The Intermediate Mindset: Directing vs. Asking

Your Prompts Are Polite. That's the Problem.

Most people approach AI the way they approach a search engine: they phrase their need as a question, hit enter, and hope the result is close enough to work with. "Can you help me write an email?" "What's a good way to structure a report?" "Do you have any ideas for this?" These prompts are polite. They're also vague — and vagueness is the single biggest reason AI output disappoints people.

The shift from beginner to intermediate isn't about learning new techniques. It's about making one fundamental change to how you think about the interaction. You are not submitting a query to a database. You are briefing a capable collaborator. And capable collaborators need direction, not questions.

The Difference Between Asking and Directing

Asking puts the AI in charge of interpreting what you need. Directing puts you in charge of defining it. When you ask, you get whatever the model's default is — which is trained to be broadly acceptable, not specifically useful for your situation. When you direct, you get output shaped by your actual requirements.

The director's mindset means treating every prompt as a brief. A brief tells someone who they are for this task, what they're producing, who it's for, what constraints apply, and what good looks like. You don't need to use all five every time. But thinking in those terms will change every prompt you write.

Before and After: The Same Task, Two Mindsets

Here is a real-world example — drafting a follow-up email after a sales call — in both modes:

Asking: "Help me write a follow-up email after a sales call."

Directing: "Write a follow-up email to a marketing director at a mid-size SaaS company. We had a 30-minute discovery call today. She's interested in our analytics platform but concerned about the implementation timeline. Keep it under 150 words, warm but professional, and end with a specific next step — scheduling a technical call for next week."

The second prompt takes maybe 45 seconds longer to write. The output it produces is often ready to send. The first prompt produces something you have to rewrite from scratch.

Four Things Directors Specify

You do not need to over-engineer every prompt. But these four elements — when included — produce measurably better results:

• Role or context: Who is the AI acting as, or what context does it need to know? "You are a copywriter specializing in B2B SaaS" produces different output than an unanchored prompt.
• Task: What specific thing should the output be? Not "help me with a report" but "write the executive summary for a quarterly performance report."
• Constraints: Length, tone, format, what to include, what to avoid. Constraints narrow the search space and make the AI's job easier, which makes your output better.
• Output specification: What does a good response look like? A numbered list? Three options? A single paragraph? Prose or bullet points? If you do not say, the model guesses.

When Asking Is Fine

Not every interaction deserves a full brief. If you are exploring an unfamiliar topic, brainstorming without a fixed destination, or asking a genuine factual question, a shorter, open-ended prompt is appropriate. The director's mindset does not mean writing a novel for every message — it means being intentional about how much direction the task actually requires.

A rough heuristic: the higher the stakes and the more specific your requirements, the more direction you should provide. Asking "what are some approaches to performance reviews?" is fine if you're genuinely exploring. Asking for a specific template you plan to actually use with your team requires a brief.

The One-Sentence Test

Before you hit enter on any prompt, ask: could a human contractor misinterpret this and still technically satisfy it? If yes, add one more sentence of direction. The goal is not to write more — it is to write precisely enough that there is only one reasonable interpretation of what you want.

This test catches most vague prompts before they produce vague output. "Write me a bio" fails it immediately — how long? what tone? what purpose? for whom? "Write a two-sentence professional bio for my LinkedIn summary" passes it.

Your First Practice

Find a prompt you've used recently that produced output you were not fully satisfied with. Look at it now and ask: was it asking or directing? Add role context, a sharper task statement, at least one constraint, and an output specification. Run it again. The difference in output quality will demonstrate the principle better than any explanation.

This is the foundation the rest of this track builds on. Every technique in the following lessons is an application of the same core idea: you get better results when you are specific about what you want.