Say you're looking for an AI tool to help you work on a proposal tomorrow.
You read comparison articles and try a feature one of them mentions. Another tool catches your eye, so you watch a video on how to use it. You've gathered plenty of new information, but the proposal is still blank.
Perhaps a little more research will turn up a better approach. Once you start thinking that way, it's hard to settle on one of the options you already have.
In that situation, "I'll get started once I know more" makes it hard to put a limit on your research. Before you begin, try narrowing the scope to what you need for today's work.
If you put "learn about AI" on your calendar, it's hard to know what would count as enough. Every new announcement gives you something else to read.
For a proposal, today's task could be to make one outline from the materials you have, then check that it covers what was requested. You can then choose what to look up based on what you need to produce that outline.
If you want to learn a new feature, try one that you'll use in your work this week. If you're revising an output, decide what you'll check before you submit it. Once you know what "make it better" means, it's easier to decide how much to change.
But if you stop at producing something, there's still work left to do.
When I build my company's website with AI, I use the pages myself, repeat the same actions, and compare the files served by the site. Checking that a button is visible and checking that it takes you to the right place are two separate tasks.
The more I ask AI to produce, the more I have to review. For me, it works better to keep each task small enough to include the necessary checks, rather than asking AI to produce more than I can review in one go.
With the proposal, too, I'd want to leave time after the outline is ready to read it against the request. When deciding "this is as far as I'll go today," include that review.
Even with a clear scope, you'll come across information that interests you along the way.
If another tool looks promising, save the link with a short note on what you might use it for, then return to the task at hand. You can open that note when you know where your current approach falls short.
You don't have to put everything you've saved on tomorrow's to-do list, either. If you don't need it, you can leave it alone.
If unfinished work is on your mind when you're about to step away from the screen, write one line about what to do next.
Next: check the outline's headings against the request. Compare other tools once I know where the current outline falls short.
That tells you where to pick things up more clearly than "continue working on the proposal." Even if the work is unfinished, you can decide what you'll leave for next time before you close it.
On days when a deadline is close or an urgent revision comes in, you may not be able to finish as planned. I don't want that to become another reason to blame yourself for failing to stop.
Once you reach your stopping point, you can use the time that's left for another task. You can also just close things down.
I don't think you need to judge time spent resting by how much work you get done afterward. When today's work is finished, you can stop there, even if there are still articles you haven't read.
If you'd like to take a moment to reflect on how you're doing, there's BlueFlow's Daily Flow Check. It's free, and you don't need to register. Four questions take about 30 seconds. Based on your answers, it suggests one small action you could try.