Suppose a tool cuts a recurring task from ten minutes to five. That sounds like five minutes saved. If every complete run really saves five minutes, the hour of setup takes at least twelve runs to recover. Checking its output, fixing mistakes, moving old work, and keeping the tool working can push that point further away. Those numbers are an illustration, not a result from my own test.
Here is a real example from this newsletter project. I asked Codex, an AI coding assistant, to help revise the website copy and subscription email flow. Its changes still needed a site build, automated tests, a delivered seed email to inspect, and a live unsubscribe test before I could trust them. I do not have a reliable manual comparison for that work, so I cannot claim a number of hours saved.
Before switching, pick one small task you actually repeat. Time one run with your current method. Then time the setup and one complete run with the new tool, including checking and corrections. Repeat the task at least once more with each method, if practical, before making a permanent switch; the first run can be distorted by learning or an unusual task. Write down anything you would have to maintain later. You do not need a perfect study to spot a tool that creates more work than it removes.
Time is only one part of the decision. A tool may still be worth using because it makes the work easier to share, more accessible, or less error-prone. It may be a poor fit if it needs data you are unwilling to give it or makes a simple task harder to understand.
Your action: Choose one tool you are considering. Write down the task, how often you do it, and the time for a complete run with each method. Include setup and rework. Decide what improvement would make a switch worthwhile for you.
The 15-Minute Time & Attention Audit can help you identify a repeat drain worth testing. You can read it without subscribing.
A question for you: Which tool has made a simple job harder for you? If you receive the Field Note by email, reply with the task and the part that took more work than expected.