A draft is not a decision

In a busy week the work arrives already sounding finished. A note to a buyer. A recommended route. A paragraph under an opportunity that reads like the forecast has been settled.

Fluency is not a commitment. A decision is a rule the room can read, with odds you are willing to repeat on the forecast call. Everything else is a draft.

The agent drafts. The rules decide.

What shows up looking finished

Lead routing suggestions. A recap of the last meeting. A proposed change to the number you will defend. These are current: work that flowed from somewhere. If you cannot point to the row that produced them, they flowed from taste.

Taste is allowed in a draft. It is not allowed to skip the switch. The switch is the table. The table says when a route may advance, when it holds for the pipeline review, and what odds you assigned before the result.

What commitment requires

• A written rule, so two people would make the same call from the same facts.
• Odds, stated before the outcome, so the forecast can be scored later.
• A draft of the work, clearly still a draft.
• Written gates, plus a named person's recorded approval, if the work would reach a customer or move money.
• A receipt, so Thursday does not depend on whoever remembers the meeting.

What you can replay

The test is dull, which is the point. Pick a decision from last month. Can you show the signals that were allowed to count, the row that matched, the odds you stated, the person who released the work, and what happened after? If yes, you can coach the rule. If no, you can only coach the story.

A model is useful on the front of that list. It can gather the signals and shape the draft. The row, the odds, and the release stay with the owner. That is the whole method, said in one pass.