A stock check once meant finding someone with the right ledger. Oh, how things have changed.
For much of the 20th century, getting an answer at work meant seeking out someone who could find it for you. Computing gradually put records, calculations and transactions within reach of employees and customers. Now AI is extending that access from retrieving information to asking systems to do something with it.
1950: The answer was somewhere in the building
In a typical office, information occupied space: filing cabinets, account ledgers, stock cards, printed catalogs and reference books. Correspondence was typed, carbon copies filed and forms carried between departments. A purchase could leave separate paper trails in purchasing, warehousing and accounts.
Checking an invoice meant locating the order and delivery record, comparing entries and asking about discrepancies. The telephone shortened the journey, but somebody at the other end still had to look. Postal correspondence carried inquiries and documents farther afield; colleagues’ memories supplied details the files never contained.
This was already a mechanized workplace. Typewriters, adding machines and calculators supported clerical work; larger organizations used punched-card equipment to sort and tabulate records. Access nevertheless depended heavily on being near the records, understanding the filing system and knowing whom to ask.
1950s–1960s: Faster work, delivered in batches
Early business computers took on work with large volumes of records and repeated calculations. In 1951, the British catering company J. Lyons & Co. used LEO to value its bakeries’ output. Payroll, accounts and billing became major applications of business computing.
Employees collected timesheets, orders and other source documents. Data preparation staff entered and checked records, often using punched cards. Operators ran scheduled jobs; printed pay stubs, invoices and management reports went back to departments. Calculations that consumed clerical hours could be completed much faster, and the same input could support several outputs.
But a computer in the organization did not put answers on every desk. An unexpected question might still require a requested report and another processing run. Programmers and operators became part of the workflow, while employees still investigated missing records and disputed figures.
1960s–1970s: Answers at the terminal
Shared computing preceded the personal computer. Early time-sharing systems, including MIT’s CTSS, demonstrated in 1961, let multiple users interact with a central machine. Transaction systems brought direct access into operational work.
American Airlines’ Sabre, fully operational in 1964, connected reservations staff to central flight information. They could check availability and make reservations during an inquiry. From 1976, Sabre terminals reached travel agencies: customers could visit or call an agent who searched and booked on their behalf.
In businesses adopting online inventory systems, staff could query stock records and allocate goods while taking orders, rather than wait for a separate stock-control inquiry. The crucial capability was updating the shared record as well as reading it. Access remained concentrated among trained employees using specialized terminals and procedures.
1980s: A computer of your own
Desktop software made everyday work more directly editable.Word processing let people move paragraphs and correct mistakes without retyping whole pages. Personal databases helped staff organize contacts and orders. Spreadsheets recalculated budgets when prices, wages or sales assumptions changed.
VisiCalc had brought that spreadsheet capability to the Apple II in 1979. IBM’s PC arrived in 1981, accelerating business adoption within an existing personal-computer market. Employees could prepare reports and explore figures without commissioning a central computing department to write a program.
Some typing and calculation tasks disappeared; others moved to the people drafting documents or making decisions. Files lived on floppy disks and local drives, traveling between desks as copies. Two people could improve the same budget independently and leave someone else to establish which improvement counted.
1980s–1990s: Connecting the desks
Local networks made shared printers, files and applications practical across offices. Ethernet, developed at Xerox PARC in the 1970s, became a foundation of this connected workplace. Links between branches extended access beyond one building.
With integrated applications and common databases, sales, warehouse and purchasing teams could work from the same order and stock records. A sale entered once could inform allocation, invoicing and replenishment. Stock at another branch became information staff could act on without first calling that branch.
Cabling alone could not deliver this. Software had to reconcile product codes, permissions and business processes. Administrators managed accounts, backups and availability; employees maintained records. Sharing an outdated file simply gave more people the same outdated answer, a problem still familiar in network documentation.
