Table of Contents
Ask Claude “Is this stock a good buy?” and you will get a confident answer built on nothing.
Claude does not know today’s price, last quarter’s margins, or what changed in the latest filing unless you give it that information. Without it, you get plausible-sounding analysis that may rest on stale or invented numbers.
The Claude prompts for stock analysis below fix this by making the data the starting point. Each one tells Claude to work only from the filings, tables, and transcripts you paste in, show its calculations, flag what is missing, and never tell you to buy or sell.
You get 10 prompts, each with a role, your context, a step-by-step process, strict rules, and an exact output format. They cover the business model, financial statements, filing changes, peer valuation, discounted cash flow scenarios, earnings calls, bear cases, competitive position, thesis tracking, and portfolio risk.
This post is educational. None of it is investment advice, and Claude is not a financial advisor.
How to Use Claude Prompts for Stock Analysis Safely
Three rules decide whether these prompts help or hurt.
Paste the data yourself. Free filings are available on the SEC’s EDGAR database: the 10-K (annual report), the 10-Q (quarterly report), and 8-Ks (major events). Copy the sections you need into the prompt. Claude analyzes what you paste, which is far more reliable than what it recalls.
Verify every number that matters. Claude can misread a table or make an arithmetic slip. Check key figures against the source and rerun important calculations in a spreadsheet.
Treat the output as research, not a recommendation. Every prompt forbids buy, sell, or hold calls. Your decision should come from your own judgment, your goals, and, where appropriate, a licensed advisor.
The prompts use XML-style tags to keep your pasted data separate from the instructions. Anthropic’s prompting guidance recommends this structure because it reduces the chance that Claude confuses your material with its task.
Claude Prompts for Stock Analysis: Understanding the Business and Its Financials
Start here. You cannot value a company you do not understand.
1. Summarize How a Company Actually Makes Money
Annual reports are long. This prompt extracts the business model from the section that explains it.
<role>
You are an equity research analyst with 15 years of experience covering a wide range of industries. You explain businesses plainly, without jargon, and you separate facts stated in the filing from your own inferences.
</role>
<context>
I am researching a company as a potential long-term investment. Below is the business description section from its latest annual report. Use only this text. Do not rely on memory for any numbers or claims about the company.
</context>
<inputs>
Company name and ticker: PASTE NAME AND TICKER
Business section of the annual report (10-K or equivalent): PASTE THE BUSINESS DESCRIPTION SECTION
Segment or revenue breakdown, if available: PASTE OR WRITE "NOT PROVIDED"
</inputs>
<instructions>
1. Explain in three sentences what the company does, who its customers are, and how it earns revenue.
2. List each revenue stream or segment with: what it sells, who buys it, how it is paid for (subscription, one-time, usage, advertising, or other), and its share of revenue if stated.
3. Identify the cost structure: the main things the company must spend money on to operate.
4. Identify what the filing says about competitive advantages. For each claimed advantage, rate how verifiable it is from the text: Strong (supported by specific facts), Moderate, or Weak (an assertion only).
5. List the key dependencies: customers, suppliers, regulations, technologies, or people the business relies on heavily.
6. List five questions the filing does not answer that an investor should research further.
</instructions>
<rules>
- Use only the pasted text. If a figure or fact is not in it, write NOT PROVIDED.
- Quote the filing for any claim of competitive advantage, and mark it as the company's own claim.
- Do not give an opinion on whether the stock is a good or bad investment.
- Use plain language. Define any technical term in the same sentence.
</rules>
<output_format>
Headings in order: Business in Three Sentences, Revenue Streams (table: Stream, Customers, How Paid, Share of Revenue), Cost Structure, Claimed Advantages and How Verifiable They Are, Key Dependencies, Questions the Filing Does Not Answer.
</output_format>
The final section is the most useful one. It gives you a research list based on what the company chose not to explain.
2. Analyze Five Years of Financial Statements for Trends and Red Flags
This prompt turns three statements into a clear picture of financial health.
<role>
You are a forensic financial analyst who reads financial statements for trends, quality of earnings, and warning signs. You are skeptical, precise, and show every calculation.
</role>
<context>
I am analyzing the financial health of a company. Below is financial data for several years. Calculate and interpret the metrics requested. Work only from the data pasted below.
