Counting the Hours You Got Back

The last lesson of the course is an accounting lesson, because a system that cannot show its savings will eventually be dismantled by a busy week. The feeling of "the AI stuff helps" does not survive a bad month; a number does. Fortunately, you have been collecting the inputs all along without noticing: the week audit was the before, and the week sheet is the after.

The baseline you already have

Your Chapter 1 audit scored each task on time cost, which means it recorded how long each task took you by hand: quoting at 45 minutes a round, inquiry replies at 15, the invoice chase at 20 of avoidance-flavored effort. That sheet is your baseline. Dig it out and keep it; owners who skip this step end up guessing their before, and guessed baselines get argued with, mostly by yourself, mostly on bad days.

The monthly count

On the first Friday of the month, inside the extended numbers session, add ten minutes:

Current minutes per run comes from the week sheet's minutes column, and it includes your checkpoint time; the count is honest or it is useless. Subtract the system's own overhead too: the Brain refresh, runbook maintenance, the count itself, perhaps an hour a month all-in. For the HVAC company, the honest number landed around 33 hours a month, most of it the office manager's, seven-ish of it the owner's, including the extinct Sunday-night quoting session.

Write the number in the week sheet. Twelve of those entries is the difference between "we use AI" and knowing what the system is worth.

What the count is for

Not self-congratulation. The number drives three real decisions:

• The keep-or-kill review. Any workflow returning less than an hour a month gets a question: fix it or retire it? Retirement is a healthy outcome; a workflow nobody runs is system rot, and rot spreads doubt to the workflows that do earn their place.
• The what-next decision. Returned hours are capacity. The HVAC owner's seven hours became sales calls to contractors, the segment the Friday numbers kept flagging as the growth lane. Capacity spent on purpose is the system compounding; capacity absorbed silently into "busy" is the scatter tax winning after all.
• The case for the next build. When you propose the next workflow, or the office manager does, and they will, the count is what makes it a business decision instead of an enthusiasm. "Reminders return 12 hours a month; review-response would likely return 6" is a sentence that survives scrutiny.

Where the system goes from here

You now hold the full loop: audit, Brain, workflows, cadence, lanes, handover, count. The loop repeats. New repeatables surface in the week sheet's notes; the toolbox has more than 100 free prompts and skills to build from; each new workflow enters through the same audit scoring, gets its runbook, earns its lane, and joins the count.

And the courses behind you are now reference material: the fundamentals when a skill needs a sharper prompt, the skills course when you build a new worker, the marketing course when the Monday batch needs a new channel. You did not take four courses; you assembled one system in four installments.

Try it now

Run your first monthly count this Friday, even if the system is two workflows old and the number is small. Small and real beats large and vague; it is the baseline your sixth month will be measured against. Then spend one line of the week sheet on the question the whole course has been building toward: what will you do, on purpose, with the first hours the system hands back?