1990s–2000s: Finding what was out there
International electronic communication and remote access existed before the web. Browsers, websites and search made a much wider range of information approachable: supplier specifications, technical manuals, research and troubleshooting advice could be found without requesting a catalog or waiting for support to respond.
Connectivity did not guarantee discovery. Sun Microsystems’ 1994 SunWeb project addressed internal information scattered across inconsistent formats and hard-to-find locations, even though employees already used networked workstations. Organizing information around users’ needs mattered alongside making it accessible.
Intranets gave policies and forms a common home. Expanding email use sped exchanges across organizations and time zones, while filling inboxes with attachments and competing demands. Research increasingly involved judging which available answer was reliable, current and applicable.
2000s–2020s: Access beyond the office
Broadband, laptops and smartphones brought work systems into homes, customer premises and journeys. Hosted applications and collaborative documents let colleagues work on a shared version instead of circulating attachments. The office increasingly became one place to access work rather than the place where work was kept.
Customer self-service, already emerging in the 1990s, became routine. Travelers compared flights and booked directly; bank customers transferred money; buyers placed orders and checked delivery status. Tasks once performed by staff moved to customers, on their own schedules. Complex cases still required assistance.
These online services did not depend on public cloud infrastructure. Cloud services expanded providers’ options for running them, allowing rented computing and storage to replace some equipment purchases. Operational work shifted partly to service providers, while customer organizations still managed access, integrations and data. Convenience also brought the expectation that people could be reached after leaving work.
2020s–2026: Asking for the work itself
Rules, scripts and automated workflows had long performed defined tasks. Generative AI assistants widened the range of requests people could express in ordinary language: summarize these documents, draft a response, analyze this spreadsheet or explain these logs.
By 2026, tools already released could combine those capabilities with browser interaction, code execution and connected applications. With appropriate access and integrations, an assistant could produce a report or update a record, rather than merely explain how. AI-assisted network management illustrates the usefulness of summarizing operational evidence, and the consequences of accepting an unsupported diagnosis.
This adds interpretation to the access problem. A fluent answer may overlook missing records or invent a connection. Employees must check consequential outputs against evidence; organizations must decide what assistants may access and which actions require review.
No call required
A traveler can now compare fares, book a flight and check in without speaking to an airline employee. Someone preparing a budget can retrieve figures, test assumptions and ask an AI assistant to draft the explanation. Work that once passed through several pairs of hands can now happen at one desk, with little or no need to talk to anyone else.
People who began work with typewriters, paper ledgers, and carbon copies now use smartphones and AI assistants. A traveler can now compare fares, book a flight and check in without speaking to an airline employee. Someone preparing a budget can retrieve figures, test assumptions and ask an AI assistant to draft the explanation. Work that once passed through several pairs of hands can now happen at one desk, with little or no need to talk to anyone else.
People who began work with typewriters, paper ledgers and carbon copies now use smartphones and AI assistants. The strange thing is, much of this no longer feels revolutionary. It seems normal. What will normal workplace computing look like 75 years from now?
Sources
- Computer History Museum: Punched-card work and the IBM 1401
- University of Warwick: Frank Land’s history of LEO
- Multicians: CTSS and the origins of time-sharing
- IBM: Sabre and the development of airline reservations
- Computer History Museum: Software that put PCs to work
- IBM: The 1981 personal computer
- Computer History Museum: Local networks and Ethernet
- Jakob Nielsen: Designing SunWeb in 1994
- Google: Collaborative documents and spreadsheets in 2006
- AWS: The origins of its infrastructure services
- Tableau: What is the history of artificial intelligence?
About NetworkTigers
NetworkTigers is the leader in the secondary market for Grade A, seller-refurbished networking equipment. Founded in January 1996 as Andover Consulting Group, the company originally built and re-architected data centers for Fortune 500 firms. Today, NetworkTigers provides consulting and network equipment to global government agencies, Fortune 2000 companies, and healthcare companies. Visit www.networktigers.com