</context>
<inputs>
Company and ticker: PASTE NAME AND TICKER
Currency and units: PASTE (for example: USD millions)
Income statement data for the last 5 years: PASTE DATA
Balance sheet data for the last 5 years: PASTE DATA
Cash flow statement data for the last 5 years: PASTE DATA
</inputs>
<instructions>
1. Calculate for each year: revenue growth rate, gross margin, operating margin, net margin, free cash flow (operating cash flow minus capital expenditure), free cash flow margin, return on equity, debt-to-equity ratio, and current ratio. Show the formula and the numbers used for at least one year so I can check your method.
2. Describe the trend in each metric in one sentence.
3. Compare net income with operating cash flow each year. Flag any year where profit is much higher than cash generation, and explain why this could matter.
4. Check for these warning signs and report each as Found, Not Found, or Cannot Assess: rising debt without matching growth, falling margins, receivables growing faster than revenue, inventory growing faster than sales, share count increasing materially, and one-off items inflating profit.
5. Give three strengths and three weaknesses of the financial profile, each tied to specific numbers.
6. List what additional data would improve this analysis.
</instructions>
<rules>
- Use only the pasted data. Write NOT PROVIDED for anything missing.
- Show calculations for transparency. Round to one decimal place.
- If the data looks inconsistent or unusual, flag it rather than correcting it silently.
- Do not recommend buying, selling, or holding.
</rules>
<output_format>
Headings in order: Metrics Table (rows are metrics, columns are years), Trend Summary, Earnings Quality Check, Warning Signs Checklist (table: Warning Sign, Status, Evidence), Strengths, Weaknesses, Data Gaps.
</output_format>
Spot-check two or three calculations yourself. If Claude’s method is correct on those, the rest is likely sound.
3. Spot What Changed Between Two Annual Reports
Companies rarely announce bad news loudly. Changes in risk language and accounting notes often show up first in the filings.
<role>
You are a securities analyst who specializes in reading changes between filings. You know that new risk language, altered accounting policies, and quietly removed disclosures often signal problems before they show in the numbers.
</role>
<context>
Below are the same sections from a company's annual report for two consecutive years. Compare them and tell me what changed and why it might matter.
</context>
<inputs>
Company and ticker: PASTE NAME AND TICKER
Last year's risk factors and key notes: PASTE PRIOR YEAR TEXT
This year's risk factors and key notes: PASTE CURRENT YEAR TEXT
</inputs>
<instructions>
1. Identify every risk factor that is new this year.
2. Identify every risk factor that was removed.
3. Identify every risk factor whose wording became stronger, weaker, or more specific. Quote the old and new wording side by side.
4. Identify changes to accounting policies, estimates, legal proceedings, customer concentration, debt terms, or related-party transactions.
5. Rank the five most significant changes by potential importance to an investor, with a one-sentence explanation for each.
6. List follow-up questions I should research for the top three changes.
</instructions>
<rules>
- Quote the exact text for every change. Do not paraphrase when reporting differences.
- Do not speculate about management's intentions. Describe what changed and why it could matter.
- If the two texts are too similar to show meaningful changes, say so.
- Do not recommend buying, selling, or holding.
</rules>
<output_format>
Headings in order: New Risks, Removed Risks, Changed Wording (table: Topic, Last Year, This Year, Why It Might Matter), Accounting and Other Changes, Top Five Changes Ranked, Follow-Up Questions.
</output_format>
If the filings are too long to paste in one go, run the prompt section by section and combine the results.
Claude Prompts for Valuing a Stock and Comparing Peers
Valuation is where many investors rely on gut feel. These prompts force assumptions into the open.
4. Compare a Company’s Valuation Against Its Peers
A multiple means little on its own. It only makes sense next to similar companies.
<role>
You are a valuation analyst who compares companies using standard multiples and understands why similar-looking companies can deserve different valuations.
</role>
<context>
I want to see whether a company's valuation looks high, low, or in line with its peers. I am pasting the data. Use only that data.
</context>
<inputs>
Target company and ticker: PASTE NAME AND TICKER
Peer companies and tickers: PASTE 3 TO 6 PEERS
Data table (price, market cap, enterprise value, revenue, revenue growth, operating margin, net income, free cash flow, P/E, EV/Revenue, EV/EBITDA, and any others you have, for each company): PASTE DATA
Date of the data: PASTE DATE
</inputs>
<instructions>
1. Build a comparison table with all available metrics for every company.
2. Calculate the peer median and average for each multiple, excluding the target company.
3. State whether the target trades at a premium, a discount, or in line with the peer median on each multiple, and by what percentage.
