Day 17: Build a Sales Playbook Your Whole Team Can Use

The Concept

Most sales playbooks fail before they're ever read. They're built the wrong way: someone compiles a deck of slides, a CRM process diagram, and a product overview, wraps it in a binder called a "playbook," and gives it to a new hire. The new hire reads it for half a day, learns nothing that helps them run a real call, and then learns the job the same way everyone else did — through six months of mistakes and pattern-matching from colleagues who have time to help them.

A playbook that actually works is not a process document. It's a judgment guide. The difference is significant. A process document tells you what steps to follow. A judgment guide tells you how to think — what signals matter, what the decision looks like at each fork in the road, what the best reps do differently than average ones. You can't write a judgment guide by documenting your CRM stages. You write it by extracting the real knowledge from people who win deals consistently.

Why Tribal Knowledge Stays Tribal

Every sales team has one or two people who just close. They're hard to explain. Their pipeline looks different from everyone else's. Their discovery calls go somewhere. Their proposals don't stall in "pending." When you watch them operate, it doesn't look like they're following a playbook — it looks like they're just good at reading people. But that's not what's happening. They've built a mental model of what works, refined over hundreds of calls, and it lives in their head rather than on paper.

AI can help extract that model. When you force yourself to articulate the specific things that win deals for you — not "I build good rapport" but "I always get a CFO intro before the final proposal" — and then run that through a structured prompt, the output becomes a skeleton of the judgment you've accumulated. New reps can read it, understand the reasoning, and start further down the curve. It doesn't replace experience. But it dramatically shortens the runway.

Short Enough to Actually Read

The other reason playbooks don't work is length. A 60-page document is a reference library, not a guide. No rep opens it before a call to check what question to ask next. The target for a working playbook is one or two pages — dense enough to contain real guidance, short enough that someone reads it before a big meeting. This is a hard constraint to hold, because the instinct is to add more. More edge cases, more product detail, more process steps. Resist it. A playbook that gets read beats a comprehensive one that doesn't.

AI helps here too. It naturally produces structured, scannable output. The discipline of the format — ICP, discovery questions, objections, proposal structure, ramp plan — forces you to be specific and concise. The real work is the editing pass you do after: cutting anything that would have been obvious to anyone anyway, and sharpening the pieces that only you would know.