A Free Chapter, Unedited

This is one real chapter from the playbook, reproduced exactly as it appears in the book. Nothing has been trimmed or dressed up for the web. We picked it on purpose: Chapter 4 introduces the CLEAR framework, the thirty-second checklist our readers run before sending any prompt that matters. It stands on its own, and by the end of it you will have something you can use this afternoon.

Read it. If it earns a place in how you work, the rest of the book delivers the full method, a library of ready-to-use templates, and a thirty-day plan you can run on your own.

Chapter 4

The CLEAR Framework

A pre-flight checklist for prompts that matter. Run it in under thirty seconds.

The six-layer anatomy in Chapter 3 is the architecture of a strong prompt. CLEAR is the checklist you run through your head before you press send. They are not in tension; they are complementary. Anatomy is what you build. CLEAR is how you inspect it.

I introduced CLEAR after observing that executives who knew the six layers cold still produced inconsistent output, because in the heat of a workday they would skip steps without noticing. CLEAR was the result: a five-letter check, deliberately memorable, that could be run in twenty to thirty seconds and would catch most of the common omissions.

The five-letter check you run before sending a prompt that matters.

C: Context

Have I told the AI what business situation it is operating in? Have I given it the numbers, the stage, the audience, the recent events that matter? Most AI output is generic because the executive forgot to make it specific. Five sentences of context, supplied in the right places, are worth ten paragraphs of editing later.

L: Language

Whose voice should this output be in? Is it a board memo or a sales email? Is the reader skeptical or sympathetic? Should the tone be measured or urgent? "Language" here is shorthand for voice, register, vocabulary, and rhetorical posture. An AI told nothing about language defaults to bland professional prose. That default is rarely what an executive wants.

E: Examples

Have I shown the AI what a good answer looks like? This is the highest-leverage CLEAR check. If you have a previous version of the same artifact (a memo, a slide outline, an investor update), paste it directly into the prompt with the instruction to match its structure and tone. The improvement is immediate and dramatic.

One example is worth a paragraph of instruction. Modern AI models imitate concrete patterns far better than they follow abstract descriptions. If you find yourself writing a long paragraph describing the tone you want, stop. Find a real document with that tone, paste it in, and say "match this." The result will be better and the prompt will be shorter.

A: Ask

What exactly am I asking the AI to do? "Ask" is the task layer rephrased as a diagnostic question. The test: can the ask be summarized in a single sentence beginning with a verb? "Draft a memo." "Critique this argument." "Identify the three weakest assumptions." If the ask is buried in two paragraphs of preamble, find the verb and surface it.

R: Refine

What is my plan to iterate on the output? CLEAR's last letter is the one that distinguishes professional users from casual ones. Casual users accept the first answer. Professionals treat the first answer as a draft and have a specific plan for the second pass.

Useful refine moves include:

• "Critique your own draft as if you were the harshest skeptic in the room. List the three weakest points."
• "Rewrite this in half the length without losing any substance."
• "Now produce a version optimized for a CFO reader. Then a version for a board chair."
• "What did I forget to ask you to consider?" (This single move surfaces remarkably useful gaps.)

A worked example

Walk through a real CLEAR application. The task: I need to draft a one-page recommendation for my CEO on whether to extend an enterprise customer's payment terms.

Final CLEAR-checked prompt

CONTEXT: We are a $40M ARR B2B SaaS company. The customer is
Northwind Insurance, $1.8M ACV, with us for three years. They
have asked to move from Net-30 to Net-90 terms citing internal
cash-flow constraints. They are at 110% NRR.

LANGUAGE: Measured executive memo. No marketing tone. Direct.
My CEO reads about a thousand of these a year; she values clear
recommendations over balanced essays.

EXAMPLE: Match the structure and brevity of this prior memo:
[paste prior commercial decision memo].

ASK: Draft a one-page recommendation. Recommendation must be
explicit (extend / decline / counter), with the three strongest
reasons and the two most material risks.

REFINE: After your draft, list the three weakest points in your
own argument. Then suggest one piece of information that, if I
had it, would meaningfully change the recommendation.

The prompt is roughly 180 words. It will produce a useable draft in one shot. The total time from blank page to a memo I can hand my CEO is under eight minutes. The equivalent memo, written from scratch, takes me forty-five.

Actions for this week

1. Print or save the CLEAR diagram somewhere you can see it. Until the framework is reflex, you need the visual. After about three weeks, you will not.
2. Pick one high-stakes prompt this week and run CLEAR on it before sending. Notice what you almost forgot. Notice how the output changes.
3. Add "What did I forget to ask?" to the end of every important prompt. It costs nothing. It surfaces gaps surprisingly often.

What the rest of the book covers

CLEAR is one checklist inside a larger system. The full playbook builds the six-layer anatomy that CLEAR inspects, walks through real prompts reworked before and after, and maps prompt patterns to the specific functions on an executive's desk. It closes with a personal prompt library, the eight pitfalls that quietly kill ROI, and a thirty-day sprint you can execute alone before involving a single other person.