How to use this page
The difference between asking an AI "what do you think of $NVDA" and getting actually-useful research is one thing: structure in the prompt. You have to tell the AI what role to play, what inputs you're giving it, what format you want the output in, and what to avoid.
The 12 prompts below do that for you. Each one is a real workflow we use. Pick one, modify the bracketed parts, and run it.
1The "Quick stock breakdown"
Best for: ChatGPT (web search) · Use when: 5-minute first look at any ticker
You are an experienced equity analyst. Give me a one-page breakdown of [TICKER]:
1. What the company does (1 sentence, plain English)
2. Current revenue, gross margin, and net income (most recent quarter)
3. Market cap, P/E, forward P/E
4. Three things going well right now
5. Three things to worry about
6. The single biggest catalyst on the calendar (earnings date, product launch, lawsuit, etc.)
Cite recent sources where possible. Don't speculate on price.
Tip: Add "respond in under 250 words" if you want it tighter. Add "compare to [PEER TICKER]" to get relative context.
2The "Bull case + bear case"
Best for: Either AI · Use when: Stress-testing your conviction before a trade
I'm considering a position in [TICKER]. Give me:
- The 3 strongest bull arguments, with the specific numbers, events, or trends that support each
- The 3 strongest bear arguments, with the same level of specificity
- One scenario in which the bull case is right and the bear case still plays out (i.e., where do you make money but eventually have to exit?)
Be balanced. I want the smartest possible version of each side, not your opinion.
Tip: That last bullet is the most important. The "right for the wrong reasons" scenario is what most retail traders miss.
3The "Earnings call summary"
Best for: Claude (long context) · Use when: You missed the call
Here's the transcript of [COMPANY]'s most recent earnings call. [PASTE TRANSCRIPT OR UPLOAD FILE]
Give me:
1. The top 5 takeaways for an investor
2. Any guidance changes (and the direction)
3. Specific moments where the CEO or CFO sidestepped a question — quote the exact language
4. The single biggest unanswered question that an analyst should follow up on
5. Anything that contradicts what the company said on prior calls
Don't summarize the prepared remarks unless they contain new information.
Tip: Earnings call transcripts are free at sites like SeekingAlpha, Motley Fool, or directly from company IR pages. Drop the whole thing in.
4The "10-K reader"
Best for: Claude · Use when: You need to read a 10-K but don't have hours
Here's the most recent 10-K for [COMPANY]. [UPLOAD]
Read the entire filing. Then give me:
1. The 3 risks I should care about most (with section references)
2. Anything new vs last year's 10-K (added risks, removed risks, language changes)
3. Any related-party transactions in the footnotes
4. Any auditor-flagged issues (changes in critical audit matters, going-concern notes)
5. Significant changes in accounting policies or estimates
Be skeptical. If something looks like it's worded to obscure rather than disclose, flag it.
Tip: This is where Claude shines — it can hold the whole filing in context. ChatGPT free tier may make you paste in chunks. More on that here.
5The "Build me a watchlist"
Best for: ChatGPT (web search) · Use when: Starting research in a new sector
Build me a watchlist of [N] [LARGE-CAP / MID-CAP / SMALL-CAP] [SECTOR] stocks that meet ALL of these criteria:
- [CRITERION 1, e.g. positive free cash flow]
- [CRITERION 2, e.g. dividend yield over 3%]
- [CRITERION 3, e.g. trading below 200-day MA]
- [CRITERION 4, optional]
For each name, give me: ticker, company in 1 sentence, why it matches, and one risk. Flag any name where you're not confident the criteria are still met — I'll verify.
Tip: Always verify the criteria in a real screener (Finviz is free) before acting. AI watchlist quality is better than nothing but not bulletproof.
6The "News interpreter"
Best for: ChatGPT · Use when: Breaking news drops and you need to react
[Paste news headline or article URL]
For this news, walk me through:
1. The first-order impact (what literally just changed)
2. Which sectors / tickers benefit most, in what direction, and roughly how much
3. Which sectors / tickers are hurt
4. The second-order impact most retail traders will miss (1-2 sentences)
5. 3 ETFs and 3 single names worth watching for the next 1-2 trading days
Don't predict price. Just frame the implications.
Tip: Best with a real article URL pasted in (so the AI works from primary text), not just a headline.
7The "What am I missing?"
