This lesson is being recorded
Every video of this course is being uploaded to my YouTube channel. Subscribe so you do not miss it.
Bruno Ashford @brunoashfordSubscribeModule 12 · Lesson 12.1 · 12 min
Measuring Hours Saved
Compare your Time Map from the start of the course with how the same work runs today, task by task.
Use ← → to move. The marker highlights text and the pencil draws on the slide; navigation locks while you annotate. Undo removes the last mark, Esc exits.
1Pull out your original Time Map
Open the Time Map you made at the start of the course: each recurring task, who does it, how often and how long it takes. If you skipped it, rebuild it now from memory and calendar history, and mark those numbers as estimates. An honest estimate is better than no baseline.
2Time the task as it runs today
For each task an agent now handles, time three real runs from start to finish, including your review. The human minutes are what matter: prompting, waiting if you cannot do anything else, checking and fixing. Machine time running in the background while you do other work does not count as your time.
Here is my Time Map from before we started (pasted below). For each task that now runs through a skill, subagent or routine, build a table with: task, frequency per month, minutes per run before, human minutes per run now (including my review), and hours saved per month. Mark any number that is an estimate. Ask me for anything you cannot infer.3Convert minutes into monthly hours
Hours saved per month equals (minutes before minus minutes now) times runs per month, divided by 60. A 45 minute weekly report that now takes 10 minutes of review saves about 2.5 hours a month. Small numbers add up quickly across ten tasks.
4Put a value on each hour
Use the loaded hourly cost of whoever used to do the task: salary plus taxes and benefits, divided by working hours. For your own time, use what you would pay someone to replace you on that work, not your dream rate. Conservative numbers survive scrutiny.
5Decide where the hours went
Saved hours are only worth money if they go somewhere. Write one line per task: reinvested in sales calls, fewer overtime hours, one hire avoided, or simply absorbed. Absorbed hours are real but soft, and you should label them that way.
A dental clinic owner told me her AI setup "saves loads of time". I asked how many hours. She did not know. Three months later her partner asked the same question at the budget meeting, and the project was nearly cancelled because nobody could answer it.
Hours saved is the first number, and the easiest one to fake. The fix is boring and it works: measure the same task, the same way, before and after. You already built a Time Map in the early modules. Now we put a second column next to it.
Key points
- Measure the same task, the same way, before and after.
- Count human minutes, including review, not machine run time.
- Value hours at the loaded cost of the person who did the work.
- Hours saved only become money when you decide where they go.
Avoid
- Counting the time the agent runs in the background as time you spent.
- Leaving out review and correction time, which makes every agent look like magic.
- Valuing your own hour at an inflated rate to make the number bigger.
- Measuring once, on a good day, instead of across several real runs.
A roofing contractor with a small office team timed quote preparation at 70 minutes per quote, about 40 quotes a month. With a quoting skill pulling measurements and price lists, the estimator now spends 20 minutes reviewing and adjusting. That is 50 minutes times 40 quotes, roughly 33 hours a month. The owner reinvested the time: the estimator now calls every quote back within 24 hours, which is where the real money showed up.