プロフェッショナルな文章スタイル

思考を構造化する文章スタイル

ジャーナリスティックスタイル：逆ピラミッド

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 in journalistic inverted-pyramid style. '
            'Lead immediately with the most important fact. '
            'No setup, no preamble. Short sentences, short paragraphs. '
            'Max 100 words.\n\n'
            'Announce: Anthropic has released Claude 4, achieving a new benchmark score of 94% '
            'on a standard reasoning test, surpassing previous models by 8 percentage points. '
            'Available via API today. Pricing unchanged.'
        )
    }]
)
print(response.content[0].text)

ビジネスコンテンツにおけるジャーナリスティックスタイル

import openai

client = openai.OpenAI(api_key='sk-your-key-here')

response = client.chat.completions.create(
    model='gpt-4o',
    messages=[{
        'role': 'user',
        'content': (
            'Write an internal engineering memo using journalistic style. '
            'Most critical information in sentence 1. Max 3 short paragraphs.\n\n'
            'Facts: We will have a 30-minute planned downtime on Saturday, March 15 at 2 AM UTC. '
            'Cause: database migration to PostgreSQL 16. '
            'Impact: all API endpoints unavailable. '
            'Action needed: engineers should not deploy during this window. '
            'Status page will show updates.'
        )
    }]
)
print(response.choices[0].message.content)

説得的なスタイル：問題解決型

import anthropic

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

response = client.messages.create(
    model='claude-opus-4-5',
    max_tokens=250,
    messages=[{
        'role': 'user',
        'content': (
            'Write landing page copy using the Problem-Agitate-Solution (PAS) persuasive framework.\n'
            'Structure:\n'
            '- Paragraph 1 (Problem): name the pain clearly\n'
            '- Paragraph 2 (Agitate): make the reader feel how serious this problem is\n'
            '- Paragraph 3 (Solution): introduce the product as the answer\n'
            'Product: A tool that automatically generates commit messages from git diffs.\n'
            'Audience: developers who write bad or no commit messages under time pressure.\n'
            'Max 150 words total.'
        )
    }]
)
print(response.content[0].text)

説得的なスタイル：AIDAフレームワーク

import openai

client = openai.OpenAI(api_key='sk-your-key-here')

response = client.chat.completions.create(
    model='gpt-4o',
    messages=[{
        'role': 'user',
        'content': (
            'Write a cold email using the AIDA framework. '
            'Label each section: [ATTENTION], [INTEREST], [DESIRE], [ACTION].\n'
            'Product: AI-powered code review tool that catches security bugs.\n'
            'Audience: CTO at a 50-150 person software company.\n'
            'Length: max 150 words. Tone: direct, professional, no hype.'
        )
    }]
)
print(response.choices[0].message.content)

説明的なスタイル：ステップバイステップ

import anthropic

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

response = client.messages.create(
    model='claude-opus-4-5',
    max_tokens=400,
    messages=[{
        'role': 'user',
        'content': (
            'Write step-by-step instructional documentation for setting up a Python virtual environment.\n'
            'Style rules:\n'
            '- Begin with Prerequisites (1-3 items)\n'
            '- Number every step\n'
            '- Start each step with an imperative verb\n'
            '- Maximum one action per step\n'
            '- Include the exact command to run in a bash code block where applicable\n'
            '- Add a verification command after every 2-3 steps\n'
            'Audience: first-year developer who knows Python basics but never used a terminal.'
        )
    }]
)
print(response.content[0].text)

説明的なスタイル：ハウツー記事

import openai

client = openai.OpenAI(api_key='sk-your-key-here')

response = client.chat.completions.create(
    model='gpt-4o',
    messages=[{
        'role': 'user',
        'content': (
            'Write a how-to article outline (headings + 1-sentence descriptions only) '
            'for "How to Set Up GitHub Actions for Python Testing".\n'
            'Include exactly:\n'
            '- 1 hook sentence\n'
            '- What You Will Learn section (3 bullets)\n'
            '- Prerequisites section (3 bullets)\n'
            '- 5 numbered steps with H3 headings\n'
            '- Final Result section (1 sentence)\n'
            '- 2 Next Steps bullets\n'
            'No body content — outline only.'
        )
    }]
)
print(response.choices[0].message.content)

技術的なスタイル：正確かつ簡潔に

import anthropic

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

response = client.messages.create(
    model='claude-opus-4-5',
    max_tokens=250,
    messages=[{
        'role': 'user',
        'content': (
            'Write technical documentation for a Python function that rate-limits API calls.\n'
            'Technical style rules:\n'
            '- Define every parameter and return value precisely\n'
            '- Include type annotations in the description\n'
            '- Quantify all behaviors (e.g., "raises ValueError after 3 retries")\n'
            '- Active voice, no passive\n'
            '- Max 15 words per sentence\n'
            '- No analogies or metaphors\n'
            'Function: takes a callable, max_calls (int), window_seconds (int). '
            'Returns a wrapped callable that enforces the rate limit.'
        )
    }]
)
print(response.content[0].text)

適切なスタイルの選び方

import openai

client = openai.OpenAI(api_key='sk-your-key-here')

# Hybrid style: journalistic intro + instructional body + persuasive CTA
response = client.chat.completions.create(
    model='gpt-4o',
    messages=[{
        'role': 'user',
        'content': (
            'Write a 3-section technical blog post intro for "How to Cut Your AWS Bill by 40%".\n'
            'Section 1 (Journalistic): 1 sentence hook — state the key finding immediately.\n'
            'Section 2 (Instructional): 3-bullet overview of what the reader will learn/do.\n'
            'Section 3 (Persuasive): 1 sentence CTA to continue reading.\n'
            'Audience: AWS-using startup CTOs. Tone: direct, no hype.'
        )
    }]
)
print(response.choices[0].message.content)

セクション間でのスタイルの一貫性

import anthropic

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

response = client.messages.create(
    model='claude-opus-4-5',
    max_tokens=300,
    messages=[{
        'role': 'user',
        'content': (
            'Write a product launch announcement combining three styles — '
            'label each section.\n\n'
            '[JOURNALISTIC LEAD]: 2-sentence news-style announcement of the launch.\n'
            '[PERSUASIVE VALUE]: 2-sentence why-it-matters with problem-solution framing.\n'
            '[INSTRUCTIONAL NEXT STEP]: 3 numbered steps to get started.\n\n'
            'Product: Sentry AI — error monitoring with automatic root cause analysis.\n'
            'Audience: developer teams who currently spend hours debugging production errors.'
        )
    }]
)
print(response.content[0].text)

プロンプトでスタイルを指定する

import anthropic

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

style_names = [
    'journalistic inverted-pyramid style',
    'PAS (Problem-Agitate-Solution) persuasive framework',
    'step-by-step instructional documentation style',
    'technical documentation style — precise and concise'
]

topic = 'how to reset a forgotten password in our web app'

for style in style_names:
    response = client.messages.create(
        model='claude-opus-4-5',
        max_tokens=80,
        messages=[{
            'role': 'user',
            'content': f'Write about {topic} using {style}. Max 3 sentences or 3 steps.'
        }]
    )
    print(f'Style: {style[:45]}...')
    print(response.content[0].text.strip())
    print()

理解度チェック

プロフェッショナルな文体 — まとめ