4. Compare growth and profitability next to the multiples. Explain whether any premium or discount seems explained by faster growth, higher margins, or lower risk, or whether it looks unexplained.
5. Identify any peer that is a poor comparison and explain why (for example: different business mix, size, or geography).
6. List the factors the multiples do not capture that would affect a fair valuation.
7. State clearly what this comparison cannot tell me.
</instructions>
<rules>
- Use only the pasted data. If a metric is missing for a company, write NOT PROVIDED and exclude it from that calculation.
- Show the median and average calculations.
- Do not say whether the stock is cheap, expensive, or a good buy. Describe the valuation relative to peers only.
- Mention the date of the data and warn that multiples change as prices move.
</rules>
<output_format>
Headings in order: Comparison Table, Peer Median and Average, Premium or Discount Analysis (table: Metric, Target, Peer Median, Difference), Do Growth and Margins Explain the Gap, Weak Comparisons, What the Multiples Miss, What This Cannot Tell Me.
</output_format>
Data providers sometimes calculate multiples differently, so pull all your numbers from the same source.
5. Build a Simple DCF With Bear, Base, and Bull Scenarios
A discounted cash flow (DCF) estimates what a company’s future cash flows are worth today. The result depends heavily on assumptions, so this prompt makes every assumption visible.
<role>
You are a valuation specialist who builds transparent discounted cash flow models. You always show assumptions, explain how sensitive the result is to them, and warn that DCF outputs depend heavily on inputs.
</role>
<context>
I want a simple, transparent DCF framework for a company. I am providing historical data and my own assumptions. Do not invent assumptions. Where I leave one blank, propose a range and label it clearly as your suggestion.
</context>
<inputs>
Company and ticker: PASTE NAME AND TICKER
Historical revenue and free cash flow for the last 5 years: PASTE DATA
Latest net debt (debt minus cash): PASTE NUMBER
Shares outstanding (diluted): PASTE NUMBER
Current share price: PASTE PRICE AND DATE
My assumptions (revenue growth by year, free cash flow margin, discount rate, terminal growth rate). Leave blank anything you are unsure about: PASTE ASSUMPTIONS OR WRITE "NONE"
Forecast period: PASTE YEARS (default 5)
</inputs>
<instructions>
1. List all assumptions used for each of three scenarios: Bear, Base, and Bull. Mark which came from me and which are your suggestions.
2. For each scenario, project revenue and free cash flow for each forecast year. Show the calculation for the first forecast year.
3. Calculate the terminal value using the perpetuity growth method. Keep terminal growth at or below a reasonable long-term growth rate, and warn me if my number is aggressive.
4. Discount all cash flows and the terminal value to present value. Show the formula.
5. Calculate enterprise value, subtract net debt to reach equity value, and divide by diluted shares for a value per share in each scenario.
6. Compare each scenario's value per share with the current price, as a percentage difference.
7. Build a sensitivity table showing value per share in the Base case at three discount rates and three terminal growth rates.
8. State what percentage of each scenario's value comes from the terminal value, and explain why that matters.
9. List the three assumptions the result is most sensitive to.
</instructions>
<rules>
- Show the full math so I can reproduce it in a spreadsheet.
- Do not present any scenario as a prediction or a price target.
- If the inputs are inconsistent (for example: negative free cash flow with high growth), flag it.
- Do not recommend buying, selling, or holding.
- End with a reminder that DCF output is only as reliable as its assumptions.
</rules>
<output_format>
Headings in order: Assumptions Table (Bear, Base, Bull), Projections for Each Scenario, Terminal Value Calculation, Value per Share by Scenario, Comparison With Current Price, Sensitivity Table, Terminal Value Share, Most Sensitive Assumptions, Reminder.
</output_format>
Rebuild the final numbers in a spreadsheet. Claude is good at the structure and logic, but a spreadsheet is a better home for the arithmetic.
Claude Prompts for Reading Management, Risks, and Competition
Numbers show what happened. These prompts help you judge what management says about what comes next.
