Specificity Wins

The single biggest improvement you can make to any prompt is to be more specific. Vague prompts produce generic output. Specific prompts produce useful output.

Here are three before-and-after examples that show exactly what this means in practice.

Example 1: Writing

Vague: "Write an email to a client about a delay"

Specific: "Write a professional email to a client informing them that their software project will be delayed by 2 weeks due to unexpected technical complexity. Tone: apologetic but confident. Include: reason for delay (technical, not blame), revised timeline, one concrete step we're taking to prevent further delays. Keep it under 150 words."

The specific version tells the AI the tone, the structure, the word count, and what to include. The vague version gives it nothing to work with.

Example 2: Analysis

Vague: "Analyse this data"

Specific: "Analyse this monthly sales data for our three product lines. Identify: which product is growing fastest, which is declining, and any seasonal patterns. Format as 3 bullet points followed by a one-sentence recommendation. Focus on trends, not raw numbers."

Now the AI knows what to look for, what to ignore, and how to present findings.

Example 3: Code

Vague: "Write a function to sort a list"

Specific: "Write a Python function that takes a list of dictionaries and sorts them by a specified key. The key should be a parameter. Handle the case where the key doesn't exist in some dictionaries (put those items last). Include a docstring and one usage example."

The specific version produces production-ready code. The vague version produces a basic example you can't use directly.

The Specificity Checklist

Before sending any prompt, ask:

• Have I specified the audience?
• Have I specified the tone or style?
• Have I specified the length or format?
• Have I specified what to include or exclude?
• Have I given enough context?

You don't need all five every time, but running through this list catches most vagueness.

Try It

Take the last prompt you wrote. Add at least two specifics you left out. Run both versions and compare the output.