companies by automating high-volume standard transactions. Employment in the order clerk category continued to grow in absolute terms through this period because total order volume grew faster than the per-order labor displaced.

• SAP R/31992–2015

  SAP R/3 and ERP-integrated order management (client-server era)

  SAP R/3, launched in July 1992, brought enterprise order management into the client-server era. By the mid-1990s, SAP and Oracle Order Management had become the systems of record for order entry at large manufacturers and distributors. The clerk now worked entirely within an ERP screen: the customer called, the clerk looked up the account, keyed in part numbers, quantities, and pricing (auto-populated from master data), and the system generated the picking list, invoice, and inventory debit automatically. The advantage was speed and accuracy; the residual job was verification, escalation, and exception handling. As ERP adoption broadened through the 2000s into mid-market companies, order entry became synonymous with ERP screen navigation, and the occupation experienced peak employment in the 2003-2006 OEWS surveys.

• AI order capture2015–present

  AI order capture and automation (Conexiom, Esker, Rossum; 2015-present)

  The current era is the first in which the core task of the order clerk, translating an incoming customer order document into an ERP transaction, is being automated at a rate that directly contracts the headcount. Conexiom, launched to commercial scale around 2015 and relaunched as the "Ideal Order Platform" in February 2025, achieves over 85% touchless autopilot processing by ingesting purchase orders from email, PDF, EDI, and image formats, running many validation checks against ERP master data, and releasing clean orders without human touch. Esker's AI order management processed over 2.87 million orders monthly by 2025 with a 92% reduction in manual entry for documented customers. What remains for human order clerks is the exception queue: orders with mismatched part numbers, credit-hold accounts, unusual quantities, or unrecognized customer formats that the AI cannot process automatically.

  Effect on the work

  BLS projects a decline in order clerk employment from 2024 to 2034. This is the most severe projected decline in the information-clerks occupational group and reflects the direct automation of the role's primary task by commercially deployed AI systems.