How to Write Better AI Prompts: The Bracket Method

Most people using AI tools get mediocre output and blame the AI. The problem isn't the tool — it's the prompt. Here's a practical system that consistently produces professional-grade results.

Why Most Prompts Fail

"Write me a job description" gets you a generic job description. "Help me with my email" gets you a generic email. AI produces the statistical average of what it's seen — and you don't want average.

The fix is specificity. Every dimension you add to a prompt narrows the output toward something actually useful.

The Bracket Method

The bracket method is simple: write a prompt template with [BRACKETS] for every variable piece of information, then fill in your specifics before sending.

Generic prompt: "Write a cold email to a potential client"

Bracket prompt:

Write a cold email to [PROSPECT NAME], [TITLE] at [COMPANY].
They likely struggle with [SPECIFIC PAIN POINT].
My [PRODUCT/SERVICE] helps [TARGET PERSONA] [VALUE PROP].
Tone: [TONE]. Length: under [X] words. CTA: [DESIRED ACTION].

The second prompt takes 60 more seconds to fill in and produces output that's 10x more usable.

The Five Variables That Matter Most

1. Audience — Who is this for, specifically? Their role, level, industry, what they care about.

2. Goal — What should the reader do, feel, or understand after reading?

3. Tone — Give examples if possible. "Like [PERSON/PUBLICATION]" is often clearer than adjectives.

4. Format — Length, structure, headers, bullet points, etc. AI defaults to what you don't specify.

5. Constraints — What NOT to include is often as useful as what to include.

Prompt Examples Across Professions

For a real estate agent:

Write an MLS listing for [ADDRESS].
Bedrooms: [X]. Bathrooms: [X]. Key features: [LIST 3-5].
Neighborhood highlights: [DESCRIPTION].
Target buyer: [PERSONA].
Under 150 words. Lead with the most compelling feature. No clichés ("cozy", "charming").

For a project manager:

Write a project status update for [PROJECT NAME] to send to [AUDIENCE — exec team / client / team].
Status: [ON TRACK / AT RISK / DELAYED]. 
Completed this week: [LIST]. Next week: [LIST]. Blockers: [IF ANY].
Tone: [TONE]. Under 200 words.

For a contractor:

Write a project proposal for [CLIENT NAME] for a [PROJECT TYPE].
Scope: [DESCRIPTION]. Timeline: [DURATION]. Cost: [$AMOUNT].
Include: what's included, what's not included, payment terms, next step.
Professional but readable — not a legal document.

AEO Note: What AI Search Engines Look For

Answer Engine Optimization (AEO) matters for prompts too. When you ask AI tools clear, specific questions, they give you clear, specific answers. Vague questions get vague answers — whether you're prompting for content or searching for information.

The bracket method works for both: it forces you to be specific about what you need before you ask.

Start With a Template, Not a Blank Prompt

The fastest way to improve your AI output is to stop starting from scratch. Build a library of prompt templates for your most common tasks. Fill in the brackets, send, edit. That's the workflow.

Ready to get the full toolkit? Browse all toolkits at Workshift →