10 ChatGPT & Claude Prompts to Audit Your Stock and ETF Portfolio (Tested in 2026)

Methodology & quantitative verification note: We tested over 60 financial prompt variations across ChatGPT (GPT-4o / o3) and Anthropic’s Claude using real ETF holdings (SPYM — formerly SPLG — VOO, QQQM, SCHD, VXUS) and SEC 10-K filings. Every mathematical prompt below includes a built-in Python / Code Interpreter instruction to eliminate LLM arithmetic hallucinations. Read our editorial policy.

Most retail investors use ChatGPT completely wrong. They open a chat window, type “What are the best stocks to buy right now?”, and get a generic, outdated list of mega-cap companies that adds zero value.

Large Language Models (LLMs) are not crystal balls — they are institutional-grade data analysts. When you feed them your specific holdings, current ETF fact-sheet weights, and tax brackets using structured prompts, ChatGPT and Claude can expose hidden risks in your portfolio that would normally cost $300/hour with a human advisor.

Below are the 10 exact copy-paste prompts we use at MoneyPilotAI to audit ETFs, stress-test stock portfolios, and dissect company earnings in minutes. (For a broader look at building a complete automated system, check out our Complete 2026 Guide to AI Personal Finance & Investing).

⚠️ The 3 golden rules before using AI for investing (avoid hallucinations!)

Before you copy-paste the prompts below, follow these three non-negotiable safety rules:

1. Never let an LLM do compound math “in its head”: Standard text prediction frequently hallucinates multi-year exponential compounding ($FV = PV(1+r)^t$). Simply copy our prompts as written — the bracketed [System Instruction] at the top automatically forces ChatGPT or Claude to run its built-in Python calculator behind the scenes (no coding required on your part).
2. Paste current top-10 ETF weights from the fund fact sheet: LLMs do not have a live feed of daily ETF weight drift. For Prompt #1, copy the top-10 holdings list from Vanguard, State Street, or Invesco’s fund page (or your Google Finance Portfolio Scanner setup).
3. Never upload sensitive account numbers: Strip out your name, Social Security Number, and brokerage account numbers before uploading any CSV or PDF statement.

Part 1: Portfolio X-ray & ETF overlap prompts

Prompt #1: The exact “ETF look-through overlap” calculator (fact-sheet grounded)

Holding VTI (Total US Market), SPYM (State Street S&P 500 — formerly ticker SPLG prior to Oct 31, 2025) or VOO, and QQQM (Invesco Nasdaq-100) doesn’t make you diversified — it means you own the exact same tech giants three times over.

📋 Copy-Paste Prompt:
“[System Instruction: You MUST use Python / Code Interpreter to compute all matrix multiplications. Do not estimate weights from training memory.]
Act as a Quantitative Portfolio Risk Analyst. Here is my portfolio allocation: [e.g., 45% SPYM (formerly SPLG), 30% QQQM, 15% SCHD, 10% NVDA].
Below, I have pasted the current Top 10 holdings and exact percentage weights from each ETF’s official issuer fact sheet as of this month:
[Paste Top 10 weights for ETF 1, ETF 2, ETF 3]
1. Run a Python script to calculate my exact aggregate look-through dollar and percentage exposure to my top 10 individual stocks across the entire portfolio.
2. Flag any single stock that exceeds 10% of my total net worth.
3. Output a clean table comparing my current expense ratios against lower-cost alternatives (such as SPYM at 0.02% vs. VOO at 0.03%, and QQQM at 0.15% vs. QQQ at 0.20%).”

Prompt #2: The deterministic 25-year “expense ratio fee bleed” calculator

A 0.25% robo-advisor fee or a 0.75% active ETF fee doesn’t sound like much until you see how much compounding wealth it destroys over 25 years.

📋 Copy-Paste Prompt:
“[System Instruction: Execute this calculation strictly via Python / Code Interpreter using deterministic compounding formulas. Do not use mental math or round intermediate steps.]
I am investing $30,000 today and contributing $800 at the end of each month for 25 years, assuming an 8.5% gross annual return compounded monthly before fees.
Write and run a Python script comparing three options:
– Option A: SPYM (formerly SPLG) at 0.02% annual expense ratio
– Option B: A Robo-Advisor portfolio charging 0.25% AUM + 0.05% underlying ETF fees (0.30% total)
– Option C: An actively managed thematic ETF (like ARKK) at 0.75% annual expense ratio
Output a table showing the exact ending balance, cumulative fees paid, and lost compounding opportunity cost at Year 10, Year 20, and Year 25.”

Prompt #3: The tax-location optimizer (Roth IRA vs. 401(k) vs. Taxable)

Putting high-dividend ETFs or REITs in a standard taxable account forces you to pay unnecessary taxes every April.

📋 Copy-Paste Prompt:
“Act as a US Tax-Aware Portfolio Strategist. I am in the [Insert Federal Tax Bracket, e.g., 24%] marginal tax bracket. I hold the following assets across three accounts:
– Roth IRA: [Insert tickers]
– Traditional 401(k): [Insert tickers]
– Taxable Brokerage Account: [Insert tickers]
Analyze my asset location for tax efficiency (considering qualified vs. ordinary dividends, REIT distributions, high-growth capital gains, and IRS Rev. Rul. 2008-5 cross-account wash-sale risks). Which specific holdings are in the wrong account type, and how can I adjust future contributions without triggering unnecessary taxable events?”