6. Analyze an Earnings Call Transcript for Tone, Guidance, and Evasion
Management’s choice of words matters. This prompt reads a transcript the way an analyst would.
<role>
You are a buy-side analyst who reads earnings call transcripts closely. You track guidance changes, tone shifts, and how management answers hard questions. You separate what was said from what you infer.
</role>
<context>
Below is the transcript of a company's latest earnings call. If I include the previous quarter's transcript or guidance, compare them. Use only the text provided.
</context>
<inputs>
Company and ticker: PASTE NAME AND TICKER
Latest earnings call transcript: PASTE FULL TRANSCRIPT
Previous quarter's guidance or transcript (optional): PASTE OR WRITE "NOT PROVIDED"
</inputs>
<instructions>
1. Summarize the main messages from management's prepared remarks in five bullet points.
2. List every piece of forward-looking guidance (revenue, margins, spending, timelines, customer or market expectations) and quote it. If previous guidance is provided, mark each as Raised, Lowered, Maintained, or New.
3. Identify language that signals confidence, and language that signals caution or hedging (for example: "we expect," "we hope," "challenging environment," "timing"). Quote examples of each.
4. In the Q&A, identify questions that management did not answer directly. For each, quote the question and the response, and explain what was left unanswered.
5. List topics that analysts asked about repeatedly, as these show where investor concern is concentrated.
6. Note any metric that management emphasized this quarter that was not emphasized before, if I provided earlier material, as a sign that the story may be shifting.
7. List five follow-up questions I should research.
</instructions>
<rules>
- Quote exact text for every claim about what management said.
- Separate Stated Facts from Your Interpretation clearly, and label each.
- Do not predict the stock's reaction or its future price.
- Do not recommend buying, selling, or holding.
</rules>
<output_format>
Headings in order: Main Messages, Guidance Table (Item, Quote, Change), Confidence and Caution Language, Questions Not Directly Answered, Repeated Analyst Concerns, Emerging Emphasis, Follow-Up Questions.
</output_format>
The unanswered questions section is usually where the most useful information sits.
7. Argue Against Your Own Investment Idea Before You Act on It
Most investors collect evidence for what they already believe. This prompt makes Claude argue the other side as strongly as the facts allow.
<role>
You are a skeptical risk analyst whose only job is to find the strongest honest case against an investment idea. You are rigorous, fair, and grounded in the evidence provided. You never invent facts to make the bear case sound stronger.
</role>
<context>
I have a thesis about a company and want to test it. Below is my reasoning and the data I have gathered. Build the strongest evidence-based case against it.
</context>
<inputs>
Company and ticker: PASTE NAME AND TICKER
My investment thesis in my own words: PASTE THESIS
The evidence supporting it: PASTE DATA, FILINGS EXTRACTS, NOTES
Time horizon I am considering: PASTE (for example: 3 to 5 years)
</inputs>
<instructions>
1. Restate my thesis in two sentences to confirm you understood it.
2. List the assumptions my thesis depends on. Mark each as Supported by the data I provided, Partly supported, or Unsupported.
3. Build the bear case: the five strongest arguments against the thesis, each tied to evidence I provided or to a clearly labeled logical risk. For each, state what the evidence shows, why it matters, and how likely the risk is to affect the thesis (Low, Medium, High) with a reason.
4. Run a pre-mortem: imagine that three years from now the investment has lost significantly. Describe the three most plausible reasons, based on the information given.
5. Identify the single weakest point in my thesis.
6. List what evidence would prove the bear case right, and what evidence would prove it wrong.
7. List the conditions under which I should re-examine or exit the idea, expressed as measurable signals.
</instructions>
<rules>
- Use only the evidence I provided and clearly labeled logical reasoning. Do not invent news, numbers, or events.
- Be direct. Do not soften the criticism, but stay fair and factual.
- Do not tell me whether to invest. Test the thesis only.
</rules>
<output_format>
Headings in order: Thesis Restated, Assumption Check (table: Assumption, Support Level), Bear Case (table: Argument, Evidence, Why It Matters, Likelihood), Pre-Mortem, Weakest Point, What Would Prove It Right or Wrong, Re-Examination Signals.
</output_format>
Run this before you commit money, not after. It is much cheaper to find the weak point on paper.
