AI Prompt Engineering · Lesson

Adding Concrete Details

Numbers, names, formats, and examples that anchor AI output.

Concrete Details Are Anchors

Concrete details are specific, measurable, verifiable pieces of information. They anchor the model's output to your actual situation instead of a generic interpretation.

Without anchors, the model generates for the imaginary average user. With anchors, it generates for you.

Word Count as a Concrete Constraint

Word count is one of the simplest and most effective concrete details you can add. Compare:

Vague: 'Write a short description'
Concrete: 'Write a 75-word description'

The model will count tokens and aim for your target. If it misses slightly, you can ask for an exact recount. Word count eliminates entire categories of rewriting.

import anthropic

client = anthropic.Anthropic(api_key='sk-ant-your-key-here')

response = client.messages.create(
    model='claude-opus-4-5',
    max_tokens=200,
    messages=[{
        'role': 'user',
        'content': (
            'Write a product description for a noise-cancelling travel pillow in exactly 60 words. '
            'After the description, output the word count in parentheses on a new line.'
        )
    }]
)
print(response.content[0].text)

All lessons in this course

← Back to AI Prompt Engineering