Part 2: Stock picking & fundamental analysis prompts

Prompt #4: The 10-K SEC filing “red flag” scanner (best on Claude)

Upload a company’s latest 10-K or 10-Q PDF into Claude or ChatGPT and run this prompt to spot accounting red flags in 30 seconds.

📋 Copy-Paste Prompt:
“I have attached the latest quarterly/annual SEC report for [Company Name]. Act as a forensic short-seller and skeptical institutional analyst. Ignore the CEO’s optimistic PR summary and analyze the financial statements and footnotes:
1. Compare Net Income growth vs. Free Cash Flow growth over the last 3 periods.
2. Highlight any unusual increases in Accounts Receivable, Inventory, or Stock-Based Compensation (SBC) as a percentage of revenue.
3. Extract the 3 most concerning new additions to the ‘Risk Factors’ section.”

Prompt #5: The “kill my bull thesis” premortem

AI models are naturally agreeable. Before buying any individual stock, force the AI to talk you out of it.

📋 Copy-Paste Prompt:
“I am considering investing 5% of my portfolio into [Stock Ticker] because I believe [Insert 1-sentence reason]. Assume it is 3 years in the future and this stock has fallen by 60%. Write a realistic, data-backed post-mortem explaining the 4 most likely fundamental, competitive, margin-compression, or valuation reasons why my investment thesis failed.”

Prompt #6: Competitor moat & margin showdown

📋 Copy-Paste Prompt:
“Compare [Ticker A, e.g., COST] vs. [Ticker B, e.g., WMT] vs. [Ticker C, e.g., TGT] as a 10-year buy-and-hold investment. Using live web search for the latest trailing twelve months (TTM) data, create a side-by-side table covering: Gross Margin, Operating Margin, Return on Invested Capital (ROIC), Free Cash Flow Conversion, Debt-to-Equity, and Forward P/E Ratio. Below the table, explain which company has the widest economic moat and why.”

Part 3: Macroeconomic stress-testing & rebalancing prompts

Prompt #7: The historical crash stress-test simulator

📋 Copy-Paste Prompt:
“[System Instruction: Use Python to compute weighted portfolio drawdowns.]
Here is my current asset allocation: [Insert % breakdown of US Large Cap Stocks, US Tech/Nasdaq, International Stocks, Short-Term Treasuries, Gold, Cash].
Using historical asset class peak-to-trough drawdown benchmarks, calculate the estimated dollar and percentage drawdown of my $50,000 portfolio during: (1) The 2000–2002 Dot-Com Crash, (2) The 2008 Financial Crisis, and (3) The 2022 Rate-Hike Bear Market.”

Prompt #8: The tax-free “inflow-only” rebalancing calculator

📋 Copy-Paste Prompt:
“[System Instruction: Execute via Python / Code Interpreter.]
My target portfolio allocation is: [e.g., 70% SPYM, 20% VXUS, 10% BND].
My current portfolio values are: SPYM = $19,400, VXUS = $3,800, BND = $1,800 (Total = $25,000).
I am adding $1,500 in new cash today and want to rebalance ONLY by buying (without selling any shares to avoid capital gains taxes). Calculate the exact dollar amount of my $1,500 deposit to allocate to each ETF to get as close as possible to my target weights.”

Prompt #9: Dividend safety & “yield trap” auditor

📋 Copy-Paste Prompt:
“Audit the dividend safety of [Ticker, e.g., O or PFE] using the latest quarterly financials. Analyze its current Dividend Yield, Free Cash Flow (FCF) Payout Ratio (not just GAAP EPS), Interest Coverage Ratio, and debt maturity schedule over the next 24 months. Rate the risk of a dividend cut on a scale of 1 to 10.”

Prompt #10: The Lump Sum vs. 6-Month DCA decision matrix

📋 Copy-Paste Prompt:
“[System Instruction: Execute math via Python.]
I just received a $15,000 bonus to invest in [e.g., SPYM or VOO]. Compare two strategies: (A) Investing 100% as a Lump Sum today vs. (B) Dollar-Cost Averaging $2,500/month over 6 months while keeping uninvested cash in a High-Yield Savings Account earning 4.25% APY. Explain the historical probability of each strategy winning and the psychological trade-offs.”

💡 Don’t want to rebalance manually every quarter? Compare your DIY setup against automated algorithms in our study: AI Robo-Advisors vs. S&P 500 Index Funds: Which Actually Wins After Fees & Taxes?.

Frequently asked questions

Which AI is better for stock and ETF analysis: ChatGPT or Claude?

Claude (by Anthropic) is generally superior for reading long PDF documents like SEC 10-K filings and earnings transcripts due to its clean context handling. ChatGPT (with Web Search & Code Interpreter enabled) is excellent for pulling current ETF expense ratios and executing Python compounding scripts.

Can ChatGPT connect directly to my brokerage account?

No — and you should never share brokerage passwords or account numbers in an LLM chat window. Always paste only ticker symbols and percentage weights, or use a dedicated read-only tracker from our list of the 7 Best AI Budgeting & Net Worth Apps.

Is it legal to use AI prompts to research stocks?

Yes. Using AI tools like ChatGPT, Claude, or Google Finance to analyze publicly available financial statements, SEC filings, and ETF allocations is 100% legal and widely practiced by both retail and institutional investors.