The JSON Prompt Mastery for Claude collection includes 70+ pre-built structured prompts organized by business workflow, so you run the prompt and get the output without building anything from scratch. If you build research processes like this regularly, a ready-made library saves hours.
8. Map a Company’s Competitive Position Within Its Industry
A company’s results make more sense when you understand the competitive forces around it.
<role>
You are an industry analyst who evaluates competitive position using established strategy frameworks. You focus on evidence, not narrative.
</role>
<context>
I want to understand how strong a company's position is within its industry. Use the industry information and company details I paste. Do not add facts from memory.
</context>
<inputs>
Company and ticker: PASTE NAME AND TICKER
Industry description, competitor list, and any market share data: PASTE INFORMATION
Company's description of its competitive advantages (from its filings): PASTE TEXT
Customer and supplier information from its filings: PASTE TEXT
</inputs>
<instructions>
1. Assess each of these five forces and rate it Low, Medium, or High, with reasoning tied to the pasted information: threat of new entrants, bargaining power of suppliers, bargaining power of customers, threat of substitutes, and rivalry among existing competitors.
2. Evaluate the company's claimed advantages. For each, state the type (cost, brand, network effects, switching costs, regulatory, scale, or technology) and say whether the evidence supports it, partially supports it, or does not support it.
3. Identify what would need to happen for each advantage to erode.
4. Describe how profitable the industry is in general, if the data supports it, and whether the company's margins suggest it is outperforming.
5. List the three biggest threats to the company's competitive position.
6. State what additional data would improve the assessment.
</instructions>
<rules>
- Use only the pasted information. Write NOT PROVIDED where evidence is missing and avoid guessing.
- Distinguish between claims made by the company and facts supported by outside data.
- Do not recommend buying, selling, or holding.
</rules>
<output_format>
Headings in order: Five Forces Table (Force, Rating, Reasoning), Advantage Evaluation (table: Claimed Advantage, Type, Evidence Level, What Could Erode It), Industry and Margin Context, Top Three Threats, Data Gaps.
</output_format>
Pair this with prompt 1. Together they give you a clear picture of the business and the arena it competes in.
Claude Prompts for Building a Thesis and Checking Portfolio Risk
The last two prompts are about discipline. They help you write down your reasoning and keep it honest.
9. Write a One-Page Thesis With Monitoring Triggers
A written thesis stops you from rewriting your reasons after the price moves. This prompt turns your research into a document you can check against later.
<role>
You are an investment research editor who helps individual investors write clear, honest, one-page investment theses. You insist on specifics, measurable signals, and an explicit statement of what would prove the thesis wrong.
</role>
<context>
I have researched a company and want to document my reasoning in a structured way. Use my notes and the analysis outputs I paste below. Do not add information that I did not provide.
</context>
<inputs>
Company and ticker: PASTE NAME AND TICKER
My research notes, including business summary, financial analysis, valuation work, and risks: PASTE ALL NOTES AND PRIOR OUTPUTS
Why I find this company interesting, in my own words: PASTE REASONS
My time horizon: PASTE HORIZON
</inputs>
<instructions>
1. Write a one-page thesis with these sections, using plain language and short paragraphs:
- The Business (3 sentences)
- Why I Am Interested (the central reason, in 3 sentences)
- The Evidence (3 to 5 specific facts or numbers from my notes)
- Valuation View (what my notes say about the price relative to value, with the main assumptions stated)
- Key Risks (the 3 to 5 most important, ranked)
- What Would Prove Me Wrong (3 to 5 measurable conditions)
2. Create a monitoring checklist of 5 to 8 specific metrics to review each quarter. For each, give the current value from my notes, the direction that would support the thesis, and the level that should trigger a re-evaluation.
3. List the dates or events I should track, such as earnings releases or debt maturities, if they appear in my notes.
4. Write a short "Decision Journal" entry I can save: today's date, my reasoning, my confidence level (Low, Medium, High) and why, and what I expect to see in the next year.
5. Flag any claim in the thesis that rests on weak or missing evidence.
</instructions>
<rules>
- Do not include information not found in my notes.
- Every trigger must be measurable. No vague conditions such as "if things get worse."
- Do not state a price target or recommend an action.
</rules>
<output_format>
Headings in order: One-Page Thesis, Quarterly Monitoring Checklist (table: Metric, Current Value, Supports Thesis If, Re-evaluate If), Dates to Track, Decision Journal Entry, Weak Evidence Flags.
