The Gap Between the Narrative and Reality Always Gets Filled by Cheaper Labour

Every generation gets told technology will replace workers. Every generation, the same thing actually happens instead.

As someone who has spent over two decades in workforce transformation, I can tell you: the script we're watching in 2026 isn't new. It's a rerun.

The current wave of AI-driven layoffs, 49,000+ tech jobs cut in the first two months of 2026 alone, a 118% increase year-on-year per Challenger, Gray & Christmas, has been framed as something unprecedented. A rupture. The moment technology finally catches up to the workforce.

It isn't. We've run this playbook three times before. And the outcomes have been remarkably consistent.

Cycle 1: "The Office of the Future" (1970s–1980s)

The promise of office automation dates back to the early 1970s, when word processing and minicomputers were going to eliminate secretaries, typists, and entire layers of administrative support.

IBM, Wang Laboratories, and Vydec sold the vision of the paperless, people-less office. Business magazines predicted that typing pools would vanish within a decade.

What actually happened was more instructive. The roles didn't vanish, they were reclassified. Typing pools became "word processing centres." Those centres eventually gave way to the title of "administrative assistant." The work moved sideways, not away. Headcount did eventually decline in these functions but over 15-20 years, not overnight, and largely through attrition rather than mass displacement.

The technology narrative ran well ahead of the technology reality. Managers refused to type their own documents. The "human factor" as researchers at the time described it proved far stickier than anyone predicted.

The pattern: Promise of mass displacement. Actual outcome: gradual role evolution over decades.

Cycle 2: The BPO/Offshoring Wave (Late 1990s–2010s)

The internet was the next transformative technology. Connectivity was going to change everything about how work gets done. And it did just not in the way the narrative suggested.

The internet didn't eliminate the work. It made it possible to move the work to Bangalore, Manila, Hyderabad, and Chennai at 20-35% of the cost.

India's BPO sector went from roughly $1 billion in revenue in 2001 to over $30 billion by 2020. GE, American Express, and British Airways were among the early movers, proving the model worked. By the mid-2000s, India commanded nearly 70% of the global outsourcing market.

The narrative sold to shareholders was "digital transformation." The operational reality was labour arbitrage. The technology was the enabler of cheaper human labour, not the replacement for it.

For anyone who lived through this era, and I spent significant parts of my career across India, Philippines, Singapore, and the US during the peak of it the parallels to 2026 are impossible to ignore.

The pattern: Technology enables new ways of working. Companies use the narrative to move jobs, not eliminate them.

Cycle 3: RPA / "Intelligent Automation" (2015–2022)

UiPath, Automation Anywhere, and Blue Prism sold the next vision: robotic process automation would eliminate swathes of back-office and finance roles. The pitch was "bots replacing humans" in accounts payable, claims processing, and data entry.

Enterprise spending on RPA surged. Every Big Four firm built an automation practice. Conference stages were filled with projections about millions of roles being automated within years.

What happened? Companies bought the tools, partially automated some processes, and then needed more humans to manage the exceptions, maintain the bots, handle edge cases, and do the work the automation couldn't handle. RPA became a supplement to existing workflows, not a substitute for existing workers.

Meanwhile, the BPO providers themselves adopted the tools not to eliminate their own workforces, but to upskill them and move up the value chain. The automation that was supposed to kill outsourcing actually strengthened it.

The pattern: Bold claims of technology-driven displacement. Actual outcome: technology augments human work, adoption takes far longer than promised, and the labour arbitrage model persists.

Cycle 4: "AI Replaces Everyone" (2024–Present)

Generative AI arrives. CEOs declare transformation. Block cuts 4,000 (40% of workforce). Amazon eliminates 30,000. Pinterest cuts 700. eBay drops 800. CrowdStrike trims 500. The stated reason, almost universally: AI.

The stock market rewards each announcement with 10-25% jumps. The narrative is clean, compelling, and investor-friendly.

But the data tells a more complicated story.

Challenger, Gray & Christmas data shows AI was cited as a factor in just 7% of January 2026 job cuts. Wharton's Ethan Mollick has observed that there is little evidence to support sudden 50%+ efficiency gains from AI at any company operating at scale. And Forrester's "Predictions 2026: Future of Work" report found that 55% of companies that cut staff citing AI subsequently rehired — at lower cost, or offshore, or both.

Perhaps most tellingly, Forrester pointed to Amazon's "Just Walk Out" retail technology — marketed as AI-powered autonomous checkout which was found to be largely managed by remote Indian workers reviewing camera feeds. They called these "ghost workers" and predicted the pattern would spread: more companies rehiring under the guise of AI at lower wages and in lower-cost geographies.

A Blind survey of 2,392 tech workers in January 2026 put numbers to the pattern: 50%+ said their companies were increasing India hiring. 38% said the offshore hiring was replacing US roles, not complementing them. 28% attributed the shift partly to H-1B visa restrictions pushing hiring offshore rather than onshore.

The Through-Line

Across all four cycles, the mechanism is the same:

Step 1: A new technology arrives. CEOs declare it will change everything.

Step 2: Companies use the narrative to justify workforce reductions that are actually driven by cost pressure, overhiring, or margin targets.

Step 3: The eliminated roles quietly reappear — in a different geography, at a different price point, under a different title.

Step 4: The technology eventually catches up, but on a much longer timeline than the narrative suggested, and the transition involves far more human labour than anyone publicly admitted.

The technology narrative consistently runs ahead of the technology reality. And the gap between narrative and reality gets filled by cheaper human labour somewhere else.

Where This Cycle Might Diverge

There is a credible argument that generative AI is fundamentally more capable than the technologies of previous cycles. Word processors, the internet, and RPA bots each had hard ceilings on what they could automate. Large language models and agentic AI do not appear to have the same constraints at least not yet.

The endgame where technology genuinely does the work may arrive faster this time.

But "faster" is relative. Forrester is also predicting that companies will delay a quarter of their AI spend into 2027 because they cannot yet find measurable returns. McKinsey reports that 33% of organisations are using generative AI for workforce reductions, but even their most optimistic timelines suggest 2-4 years before meaningful process-level automation is reliably deployed.

In the meantime, what I'm observing is a three-phase pattern that mirrors the historical playbook with a distinctly modern twist:

Phase 1 (now): Cut high-cost roles in expensive markets. Frame it as AI-driven transformation. Generous severance buys goodwill.

Phase 2 (12-24 months): Backfill at 30-70% lower cost in India, Southeast Asia, and Eastern Europe. Maintain output. Free up capital. This is the phase that doesn't appear in earnings calls or CEO letters.

Phase 3 (2-4 years): Use the freed capital to build the AI tooling that eventually automates the offshore roles too. The offshore workforce becomes the transitional bridge, not the destination.

Oracle's recent moves provide an interesting proof point from the other direction. They cut ~3,000 India-based roles while simultaneously investing $5 billion in UK cloud infrastructure and hiring selectively in Virginia — suggesting they may be further along the curve where AI is now displacing the offshore bridge workforce itself.

What This Means for Leaders

If you're a CEO, CHRO, or board member navigating this moment, the historical record suggests a few things worth keeping in mind.

First, be honest about what you're actually doing. If the cuts are about correcting pandemic-era overhiring and improving margins, both legitimate business decisions, say that. Dressing up a cost play as an AI transformation erodes trust with the very workforce you'll need to execute the actual transformation later.

Second, understand the sequencing. The companies that will navigate this well are the ones that treat Phases 1 through 3 as an integrated workforce strategy, not a series of disconnected moves. The ones that will struggle are the ones making Phase 1 cuts today with no clear plan for how AI actually changes their operating model in Phase 3.

Third, respect the timeline. No company has demonstrably proven that AI can replace 40% of a knowledge workforce overnight. Klarna reduced headcount by a similar proportion through natural attrition over two years, not a single-day shock. The difference between those approaches isn't just optics. It's operational discipline versus performance.

And finally, remember that history has been remarkably consistent on this point: the technology always arrives. But it arrives slower than the narrative.

The data and research referenced in this article draws from Forrester's "Predictions 2026: Future of Work" report, Challenger, Gray & Christmas employment data, Blind's January 2026 tech worker survey (n=2,392), and company-specific filings and reporting from CNBC, Reuters, Bloomberg, and TechCrunch.