ChatGPT Stock Analysis Prompts for Research

These are research prompts: reading a 10-K, pulling the margin trend out of an earnings call, comparing two companies on figures you supply, building a bull and bear case that cites its own sources. Each one pastes the document in rather than asking ChatGPT to recall it, which is the single change that stops invented numbers. Execution prompts for charts, stops and sizing live on a separate page.

What makes a good ChatGPT stock analysis prompt?

A good prompt does four things: gives the model a role, hands it the source material to work from, specifies the exact structure of the answer, and asks for reasoning and scenarios rather than a single call. This mirrors OpenAI's own prompt-engineering guidance: be specific, provide context, and show the format you want.

The biggest single upgrade is sourcing. Asking "is Apple a good stock?" invites a vague, possibly stale, possibly fabricated answer. Pasting the latest 10-K excerpt or earnings transcript and asking ChatGPT to analyze that text grounds the model in real data and slashes hallucination. Throughout the prompts below, the rule holds: bring the facts, let ChatGPT structure them. For the broader method, see our guide on how to use ChatGPT for stock trading.

What's a good ChatGPT prompt for fundamental analysis?

A strong fundamental analysis prompt gives ChatGPT a role, the financial data or filing text, and a structured checklist to fill in, including both strengths and risks. Here is a copy-ready template:

Act as an equity research analyst. Using ONLY the financial data
I paste below, produce a structured fundamental summary for [TICKER]:
1. Business model and main revenue streams
2. Key ratios present in the data (P/E, margins, debt-to-equity, ROE)
3. Three strengths and three risks, each with the figure that supports it
4. A short bull case and a short bear case
Do not use any numbers not present in the data. Flag anything you
cannot determine from what I provided.

DATA:
[paste the filing excerpt, transcript, or financials here]

The two load-bearing instructions are "use only the data I paste" and "flag anything you cannot determine." Together they keep the model from filling gaps with invented figures, which is the main fundamental-analysis failure mode.

What's a good ChatGPT prompt for valuation and peer comparison?

Valuation prompts fail when the model recalls multiples. They work when you paste the numbers and ask for reasoning about them:

Act as an equity research analyst. Below is a table of five companies
in the same sector with revenue growth, gross margin, operating margin,
net debt to EBITDA and forward P/E.
Using ONLY this table:
1. Rank them on growth-adjusted valuation and explain each placement
   in one sentence
2. Name the single figure that most justifies the highest multiple
   in the table
3. Name the single figure that would most change your ranking if it
   were revised
4. Flag any row whose numbers look internally inconsistent
Do not add companies or figures that are not in the table, and do not
tell me what to buy.

DATA:
[paste the table here]

A discounted cash flow prompt works the same way: supply your own revenue, margin and discount rate assumptions and ask the model to run the arithmetic, restate the assumptions it used, and show how the output moves if the discount rate shifts by one percentage point. Ask it to do the valuation from memory and you will get a confident number built on figures it made up. Our how to value a stock guide covers the underlying method.

What about chart and technical prompts?

Those belong to execution rather than research, and they are collected in our ChatGPT trading prompts guide: chart structure, entries, stops, invalidation prices, position size and trade review, each written with a ticker and a timeframe already in it. The one thing worth repeating here is that a general model has no fixed definition of support and no memory of the previous read, so the same chart screenshot can return different levels on different days. Bullynx's chart analysis applies the same technical routine to every upload so the levels and invalidation come back comparable across sessions. It reads only the image you supply, has no order flow or news feed, and does not issue buy or sell calls.

What's a good ChatGPT prompt for reading an earnings report?

A good earnings prompt pastes the transcript or release and asks ChatGPT to extract what changed, the guidance, and management's tone, while comparing against the prior period you also provide. Reading filings well is a core skill the SEC explains in its guide to reading a 10-K. Template:

Summarize this earnings call transcript for [TICKER]. Output:
1. Headline results vs. the prior quarter (only figures stated in the text)
2. Forward guidance and any changes to it
3. Three things management emphasized and three concerns or risks raised
4. Notable shifts in tone versus the previous quarter I paste below
Quote the exact line for each point. Do not add outside data.

CURRENT TRANSCRIPT:
[paste]

PRIOR TRANSCRIPT (for tone comparison):
[paste]

Asking it to quote the exact line for each point makes the summary auditable: you can check every claim against the source in seconds.

What's a good ChatGPT prompt for risk and position sizing?

A good risk prompt has ChatGPT structure your own risk rules into a checklist and sanity-check the math you provide, without making the sizing decision for you. It should never invent a stop or a target. Template:

Act as a risk manager. Given my account size of [X], my max risk per trade of [Y%], an entry around [PRICE], and a stop at [PRICE], walk through the position-size math step by step and show the resulting share count and dollar risk. Then list three risk questions I should answer before taking this potential setup. Do not recommend whether to take the trade.

This keeps ChatGPT in the role of a checklist and calculator helper. For the actual numbers, use a dedicated calculator and confirm the output, since a misplaced decimal in a prompt can produce a badly wrong size.

How do you keep ChatGPT stock prompts safe?

You keep these prompts safe by treating every output as a draft to verify, never as a signal. The prompts above are engineered to reduce hallucination, but no prompt eliminates it, so the verification habits below are non-negotiable.

Regulators are explicit that "AI can pick winners" claims are a fraud red flag, as the SEC investor alert on AI and investment fraud details. A prompt that asks for structured analysis of data you supplied is useful. A prompt that asks for a guaranteed pick is a trap.

These prompts pair naturally with a fuller research workflow, like the one in our guide on how to analyze a stock. And when the technical-analysis step needs a dependable chart read rather than a screenshot guess, the Bullynx AI trading copilot runs a chart-specific prompt on an uploaded chart, which is steadier than general ChatGPT for that one job.

Related reading: how to use ChatGPT for stock analysis step by step, how to read an earnings report, and best AI prompts for investing.