</output_format>
Save the decision journal entry. When you look back in a year, it shows how your reasoning actually held up.
10. Review Your Portfolio for Concentration and Hidden Risk
A portfolio can look diversified and still depend on one idea. This prompt tests for that.
<role>
You are a portfolio risk analyst who reviews holdings for concentration, overlap, and exposure to common risk factors. You explain risks clearly and never tell the investor what to buy or sell.
</role>
<context>
I want to understand the risk structure of my portfolio. Review the holdings below. Use only the data I provide, and tell me what additional data would sharpen the analysis.
</context>
<inputs>
Holdings (ticker, company or fund name, approximate value or percentage weight, and sector or description if known): PASTE HOLDINGS
Total portfolio size (approximate): PASTE VALUE
My time horizon and how I would react to a 30 percent fall (for example: I would hold, I would be worried, I would need to sell): PASTE HONEST ANSWER
Any income needs from the portfolio in the next 3 years: PASTE OR WRITE "NONE"
</inputs>
<instructions>
1. Calculate the weight of each holding and identify the top 5 positions and their combined share of the portfolio.
2. Group holdings by sector, geography (if determinable), and style (growth, value, income, or other). Show the percentage in each group.
3. Identify overlap: funds or stocks that likely hold the same underlying companies. Mark each as Confirmed (from data given) or Possible (needs checking).
4. Identify concentration risks: any single stock, sector, or theme that makes up an outsized share, and explain what scenario would hurt most.
5. Run four simple stress scenarios in plain terms, showing approximate portfolio impact using the weights given: a broad market fall of 20 percent, a sharp fall in the largest sector, a fall in interest-rate-sensitive holdings, and a fall in the largest single holding. State the assumptions clearly. These are illustrations, not predictions.
6. Compare the risk profile with my stated time horizon and reaction to a 30 percent fall, and describe any mismatch plainly.
7. List the questions I should discuss with a licensed financial advisor.
</instructions>
<rules>
- Do not recommend buying, selling, rebalancing, or any specific allocation.
- Do not describe any scenario as likely or predicted.
- If holdings are unclear or the data is incomplete, state your assumptions.
- State that this analysis is educational and is not personalized financial advice.
</rules>
<output_format>
Headings in order: Position Weights (table), Group Exposure (table), Overlap Check, Concentration Risks, Stress Scenarios (table: Scenario, Assumption, Approximate Impact), Fit With My Stated Horizon and Tolerance, Questions for an Advisor, Disclaimer.
</output_format>
Be honest in the field about how you would react to a large fall. The most useful part of this prompt is the mismatch check, and it only works with a truthful answer.
Mistakes That Make Claude Stock Analysis Unreliable
Most bad results come from a small set of avoidable habits.
Asking for a verdict. “Should I buy this?” produces an opinion without evidence. Ask for analysis of specific data instead.
Relying on Claude’s memory for numbers. Financial figures change every quarter. Paste them from the filing.
Skipping verification. A misread table or arithmetic slip can flow through the whole analysis. Check the numbers that drive your conclusion.
Treating a DCF as a price target. The output depends on the assumptions you choose. Change them slightly, and the answer shifts a lot.
Using it to confirm what you already believe. Run prompts 7 and 3 deliberately, since they are built to find problems.
How to Combine These Claude Prompts Into a Research Workflow
You do not need all ten for every company. A clean order works for most research.
Start with prompts 1 and 2 to understand the business and its financial health. Run prompt 3 to check what changed in the filings, then prompts 4 and 5 for valuation.
Add prompts 6 and 8 for management and competitive context. Always run prompt 7 before you decide anything, and finish with prompt 9 to document your reasoning. Use prompt 10 on its own whenever you review your portfolio as a whole.
Keep every output in one document per company. Over time, that file becomes a research record you can check against what actually happened.
Good Analysis Starts With Your Data, Not Claude’s Memory
The pattern behind all ten prompts is the same: real data from you, a specific task, strict rules, and a fixed output format. That is what turns Claude from a confident guesser into a useful research assistant.
Pick one company you already follow and run prompts 1 and 2 on its latest annual report. Compare the output with your own understanding and see where it agrees, where it disagrees, and what it missed.