You budget money to the penny. And hours not at all.

Every company I work with can tell me its budget to the pound. Very few can tell me how their hours are spent, and none of them have ever divided them up on purpose.

Which is strange, because the hours are the company. The money buys them.

The block

A hundred people, working a normal year, is roughly 184,000 hours.

That number is the company. Not the org chart, which shows reporting lines. Not the strategy deck, which shows intentions. The hours are what actually gets spent, and every plan you make is a claim on them.

The first cut

Most of those hours go on keeping the promises you have already made. Serving customers, running the platform, closing the books, fixing what breaks. Call it seventy percent, and do not call it waste, because it is not. It is the business working.

Which leaves about thirty percent to actually change the company.

Thirty percent of a hundred people is thirty person-years. That is your entire strategy budget for the year. Everything you want to become gets built out of thirty person-years.

Most leaders have never seen that number, and it lands differently than a budget does.

Now divide it

Thirty person-years across four quarters is roughly 7,500 hours a quarter for strategic work.

Three goals, and each gets 2,500 hours. That is real weight behind each one.

Ten goals, and each gets 750. A third of one person, per goal, per quarter.

Focus is not a philosophy. It is division. And this is why the advice to have fewer goals, which every leader has heard and most find slightly annoying, turns out to be arithmetic rather than opinion.

The second cut, which nobody makes

Here is the part that does not appear in any capacity model I have seen.

The hours are fixed. What each hour is worth is not.

Every one of those 184,000 hours belongs to a person, with their own goals, their own way of working, their own good and bad weeks. An hour from someone who understands why the work matters, and has the room to think, and is not carrying three other priorities, is worth several from someone who is depleted, unclear, or quietly certain the plan will change again.

Same payroll. Wildly different output.

Which means there are two entirely separate levers, and most companies only pull one. You can change where the hours go, which is strategy and goals. Or you can change what each hour is worth, which is everything about the environment those people work in. The second one is usually untouched, and it is not soft, it is arithmetic with a multiplier in it.

And AI, which is not doing what people think

The pitch is that AI gives you time back. Sometimes it does. Often what it does is remove a two-hour task and add a thirty-minute task plus ninety minutes of checking whether the answer is right.

So here is the test I would apply to every AI decision, and it is a single question:

Did it move time from the mechanical to the human, or did it just make the machine work louder?

The genuine prize is real, and it is worth stating in the same units as everything else. Every percentage point of routine work you genuinely hand to a machine is about 1,840 hours. Roughly one person-year, minted without hiring anyone.

Pointed carelessly, the same technology does the opposite: more output to check, more noise to read, thicker routine work with a faster clock.

Three questions worth an hour of the 184,000

Where do your hours actually go? Not roughly. If nobody can answer within ten percent, that is the finding.

What is each one worth? And what in how you run is currently reducing it?

And is your AI moving the line, or thickening it?

The honest footnote

Your numbers will be different from mine. A hundred people, forty-six working weeks, forty hours: adjust it for your business and the shape will hold even as the figures move. That is rather the point. This is not a benchmark, it is a calculation you can do this afternoon, and most companies never have.

The Business OS decides where the hours go. The Human OS decides what each hour is worth.