How a simple framework helped me communicate more effectively with AI while building an AI-powered investing platform.

💡 Have You Ever Wondered Why AI Gives Great Answers Sometimes—and Average Ones at Other Times?

I certainly have.

Over the past few years, while building an AI-powered investing platform, I’ve spent countless hours working with ChatGPT, Claude, Gemini, Perplexity, Copilot, and other AI assistants. They have helped me write code, analyze financial statements, design software architecture, prepare documentation, create content, and brainstorm new ideas.

One observation stood out very quickly.

The biggest difference between an average AI response and an excellent one usually wasn’t the AI model—it was how I asked the question.

Like many people, I initially wrote short prompts and hoped the AI would infer what I wanted. Sometimes it did. Often, it didn’t.

As my projects became more complex, I found myself repeatedly including the same information in my prompts: the role I wanted the AI to assume, the background, the objective, the constraints, and the preferred output.

Instead of rewriting those elements every time, I organized them into a simple, reusable framework.

This article shares that framework.

I don’t claim it’s the only way to write prompts, nor do I claim to have invented prompt engineering. It’s simply the approach that has worked well for me across different AI models and different types of work.

If you find it useful, feel free to adapt it to your own workflow. 😊

📋 My AI Prompt Template

Role:

Context (including the objective):

Task (including inputs, if any):

Assumptions (if any):

Constraints (including limitations, restrictions, dependencies, or requirements):

Output Format (optional):

Output Style (optional):

Output Restrictions (optional):

The framework is intentionally simple. Most prompts don’t need every section, but having a consistent structure makes it easier to communicate your requirements clearly.

🔍 Why Each Section Matters

👤 1. Role

🎯 Purpose

Tell the AI what perspective to adopt.

✅ Examples

• Act as an experienced financial analyst.
• Act as a software architect.
• Act as a travel planner.
• Act as a marketing consultant.

⚠️ Common Mistake

Leaving the role undefined and expecting the AI to know the level of expertise you want.

🌍 2. Context (Including the Objective)

🎯 Purpose

Context explains the background, while the objective explains what success looks like.

✅ Example

I am comparing three Indian banking companies for long-term investing. My objective is to identify businesses with strong fundamentals and sustainable dividend growth.

The more relevant context you provide, the more focused the response is likely to be.

💡 Tip

Good context saves AI from making unnecessary assumptions.

📝 3. Task

🎯 Purpose

Clearly describe what you want the AI to do.

Instead of:

Tell me about HDFC Bank.

Try:

Compare HDFC Bank, ICICI Bank, and Kotak Mahindra Bank using profitability, asset quality, valuation, and long-term growth prospects. Conclude with a ranked recommendation.

Specific tasks produce more useful answers.

🤔 4. Assumptions

🎯 Purpose

Sometimes information is missing. Rather than letting AI make hidden assumptions, either state them yourself or ask the AI to list its assumptions before answering.

✅ Example

Assume I am a long-term dividend investor with a 15-year investment horizon.

💡 Tip

Making assumptions explicit increases transparency and makes responses easier to validate.

🚧 5. Constraints

🎯 Purpose

Constraints define the boundaries.

✅ Examples

• Budget below ₹2 lakh.
• Recommend only open-source software.
• Use publicly available information.
• Avoid speculative assumptions.
• Explain in simple English.

Constraints keep the response aligned with your requirements.

⚠️ Common Mistake

Don’t confuse constraints with output formatting.

📊 6. Output Format

🎯 Purpose

Choose how you want the answer organized.

✅ Examples

• 📋 Table
• 🔹 Bullet points
• 📝 Markdown
• 📄 Report
• 🔧 JSON
• ✔️ Checklist
• 🔢 Numbered list

A good format often makes the response easier to understand than adding more detail.

🎨 7. Output Style

🎯 Purpose

This controls the tone and depth of the response.

✅ Examples

• 👶 Beginner-friendly
• 👔 Executive summary
• 💻 Technical
• 🎓 Academic
• 💬 Conversational
• 🏢 Professional

The same information can be presented very differently depending on the intended audience.

📏 8. Output Restrictions

🎯 Purpose

These define limits on the final response.

✅ Examples

• Maximum 500 words.
• No tables.
• No code.
• Five recommendations only.
• Keep the explanation concise.

Restrictions help prevent overly long or unnecessarily detailed answers.

💼 A Practical Example

Instead of writing:

Analyze these companies.

I would write:

👤 Role: Act as an experienced equity research analyst.

🌍 Context: I am building a long-term investment portfolio focused on Indian companies with strong fundamentals and consistent cash generation.

📝 Task: Compare the attached annual reports and rank the companies based on financial strength, profitability, valuation, and long-term investment potential.

🤔 Assumptions: Assume I have a 15-year investment horizon and prefer dividend-paying businesses.

🚧 Constraints: Use only the information provided. Clearly mention any assumptions.

📊 Output Format: Present the results as a comparison table followed by a ranked summary.

🎨 Output Style: Professional but easy to understand.

📏 Output Restrictions: Limit the report to two pages.

This prompt leaves very little room for ambiguity.

⭐ Why I Continue to Use This Template

I use this framework because it helps me think more clearly before asking AI to solve a problem.

It isn’t about making prompts longer.

It’s about making them more structured.

Sometimes I use only three sections.

For larger projects, I use all eight.

The flexibility is what makes the framework practical.

✅ Advantages

• 🚀 Works across multiple AI models.
• 🧠 Encourages clear communication.
• 🎯 Reduces ambiguity.
• 💼 Suitable for finance, software, research, writing, and everyday tasks.
• 📚 Easy to remember.
• 🔄 Flexible enough for both simple and complex prompts.

⚠️ Limitations

No template can guarantee perfect responses.

The quality of the answer still depends on:

• 🤖 The capabilities of the AI model.
• 📄 The accuracy of the information provided.
• 🧩 The complexity of the task.

Think of this framework as a strong starting point rather than a rigid rulebook.

🎯 Final Thoughts

Smart Investing India focuses primarily on investing and artificial intelligence. While building the platform, I’ve learned that structured communication makes a meaningful difference when working with AI.

This prompt template is one of the practical techniques I use every day.

It may evolve over time as AI continues to improve, and I expect future versions to incorporate new capabilities and lessons learned.

If this framework helps you write better prompts or simply encourages you to think more clearly before asking AI a question, then it has achieved its purpose.

Thank you for reading, and happy prompting! 🚀🤖