The Unexpected Lessons from Letting AI Plan My Workdays
By eleven fifteen on the second day, the morning’s writing was done. Not the kind of “done-for-now” that I would come back to when I felt braver, but genuinely finished. The experiment’s schedule, designed with ChatGPT’s guidance, instructed me to break, eat, walk, and avoid returning to my desk for any producing work. That pause, longer than I anticipated, marked a shift in how the rest of the week unfolded and how I understood my own working habits.
In preparation, I had shared quite a bit of context with ChatGPT: I’m a freelance writer with control over my schedule, and I realistically have about three solid hours of deep writing energy each day. The rest of my time is consumed by ideation, research, editing, administration, and other supporting tasks. I asked the AI to plan my workdays down to the minute for a week, urging it to be detailed and precise. What I received back was unexpectedly specific.
The schedule featured two deep-writing blocks each morning—from eight to nine thirty, and then nine forty-five to eleven fifteen—with a short walk scheduled between them. Administrative tasks were batched once, mid-afternoon, but never before the first writing block. Walks were treated as meaningful transitions rather than wasted moments, a practice I had already been leaning toward. Importantly, there was a hard stop at the end of the day.
Each writing block began with a sentence written beforehand, such as “By the end of this block, I will have…,” and ended with a “Next move:…” breadcrumb at the top of the document. This simple yet effective strategy smoothed the transition between work sessions, reducing the friction of getting started. The underlying logic was clear: protect the best three hours for deep work, batch shallow tasks, and recognize that not all work carries equal value.
Trying the AI-Designed Schedule
I did my best to follow the plan. Some days were closer to the ideal than others. On Wednesday, a message pulled me away for an hour, breaking the afternoon’s rhythm. Friday morning’s first writing block started late because I checked email while grabbing coffee. I’m not someone who can run a schedule with Swiss-watch precision, and I knew that from the start.
Surprisingly, the output was anticlimactic. Roughly the same volume as a typical week, perhaps with slightly cleaner drafts thanks to genuinely uninterrupted mornings. If the experiment had ended there, I might have written a paragraph about it and forgotten the whole thing by the weekend. But the real shock came from somewhere else.
Confronting the Reality of “Pseudo-Productivity”
For years, I had been working roughly the same hours as I did in Irish finance—possibly more—but on my own terms. About three of those hours were deep, producing work, while the rest was editing, sourcing, admin, and ideation. That was nothing new. What the AI-driven schedule forced me to confront was my unvarnished relationship with the rest of the day—the afternoon and early evening hours when the screen was on, the browser tab said “research,” but the actual activity was professional restlessness.
Half of what I had been calling research was just browsing. The feeling of “still working” didn’t equal productive work. Cal Newport, a bestselling author and productivity expert, calls this “pseudo-productivity.” He describes it as a management philosophy that “leverages visible activity as a crude proxy for useful effort,” primarily in office settings. In my case, working solo in cafés, it was a freelance iteration of the same phenomenon—the visible effort was mostly for my own benefit, a performance to convince myself I had earned the day’s workload.
Gloria Mark’s research at UC Irvine, which I’ve carried with me for years, offers a concrete data point: it takes an average of 23 minutes and 15 seconds to regain focus after an interruption. Armed with that knowledge, I had already cultivated habits to protect my focus—phone in another room, closed tabs, rain sounds on YouTube instead of music, and time blocking. These were doing their job, but the AI schedule added a new layer of awareness: not all the hours I protected were pulling their weight. Some were “performance hours” masquerading as real work.
The Bigger Truth About Freelance Work and Identity
This isn’t exactly novel. The concept that work expands to fill the time available—often called Parkinson’s Law—has been around since the 1950s. What was new to me was experiencing it through a schedule I hadn’t chosen. When an external plan said “stop,” I stopped. And I had to sit with what was left over. The afternoons I usually filled with “research” transformed into walks, errands, and genuinely slower evenings. The world didn’t end. My output didn’t diminish. My drafts weren’t worse. The hours I had been keeping on were, it turned out, not essential.
Why had I been holding on to those hours? The financial bargain of freelance work partly explains it—more hours feel like a hedge against instability. But the bigger driver is identity. Stopping early, even when the producing work is complete, feels like failure. Without an office to leave or a colleague signaling the end of the day, I kept the screen on. This is likely a common experience among remote and freelance workers, who often lack external cues to define “work time.”
What Stayed After the Experiment
I didn’t keep the AI-designed schedule after the week ended. I reverted to my own rhythm: a concentrated morning session, a walk, switching cafés, and an afternoon mixing edits and lighter tasks. But I adopted one practice from the AI plan that has stuck—the “Next move:…” breadcrumb at the top of my documents.
This simple sentence has become a cue I didn’t have before. When I can write it honestly, I know the producing work for the day is done, and I can turn off the screen. When I can’t, I realize I’ve been performing for an hour already. It’s not a flawless system, but it’s a door I can close—a small but meaningful boundary in the blurry landscape of freelance work.
For freelancers, remote workers, and anyone struggling with self-managed schedules, this experiment offers a reminder: the hardest work may be not the tasks themselves, but confronting the invisible scripts we tell ourselves about what “working” should look like.
Source: Here
