Society Issue #120 ¡

History Repeats? The 📠 Office Automation Era of 1979

Workers feel 3x faster while executives measure 1.8% — the exact gap that opened when word processors first hit the office.

Oswarld (Kwangseob Ahn) ¡ GTM & AI strategist working the seam between technology and business

History Repeats? The 📠 Office Automation Era of 1979

Opening

Dear reader, in 1979 the BBC aired a documentary — a 10-minute segment titled “Will Word Processors Start a Home Working Revolution?” The reporter declares to the camera:

“An office with no clerks, no secretaries, no mail carriers, and no paper is already up and running!”

47 years later, in 2026, we are hearing a strikingly similar sentence. “AI agents handle your code, your reports, and your email on their own.” Isn’t that totally AI-Native? And the same question lingers now as it did then: if the individual experience is so dramatic, why won’t productivity across the whole economy budge?

Let me give you the conclusion up front: this is not a technology problem, it is an organization problem. And the evidence was already there 47 years ago.

🖨️ 1979: The Office Where the Typewriters Went Silent

Start with the numbers — they were overwhelming. Britain’s Bradford City Council introduced a word processor system in 1977, then cut its staff in half and lifted productivity by 40%. It sent out the 25 letters required to process a mortgage in 2 minutes, and a single department saved £60,000 a year. 37,000 A4 pages fit into one box of magnetic disks, and entire shelves of paperwork simply disappeared.

Remote work had already begun, too. At a company called F International, more than 600 freelancers were programming computers from home; programmer Linda English wrote programs on a bubble-memory terminal while looking after her child, then transmitted them over the phone line. And the French government was giving away 30 million terminals to households for free, on the grounds that it was cheaper than printing phone books.

The documentary made this forecast: “As word processors become the connecting links of information networks, the layer of middle managers could become unnecessary.” And it left one final sentence: “Rather than destroying the traditional notions of home and family, the silicon chip is helping to rebuild them.”

What is more interesting is that even the frame used to interpret the change resembles today’s. The documentary described office work migrating into the home as “a revival of the cottage industry that predated the Industrial Revolution.” It also covered the skeptics’ worry — that “this technology will make people uniform” — to which the reporter countered that “in practice it is helping people live more freely and individually.”

The paperless office. Remote work. The death of middle management. Work-life balance. The fear that technology homogenizes people, and the rebuttal that it actually sets them free. Don’t these promises and debates sound awfully familiar?

💻 2026: The Office Where Copilot Moved In

47 years later, the surface numbers are once again overwhelming.

According to METR’s survey of technical workers published this May, developers using AI tools reported that their work felt 3x faster. Converted to value terms, it was still 2x. At the level of individual tasks, AI is clearly changing the speed of work.

But turn the camera toward organizations and the economy as a whole, and the landscape changes completely.

In a joint study of roughly 750 corporate executives by Duke University’s business school and the Richmond and Atlanta Federal Reserve Banks, the AI-driven productivity gain executives reported averaged 1.8%. Yet when the researchers worked backward from actual revenue and employment data, the figure was far smaller. The team named this gap the “productivity paradox.” Perception has changed; revenue hasn’t caught up.

PwC’s 2026 AI performance report is even more sobering. Surveying 1,217 executives across 25 industries, it found that the top 20% of companies were capturing 74% of the economic value AI has created. The remaining 80% of companies were still stuck at the pilot stage. Goldman Sachs’s analysis cuts sharper still: 70% of S&P 500 companies mention AI, but only 1% have actually quantified AI’s contribution to earnings.

The long-range forecast from the University of Pennsylvania’s Wharton School is telling as well. It projects that AI’s contribution to GDP will stay at 1.5% through 2035. That is a long way from the ‘explosive change’ individuals say they feel.

This May, Fortune summed up the situation in one sentence: “AI is visible everywhere, but not yet in the productivity statistics.” Doesn’t that overlap with the rosy numbers echoing out of the 1979 office?

🔄 The Solow Paradox, Season Three

In 1987, Nobel laureate economist Robert Solow said this:

“You can see the computer age everywhere but in the productivity statistics.”

