Iterating Prompts to Refine Results

Your first prompt won’t always deliver the perfect shortlist — and that’s okay.

The real strength of AI screening is iteration.

Refine the prompt, add a condition, remove a constraint, try again.

Treat it like a conversation with an assistant: each follow-up makes the output sharper and closer to what you need.

Why Iteration is Important

• Markets are complex. A single prompt can’t always capture everything you want.
• Broad prompts give too many results; narrow prompts can give none.
• Iteration helps you adjust in small steps instead of starting from scratch every time.

Start Broad, Then Narrow Down

It’s often better to start with a broad prompt and then refine.

Example:

• Step 1: “Show me Nifty 500 stocks with P/E below 20.” → You may get 40 companies.
• Step 2: “Now filter only those with ROE above 15%.” → List shrinks to 15.
• Step 3: “Remove companies with promoter pledges.” → You’re left with 8.

By layering conditions, you end up with a practical shortlist.

Add Metrics One at a Time

Instead of writing a single complicated prompt with 10 conditions, add metrics gradually.

• First: “List IT stocks with sales growth in the last 3 years.”
• Next: “Now show only those trading above 200-day moving average.”
• Then: “Add debt-to-equity ratio in the table.”

This way, you control the process and avoid overwhelming the AI.

Reframe When Needed

If the output still looks off, reframe the wording.

• Weak prompt: “Find strong stocks in pharma.”
• Better prompt: “Show me pharma stocks with profit growth above 12% for 3 years and RSI below 30.”

Iteration is not just about adding filters — sometimes it’s about replacing vague words with measurable ones.

Example Workflow

Suppose you’re exploring the auto sector.

1. Prompt: “Find auto stocks under ₹1,000.”
2. Refine: “Now show only those with ROE above 12%.”
3. Refine again: “Add RSI values to the table.”
4. Narrow further: “Remove companies with debt-to-equity above 1.”

At the end, you get a shortlist of auto companies that are affordable, financially sound, and technically positioned well.

Avoid Overfitting

Be careful not to keep refining until you’re left with just one or two names that look “perfect.” In real markets, there’s no such thing as a flawless stock. Too many constraints can make you miss out on opportunities.

Final Takeaway

Iteration is the real secret to using AI screeners well.

Start broad, refine step by step, and keep prompts clear and measurable.

Each round of refinement makes your shortlist sharper without overwhelming the system.

But don’t chase perfection — the goal is a focused, practical list, not an endless search for the “perfect” one