The Quiet Rewrite

By Sachi Hegde

How AI Actually Changed My Working Week, 18 Months In

The impact of AI on Sachi Hegde’s work was subtle rather than immediate. Over time, it helped free up hours spent on research and routine tasks, allowing her to focus more on creativity, strategy and the human side of communication.

It’s a Tuesday night and I’m arguing with a machine about a revenue chart. Not the numbers which I know cold, but the story I’m telling with them and whether the version in my head is the honest one or just the convenient one. It pushes back. I push back harder. Somewhere in there the framing gets sharper, and then I close the laptop. And that’s the day.

Eighteen months ago, that would have been a conversation I had with myself or with a colleague if I were lucky enough to find one free at nine in the evening. I run my own communications practice, which means most days the room is empty. That is the part of my work AI has actually changed. Not the headline part, where it supposedly writes the strategy and renders me redundant, but the quiet, structural part. The empty room is less empty than it used to be.

It would be dishonest to tell this story without naming where it comes from. I’m writing as someone the technology has helped, and that is not the only story being lived right now. The same wave that arrives in my week as a few reclaimed hours is arriving in other people’s as a redundancy notice. The World Economic Forum found that around four in 10 employers expect to cut roles where AI can do the work and whole functions are being quietly thinned out. So, what follows is one vantage point, from a desk where the change has so far been a gift. I know that isn’t everyone’s desk and the gap between the two is going to be one of the defining questions of the next few years.

I had expected something more cinematic. Everyone did. Instead, the change arrived as a series of small subtractions. Less time staring before the first sentence comes, fewer days lost to the grind of benchmarking, a draft that begins at rough rather than nothing. None of it dramatic. All of it adding up. And I’m perhaps not alone in this. About four in 10 American workers now use these tools for their jobs, the Federal Reserve reported late last year, with the steepest growth in the most recent months. Most of us, I suspect, are doing some version of what I’m doing and saying very little about it.

What I use it for is less interesting than what it has quietly rearranged. Research that used to eat a full day now takes an afternoon, as long as I treat the first pass as a hypothesis and not an answer. It is confident in exactly the places it is wrong, which is its most human quality and its most dangerous one. Data, where it is genuinely strong, has become a sparring partner rather than a calculator. I hand it my framing of a set of numbers and let it go looking for what doesn’t hold up. First drafts of things, whether words or the occasional rough deck or social layout, start somewhere instead of nowhere. None of it ships as it arrives. It simply lowers the cost of beginning, dozens of times a week.

The hours it gives back are real, and larger than I would have guessed. Researchers at the London School of Economics put the average at around seven and a half hours a week for people who use AI at work, close to a full day reclaimed, worth roughly £14,000 a head a year in productivity. But the finding I keep returning to sat just beside it. In one survey, nearly half of workers said that recovered time was theirs to keep, not their employer’s. I’m firmly in that camp. The hours don’t vanish into a longer task list by some law of productivity. Some days they mean I finish early. Some days they mean I take on work I couldn’t have fit before. The time saved is the point, not a down payment on more of it!

What the machine cannot do is the part I’d defend if it came to that. It has no taste. It doesn’t know what a particular client actually needs, or what will land in a particular room, or which idea is brave and which is merely safe. It can’t hear the silence on a call or what it means. It doesn’t care, and it can’t feel for the person on the other end. And caring and empathy are the part that can’t be handed off. Microsoft’s latest workplace research keeps circling a point I find myself agreeing with. As machines take on more of the execution, the human part doesn’t contract, it expands. That has been exactly my experience. The slow, heavy work is quieter now, and the judgment it used to crowd out has more room than it did.

Which is what innovating the playbook has looked like for me. Not a new game. The same one, with a handful of plays rewritten so quietly that you only notice when you look up and find you no longer work the way you used to.

References

Allen, J. S. (2026, April 3). Monitoring AI adoption in the U.S. economy. FEDS Notes, Board of Governors of the Federal Reserve System. https://www.federalreserve.gov/econres/notes/feds-notes/monitoring-ai-adoption-in-the-u-s-economy-20260403.html

London School of Economics and Political Science. (2025, October). AI boosts productivity by the equivalent of one workday per week, new report finds. (Reporting on Bridging the Generational AI Gap, The Inclusion Initiative at LSE with Protiviti.) https://www.lse.ac.uk/news/ai-boosts-productivity-by-the-equivalent-of-one-workday-per-week-new-report-finds

Korn Ferry. (2026, January 30). AI saves time: Do you work more—or take Fridays off? Korn Ferry Insights. https://www.kornferry.com/insights/this-week-in-leadership/ai-saves-time-do-you-work-more-or-take-fridays-off 

Microsoft. (2026, May). 2026 Work Trend Index annual report: Agents, human agency, and the opportunity for every organization. Microsoft WorkLab. https://news.microsoft.com/annual-work-trend-index-2026/# 

World Economic Forum. (2025, January 8). Future of Jobs Report 2025 [Press release]. https://www.weforum.org/publications/the-future-of-jobs-report-2025/