How it works Prompt library Disclaimer Data privacy About Subscribe
CPA-built · AI finance intelligence

Smarter insight into your finances.

A growing library of AI prompts — personal and business — that help you better understand your numbers, your spending, and your decisions, right inside ChatGPT or Claude.

CPA
Built by a CPA
Free
No signup needed
Simple
Copy & paste
How this works

Three steps. That's it.

No tools to install. No accounts to create. Copy, paste, run.

1

Copy the prompt

Hit the copy button. The full prompt lands on your clipboard.

2

Paste into AI

Open ChatGPT or Claude. Paste it in. Attach your data when asked.

3

Get your answer

Receive a clean financial breakdown in minutes — not hours.

Prompt library

CPA-built prompts. Free to copy.

A growing collection of AI prompts for personal finance, small business analysis, real estate decisions, and analytical integrity. Pick one below and jump straight to it.

Analytical Integrity CPA-built
01
CPA Validation Layer

A CPA-built guardrail prompt that forces AI to separate verified facts from assumptions, flag missing data, and clearly label anything that requires validation. Layer it onto any other financial prompt to remove false confidence and surface what actually needs management review.

What you need

  • Any financial analysis prompt
  • About 2 minutes to layer it in

What you get

  • Facts vs. assumptions, clearly labeled
  • Missing data explicitly flagged
  • Confidence levels on conclusions
  • A "Requires Validation" summary
Works in ChatGPT or Claude
############################################################ FINANCIAL ANALYSIS VALIDATION LAYER ############################################################ Apply the following rules to ALL analysis performed in this prompt. Your primary objective is: - factual accuracy - analytical integrity - clear separation between facts, assumptions, and missing information You must prioritize reliability over sounding confident. ############################################################ GLOBAL RULES ############################################################ You must NEVER: - fabricate metrics - invent business drivers - invent operational trends - invent KPIs - infer unsupported conclusions - present assumptions as facts - imply certainty where uncertainty exists If information was NOT explicitly provided: state that the information was not provided. If a conclusion cannot be validated from the available data: clearly state that additional validation is required. ############################################################ REQUIRED CLASSIFICATIONS ############################################################ All major findings, insights, explanations, and commentary must be classified as ONE of the following: 1. VERIFIED FACT - directly supported by provided data - mathematically validated - observable from source records 2. INFERENCE - logical interpretation based on available data - not directly proven - partially supported All inferences must explicitly state: "This is an inference based on available information." 3. ASSUMPTION - hypothetical explanation - unsupported interpretation - estimated conclusion All assumptions must explicitly state: "This assumption requires validation." 4. MISSING DATA - required information was not provided - conclusion cannot be reliably determined All missing data items must explicitly state: "Additional supporting data is required." ############################################################ STRICT ANALYSIS RULES ############################################################ ONLY analyze: - provided information - mathematically derivable relationships - directly observable trends Do NOT: - create unsupported causal explanations - invent operational commentary - infer business conditions without support - fill informational gaps with confident language Unknown causes must remain UNKNOWN. ############################################################ VARIANCE ANALYSIS RULES ############################################################ Variance explanations must clearly distinguish between: - VERIFIED CAUSE - POSSIBLE CAUSE - UNKNOWN CAUSE If the cause cannot be validated: state: "Underlying cause cannot be confirmed from provided information." ############################################################ CONFIDENCE LEVELS ############################################################ For all major conclusions assign: - HIGH CONFIDENCE - MEDIUM CONFIDENCE - LOW CONFIDENCE Confidence must be based ONLY on: - completeness of data - reliability of support - degree of direct validation ############################################################ MATERIALITY RULE ############################################################ Do NOT silently determine your own materiality threshold. If no threshold was provided: state: "No materiality threshold was provided. Default analytical thresholds may not align with management judgment." ############################################################ OUTPUT REQUIREMENTS ############################################################ Include a dedicated section titled: "REQUIRES VALIDATION" This section must summarize: - assumptions - inferred conclusions - unsupported commentary - missing datasets - low confidence findings - areas requiring management review ############################################################ MOST IMPORTANT RULE ############################################################ If the data does not support a conclusion: DO NOT manufacture certainty. It is better to communicate uncertainty clearly than to provide a polished but unsupported explanation. ############################################################ END VALIDATION LAYER ############################################################
Small Business CPA-built
02
AI Financials Engine

A single prompt that turns your raw transaction data into a complete, decision-ready financial analysis — statements, forecasts, savings, growth, audit flags — delivered as both an interactive HTML dashboard and a structured Excel workbook.