That is the Solow paradox. Through the 1970s and 80s, American companies poured massive investment into IT, yet productivity growth actually fell from 3% a year to 1%. Personal computers landed on every desk, and the economy as a whole got slower.

The paradox wasn’t resolved until the mid-1990s. Productivity jumped only 10 to 15 years after the technology entered the office. Analyzing that period, McKinsey pointed out that the needle for the whole economy moved only after a handful of sectors (tech, retail, wholesale) fundamentally redesigned their work processes around the technology.

Economic historian Paul David found the same pattern further back in the past. The electric motor was invented in the 1880s, but US manufacturing productivity didn’t actually leap until the 1920s — 40 years later. The reason is clear. Factories kept the steam-age design — every machine belted to a single giant power source — and simply swapped in motors. Electricity’s real potential didn’t materialize until factories themselves were redesigned around distributed power.

In economics this is called the productivity J-curve. When a new technology arrives, productivity initially stalls or even falls because of learning and adaptation costs. Only after the organization restructures itself around the technology does the curve shoot upward.

AI may be passing through the bottom of that J-curve right now. Individual speed is 3x; organizational productivity is 1.8%. Structurally, this is the same point as the 1900s, when electric motors were bolted into steam-engine factories, and as 1979, when typewriters gave way to word processors but the workflow stayed the same. If one thing is different, it’s that the lag is shrinking. Electric motors took 40 years; computers took 15. How long will AI take?

Oswarld’s Take

Honestly, I have seen this pattern far too many times.

Working on technology strategy, I have reviewed countless proposals claiming “adopt this technology and productivity rises N%.” Productivity never resolves that neatly in the first place… and it doesn’t measure that cleanly either. And in the cases where the promise actually came true, it wasn’t because the technology was brilliant — it was because the organization changed how it worked to fit the technology. If the tool changes but the meeting structure stays the same, the decision paths stay the same, and the performance metrics stay the same, then perception changes and measurement doesn’t.

Look closely at why Bradford City Council succeeded in 1979. It wasn’t because the word processor was brilliant. They built standard templates, redesigned the workflow around a central memory system, and even restructured the workforce. They didn’t adopt a technology — they rebuilt the organization around it. The top 20% of companies capturing 74% of AI’s value in 2026 are almost certainly doing the same thing.

One caveat, though: please don’t read this as optimism that “everything works out in the end.” The electric motor took 40 years; the computer took 15. Across that long lag, the organizations and individuals who failed to redesign were weeded out. The direction is right, but the timing and the path always defy expectations. That is the real lesson of 47 years of repetition.

Closing

Whether it’s 1979’s 40% or 2026’s 3x, there is a long lag before dramatic improvement in individual tasks translates into the organization as a whole. And what shortens that lag is not better technology but organizational redesign. Most of the predictions in that 47-year-old documentary came true — but every path they imagined was wrong. Big change doesn’t happen in an instant, and it doesn’t happen through a few lectures and a few seminars. Isn’t this trendy word “AI-Native” amusing, too? What exactly is supposed to be Native? Confidently claiming you’ve already implemented something you can’t define looks a lot like the Agentic-versus-Agent terminology fight that was burning hot until just a few months ago. Don’t get eaten by the vocabulary. New technology will keep arriving regardless. What matters is how you absorb it and how you apply it.

Take stock this week. Is the AI tool you’re using right now merely ‘sitting on top of’ your existing work, or is it changing the work itself?

Tell me in the comments about the gap between what you “feel” and what you can “measure.”

References & Further Reading

Primary sources

Background

The author, Kwangseob Ahn, is a professor of business administration at Sejong University and lead consultant at OBF (Oswarld Boutique Consulting Firm). He teaches statistics and data analysis — including business data management and business analytics — at the university, while leading GTM strategy and AI strategy consulting in the field, designing the interface between technology and business. He has published academic research on memory architecture for AI dialogue systems (HEMA) and runs Daily Arxiv, a project curating global AI papers every day. He graduated from the master’s program at Korea University’s Graduate School of Management of Technology and from its KMBA program. He is the author of “Homo Brainless: The People Who Outsource Their Thinking”.