Prompt Engineering for Financial Writing

To make the most of AI, you need to master prompt engineering. That means carefully crafting the instructions you give to the AI model. In finance writing, start by specifying exactly what you want: the format (bullet points, essay, chart caption), the audience level, and any data to use.

💡 Key Insight

Deloitte's examples are instructive: you might prompt an AI to "Summarize the quarterly financial performance during FY2023" or to "Write a variance analysis report for June 2023 comparing actual results to budget". These precise instructions yield structured, useful outputs rather than vague paragraphs.

Best Practices for Financial Prompt Engineering

Here are some proven techniques to get better outputs from AI when creating financial content:

🎯 1. Be Specific in the Task

Instead of "Tell me about Company X," say "Draft a 200-word investor update on Company X's Q2 earnings, focusing on revenue growth and market outlook." This limits the AI's scope and ensures you get exactly what you need.

❌ Vague Prompt:

"Write about Tesla's performance"

✅ Specific Prompt:

"Write a 150-word analysis of Tesla's Q3 2024 delivery numbers, comparing them to analyst expectations and explaining the impact on stock price. Use a professional tone for institutional investors."

📊 2. Provide Data and Context

If possible, include the actual numbers or facts in your prompt (e.g. past P&L data or quotes) so the AI doesn't have to guess. You can even copy-paste a mini dataset and say "Use this for reference."

Example with Data:

"Using the following data, write a quarterly performance summary:

Q3 Revenue: $5.2B (vs $4.8B Q2, +8.3%)
Net Income: $420M (vs $380M Q2, +10.5%)
Operating Margin: 18.5% (vs 17.2% Q2)
Customer Growth: 2.1M new accounts

Focus on positive momentum and margin expansion. Keep it under 200 words."

🎨 3. Set the Style and Tone

Indicate the voice you want (e.g. "friendly tone for beginners" or "formal analysis for professionals"). For example, prompt "Explain the concept of an ETF in simple terms."

Tone Variations:

For Beginners:

"Explain dividend investing in simple terms that a college student could understand. Use everyday analogies and avoid jargon."

For Professionals:

"Write a technical analysis of the Fed's monetary policy stance for institutional investors. Use industry terminology and reference recent FOMC statements."

For Social Media:

"Create an engaging eToro post about renewable energy stocks. Use a conversational tone, include 2-3 emojis, and keep it under 280 characters."

📝 4. Use Examples and Formatting Cues

Ask the model to output bullet lists, tables, or headings by specifying it ("Summarize this in 3 bullet points" or "Output as a short report with sections").

Formatting Example:

"Create a market update with the following structure:

Opening paragraph (2-3 sentences on overall market sentiment)
3 bullet points highlighting key market movers
Brief analysis paragraph (3-4 sentences)
Closing sentence with outlook
Total length: 200-250 words"

✅ 5. Fact-Check Outputs

Treat the AI draft as a first pass. Always verify any facts, figures or claims it generates. The AI can synthesize plausible content, but errors are possible, especially in finance where accuracy is critical.

⚠️ Critical Reminder

AI models can "hallucinate" facts, dates, or statistics. In financial writing, a single error can damage your credibility. Always cross-reference numbers with official sources, verify dates, and check that the AI hasn't confused different companies or time periods.

Try StockStudio's Free Editor

Easily format your financial content for social media using our free editor.

Try Free Editor →

Putting It All Together

Think of prompt engineering as guiding the AI. Good prompts (with clear instructions like those in Deloitte's guide) will yield better, more relevant financial copy. As you experiment, you'll learn how to phrase your prompts to get precisely the market commentary or analysis you need.

Quick Prompt Template for Financial Content:

Task: [Specific output you want - e.g., "Write a stock analysis"]

Format: [Structure - e.g., "3 paragraphs with bullet points"]

Length: [Word count - e.g., "200-250 words"]

Tone: [Voice - e.g., "Professional but accessible"]

Audience: [Who's reading - e.g., "Retail investors"]

Data: [Include relevant numbers/facts]

Focus: [Key points to emphasize - e.g., "Revenue growth and market share"]

With practice, you'll develop an intuition for what works. Start with these best practices, experiment with different prompt structures, and refine based on the outputs you receive. The goal is to make AI your co-pilot in financial content creation: one that consistently delivers high-quality, accurate, and engaging content.

Reference

[1] Prompt Engineering for Finance 101 | Deloitte US

https://www.deloitte.com/us/en/services/consulting/articles/prompt-engineering-for-finance.html

Prompt Engineering for Financial Writing: Get Better AI Outputs