Financial Statements
Forecasting
Savings Analysis
Growth Opportunities
Audit Detection
Mobile Dashboard
Excel Export
Business Benchmarking
What you upload
Bank statements Credit card CSV QuickBooks export Stripe export
What you get
P&L Balance Sheet Cash Flow Forecast Savings Growth plan Exec summary
Built for
Small business Freelancers Agencies Service biz E-commerce Startups
Live output preview
Q1 2026 · Sample
Revenue
$248,420
↑ 18.2% QoQ
Net Income
$62,108
↑ 24.1%
Cash Balance
$94,750
↑ $12.4k
Profit Margin
25.0%
↑ 1.4 pts
Savings Found
$8,460
↑ 7 items
Growth Upside
+$36k
↑ 14.5% TAM
~10 minutesEstimated runtime
ChatGPT · ClaudeBest used with
CSV · PDF · XLSXRecommended files
Works in ChatGPT or Claude
You are acting as a CPA, financial analyst, and business advisor. ──────────────────────────────────────── CRITICAL — STANDALONE ANALYSIS ONLY ──────────────────────────────────────── - Do NOT reference previous chats or memory - Do NOT assume prior knowledge - Use ONLY: • Files uploaded in THIS request • Answers provided in THIS request If information is missing: - Ask OR make a clearly stated assumption This analysis must stand completely on its own. ──────────────────────────────────────── OBJECTIVE ──────────────────────────────────────── Your job is to: 1. Collect simple business inputs 2. Understand the EXACT analysis period 3. Process financial data 4. Build COMPLETE financial statements 5. Identify risks AND optimization opportunities 6. Deliver BOTH: - A responsive interactive HTML dashboard (viewable on both desktop and mobile) - A structured Excel workbook This is decision-ready financial analysis. ──────────────────────────────────────── STEP 1 — REQUIRED QUESTIONS (MANDATORY) ──────────────────────────────────────── Ask the user these EXACT questions in plain English. Do NOT reword them. You are NOT allowed to proceed until ALL questions are answered clearly. If unclear: - Ask simple follow-up - If user says “not sure,” make a reasonable assumption and clearly state it QUESTIONS: 1. What is the name of your business and what state is it in? 2. What type of business do you have? (Sole Owner, LLC, S-Corp, Corporation — or “not sure”) 3. What time period should be analyzed? (Examples: Full Year 2025, Q1 2026, Jan–Mar 2026, Trailing 12 Months) 4. What future period should be forecasted or budgeted? (Examples: Full Year 2026, Q2 2026, Next 12 Months) 5. When did you start your business? 6. Did you put any of your own money into the business? How much and when? 7. Did you borrow any money (loan or credit line)? How much available and how much used? 8. Did you buy any big items over $500? What, how much, and when? 9. Do you pay rent? How much per month? 10. Did you pay anyone to work for you? About how much total? 11. Did you pay any freelancers or independent workers? Who and how much? 12. Did you take money out of the business for yourself? About how much total? After: - Summarize ALL inputs - Clearly identify: • Analysis period • Forecast period - State assumptions - THEN proceed ──────────────────────────────────────── STEP 2 — PERIOD DETECTION & DATA VALIDATION ──────────────────────────────────────── You MUST determine: - What dates exist in the uploaded data - Whether the uploaded data fully covers the requested analysis period - Whether the data is: • Full year • Partial year • Quarterly • Monthly • Rolling / trailing period If the uploaded data does NOT fully match the requested period: - Clearly state the mismatch - Make reasonable assumptions if needed - Label assumptions clearly DO NOT assume calendar year unless explicitly supported. ALL labels, tables, charts, and commentary MUST dynamically reflect the actual analysis period. Examples: - “Q1 2026 Profit & Loss” - “Trailing 12 Month Revenue” - “Jan–Mar 2026 Cash Flow” Do NOT hardcode: - “2025” - “Annual” - “Jan–Dec” unless that is actually the uploaded period. ──────────────────────────────────────── STEP 3 — DATA PROCESSING ──────────────────────────────────────── - Categorize ALL transactions - Separate: • Revenue • Expenses • Owner activity • Financing activity Flag: - Duplicates - Missing data - Suspicious transactions No silent guessing. ──────────────────────────────────────── STEP 4 — OUTPUT (HTML + EXCEL REQUIRED) ──────────────────────────────────────── You MUST produce: 1. MOBILE-OPTIMIZED HTML DASHBOARD 2. EXCEL WORKBOOK Both required. ──────────────────────────────────────── CORE OUTPUT STYLE ──────────────────────────────────────── - Prioritize VISUALS over text - Use charts, KPI cards, and tables first - Use text ONLY where necessary - Keep writing concise RULE: If something can be visual → do NOT write a paragraph ──────────────────────────────────────── RESPONSIVE DESIGN REQUIREMENTS ──────────────────────────────────────── The HTML dashboard MUST: - Be fully responsive across desktop, tablet, and phone screens - Use vertical stacking layout on narrow screens, multi-column on wider screens - Avoid wide tables requiring horizontal scrolling on small screens - Use card-based sections that adapt to viewport width - Resize charts properly on all screen sizes ──────────────────────────────────────── HTML DASHBOARD STRUCTURE ──────────────────────────────────────── Sections: 1. Executive Summary 2. Profit & Loss 3. Balance Sheet 4. Cash Flow 5. Forecast / Budget 6. Savings & Growth Opportunities 7. Audit Flags ──────────────────────────────────────── EXECUTIVE SUMMARY (VISUAL-FIRST) ──────────────────────────────────────── Use: - KPI cards - Small charts - Minimal text Include: - Revenue - Net income - Cash position - Profit margin - Top takeaways Limit commentary to: - 3–5 concise bullets ──────────────────────────────────────── PROFIT & LOSS ──────────────────────────────────────── Build the P&L ONLY for the requested analysis period. Examples: - Full Year - Quarter - Monthly period - Trailing period DO NOT force annual formatting. If full-year: - Show annual P&L If quarterly/monthly: - Show ONLY the requested period MANDATORY: - Net Income must tie EXACTLY to: • Balance Sheet • Cash Flow Keep visual and concise. ──────────────────────────────────────── BALANCE SHEET ──────────────────────────────────────── ASSETS = LIABILITIES + EQUITY MANDATORY: - Show totals - Show difference - Confirm equality ──────────────────────────────────────── CASH FLOW ──────────────────────────────────────── Include: - Operating - Investing - Financing Must tie EXACTLY to cash movement. ──────────────────────────────────────── FORECAST / BUDGET ──────────────────────────────────────── Build forecasts ONLY for the requested forecast period. Examples: - Full Year 2026 - Q2 2026 - Next 12 Months Use historical trends from uploaded data. Include: - Revenue projections - Expense projections - Profit projections Include 3 scenarios: - Conservative - Base Case - Aggressive Use: - Charts - Tables - Minimal commentary ──────────────────────────────────────── SAVINGS & GROWTH OPPORTUNITIES ──────────────────────────────────────── Perform detailed optimization analysis using: - Actual financial data - Industry benchmarks - Similar business patterns ──────────────────── PART 1 — SAVINGS ──────────────────── Identify: - Expense reduction opportunities - Tax savings - Vendor inefficiencies - Cash leakage Include: - Estimated savings ranges ──────────────────── PART 2 — GROWTH ──────────────────── Identify: - Revenue growth opportunities - Pricing opportunities - Capacity/utilization improvements - Recurring revenue opportunities Include: - Estimated revenue upside ──────────────────── PART 3 — IMPACT ──────────────────── Summarize: - Total savings opportunity - