Remote Work, Not AI, Is Driving the Five-Year Productivity Boom, New York Times Reports

There's a story corporate America has been telling itself for the past year. Productivity is surging, AI is everywhere, so AI must be driving productivity. It's a clean narrative. It justifies the capex and it looks great on an earnings call.

The problem is that the dates don't line up.

As a new New York Times article highlights, the current productivity acceleration began roughly five years ago. U.S. nonfarm business output has grown about 2% annually over that stretch, roughly double the pace of the 2010s, a run strong enough that Fed Chair Jerome Powell has said he never expected to see this many consecutive years of high productivity. Widespread enterprise AI adoption, meanwhile, didn't kick in until around 2025. You can't credit a technology for a boom that started three years before anyone was using it. And when researchers survey firms directly, companies say the same thing: they expect AI to boost productivity in the future, but the measured effects so far are limited.

So what actually moved the needle? The Times points to two forces that are far less fashionable right now: digitization and remote work.

The mechanism is boring, which is exactly why it's believable

The remote work argument isn't mystical. KKR, cited in the piece, put it plainly: distributed work let companies hire from a bigger talent pool. When your labor market is a 40-minute commute radius, you hire the best person within driving distance. When it's a whole country, you hire the best person, period. Better matches between people and jobs is one of the most reliable productivity mechanisms in economics. No hype cycle required.

Add a decade of forced digitization on top of that, with every process, document, and meeting dragged online between 2020 and 2023, and you have a plausible, unglamorous explanation for the boom that doesn't require a single GPU.

The aggregate data also lines up with a growing stack of research. A Bureau of Labor Statistics analysis found that industries with bigger increases in remote work saw faster growth in total factor productivity, even after controlling for pre-pandemic trends. Stanford economist Nicholas Bloom, whose team runs the ongoing survey work at WFH Research, points to a clear post-2020 surge in productivity growth arriving exactly when work from home ramped up. And randomized controlled trials, including Bloom's large hybrid-work experiment published in Nature, keep finding that well-designed hybrid and remote arrangements can improve productivity and employee well-being at the same time. Not "remote work is fine, actually." Better on both counts, when it's done deliberately.

The awkward part

Hold this finding up against the last eighteen months of corporate behavior and the dissonance is loud.

The same companies mandating five days back in the office "for productivity" may have been unwinding one of the actual drivers of the productivity boom. The same CEOs attributing headcount reductions to AI efficiency gains are crediting a technology that, by firms' own survey responses, hasn't delivered measurable productivity effects yet. We've written before about executives hiding behind the AI narrative. This is more evidence that the narrative and the data live in different buildings.

None of this means AI won't matter. It almost certainly will, and probably enormously. The honest read is that AI's productivity story is still in front of us, not behind us. But "will matter" and "already explains the boom" are different claims, and conflating them is how companies end up making billion-dollar real estate and workforce decisions on vibes.

What this means if you're building a brand

Your RTO mandate is a brand statement, whether you meant it or not. If the evidence increasingly says well-designed flexibility drives productivity, then a blanket mandate reads as what it usually is: a control preference dressed up as a performance policy. Candidates can read data too. In a market where the best talent has options, "we trust you to work where you work best" is a recruiting asset with actual economic evidence behind it.

Stop borrowing AI's credibility for outcomes it didn't produce. Attributing your efficiency gains to AI when the real driver was hiring reach, digitized workflows, or plain old operating discipline is a trust withdrawal. When the AI results eventually do show up, you'll have already spent the credibility you need to announce them.

The unglamorous explanation is usually the durable advantage. Talent-pool expansion isn't a keynote topic. Nobody's putting "we hire from more zip codes now" on a billboard. But boring, structural advantages compound quietly while everyone else chases the shiny narrative. The companies that figured out distributed work as an operating system have been banking the returns for five years while the discourse argued about badge swipes.

The productivity boom is real. The question is whether companies will learn the right lesson from it, or keep crediting the technology with the better publicist.