Best for: Claude · Use when: You're about to pull the trigger
I'm about to take a position in [TICKER] at around [PRICE]. My thesis is:
[PASTE YOUR THESIS — 2-4 sentences on why you think this works]
Tell me what I'm missing. Specifically:
1. Any obvious counter-argument my thesis doesn't address
2. Any risk specific to my entry timing (earnings in 3 days? technical resistance overhead?)
3. Any historical pattern where this exact setup has failed
4. The smartest single question I should answer for myself before clicking buy
Don't validate me. Try to talk me out of the trade. If you can't, that's a signal.
Tip: Run this twice with two different AIs. If both can't talk you out of it, that's stronger conviction.
8The "Sector comp"
Best for: ChatGPT · Use when: Choosing between similar names
Compare [TICKER A], [TICKER B], and [TICKER C] across:
- Revenue growth (TTM and 5Y CAGR)
- Operating margin
- Net cash / net debt
- Forward P/E
- Return on invested capital (ROIC)
- Insider ownership %
Then tell me: at current prices, which is the best risk/reward and why. If they're roughly equivalent, say so.
Tip: Always verify the numbers — finance data in AI training sets ages quickly. Cross-check on Stock Analysis or Yahoo Finance.
9The "Position size sanity check"
Best for: Either AI · Use when: Sizing the trade
My account size is $[AMOUNT]. I want to buy [TICKER] at $[ENTRY] with a stop at $[STOP]. My max loss tolerance per trade is [X%] of account.
Calculate:
1. The number of shares I should buy
2. The dollar risk if my stop hits
3. As a % of account
4. Whether this size is consistent with my [X%] rule
If the position size is uncomfortable for the setup (e.g. wide stop, big size), flag it.
Tip: Boring but underrated. Fixed-fractional position sizing is the single most underused risk control in retail trading.
10The "Macro context"
Best for: ChatGPT (web search) · Use when: Markets feel weird and you can't tell why
Give me a macro briefing for the current trading session:
- Major economic data releases today (with consensus and prior)
- Notable Fed speakers and what they're likely to say
- Significant earnings reports (after-hours and pre-market)
- Geopolitical events that could move markets
- What sectors are leading/lagging year-to-date and this week
- Anything unusual in the bond market, dollar, or commodities
Keep it under 300 words. I just need the lay of the land before I start trading.
Tip: Run this every morning. After two weeks you'll have built a habit and a macro framework simultaneously.
11The "Investment thesis writer"
Best for: Claude · Use when: Forcing yourself to articulate before buying
Write a 1-page investment thesis for a long position in [TICKER] held for [HOLDING PERIOD]. Use this structure:
1. The setup (why now, what changed)
2. The bull case (3 reasons, with specifics)
3. The risks (3 things that would invalidate the thesis)
4. The catalyst path (what needs to happen for the thesis to play out)
5. The exit plan (when do I take profits, when do I cut losses)
Write it like an analyst recommendation memo, not a Reddit post. Be specific and falsifiable.
Tip: Save the output. Six months later, re-read it. The retrospectives are how you actually get better at this.
12The "Trade journal coach"
Best for: Claude · Use when: Reviewing trades at week's end
Here are my trades from this week. [PASTE LIST: ticker, entry, exit, P&L, reason for entry, reason for exit]
Critique my trading. Specifically:
1. Patterns in what's working vs not
2. Cases where my entry reason and exit reason are inconsistent (e.g., entered for momentum, exited for value)
3. Cases where my position sizing didn't match my conviction
4. The single behavior I should change next week
Be honest. Don't soften the feedback.
Tip: This is the highest-ROI prompt on the list. Most traders never review trades systematically. The ones who do, improve.
Two general rules that make every prompt better
Rule 1: Tell the AI to admit uncertainty. Add "Flag anything you're not 100% sure about" to any research prompt. Both ChatGPT and Claude will hallucinate confident-sounding numbers if you don't ask them not to.
Rule 2: Always verify the numbers. AI is great at framework, structure, and synthesis. AI is bad at recent specific data points (training data cutoff). Use the AI for the analysis, your screener for the numbers, your eyes on the chart for the price.
Which AI for which prompt
The shorthand:
- ChatGPT for anything that needs fresh data — news interpretation, watchlist building, macro briefings, sector compares.
- Claude for anything long-form — earnings transcripts, 10-Ks, multi-page filings, careful thesis writing.
- Either works for the rest. Try both, see which voice you prefer.
Both have free tiers that are more than enough for these prompts. See our full ChatGPT vs Claude comparison →
If you read nothing else
The one prompt that matters most.
If you only adopt one prompt from this list, make it #7 — "What am I missing?" Run it before every position. The reason it matters: AI's biggest comparative advantage over you isn't intelligence. It's that it doesn't have your emotional attachment to the trade. Asking an AI to talk you out of a trade is the cheapest second opinion you'll ever get.