The Two Kinds of Blank

My business plan went through eleven versions today. An AI wrote every draft; my job was to judge them. By evening the plan had gone from twenty-one pages to nine, and I thought the document was in good shape.

Then I asked the AI to list everything in the plan that was still undecided. It found seven items. "Under consideration." "To be negotiated case by case." "Method not yet designed." "To be determined at contract time." Seven blanks, sitting inside the cleanest document I had ever produced. A page that has declined to decide still looks finished. That is what makes these phrases dangerous: they read as diligence and function as delay.

Looking at the seven, a distinction became obvious. Blanks come in two kinds, and the two kinds deserve opposite treatment.

Fact-gaps and decision-gaps

A fact-gap is a blank that only the world can fill. What the data vendor actually charges. How many hours a client actually consumes. What a customer will actually pay. Leaving a fact-gap open is honest — it is the correct notation for "not measured yet." Filling it with an invented number is the failure. A fact-gap earns its place by naming its own resolution: who measures it, when, with what instrument.

A decision-gap is a blank I could fill right now by deciding. "Under consideration" — considered by whom? On what schedule? In my plan, the answer was: by nobody, on no schedule. The phrase was procrastination written in the vocabulary of process. All seven of my blanks were this kind. Each one had a decision available at the moment of writing, and each had dodged it.

Why the old habit no longer pays

Deferring decisions used to be economically rational. Revising a plan cost a meeting and a week of everyone's patience, so committing early meant expensive rework, and "to be discussed" kept the rework off the books.

That justification has died in my company, because an AI now rewrites the entire document in minutes. A wrong provisional decision costs one revision cycle — nearly nothing. An open blank still costs what it always cost: an agenda item, a meeting, a slice of a partner's attention. The prices have flipped. Deferral used to be the cheap option; now it is the expensive one. Most planning habits were formed under the old prices.

There is a second reason to decide provisionally, independent of cost. A provisional decision produces information: a number on the page can be struck out, argued with, flinched at. A blank produces nothing, because nobody can disagree with a blank. Write the tentative answer and the review meeting changes shape by itself — from an open discussion that schedules another discussion, into a row of concrete answers waiting to be approved or corrected.

What the plan looks like now

The seven blanks became seven provisional decisions. Each one carries a label marking it as tentative, and a trigger — the named event that reopens it. The referral terms that were "negotiated case by case" now have a default. The pricing floor has an anchor number. The data-handover clause that was "to be determined at contract time" is determined, four months before any contract exists, precisely so that a real contract has something to disagree with.

Fact-gaps stayed open. They are supposed to.

Did converting the other seven take nerve? Less than it should have. Deciding felt easy because I knew the machine could unwind any of these answers in minutes. Whether that is courage with better tools, or cowardice that got cheaper, I have not decided — and I notice that one is a fact-gap about myself, still waiting on measurement.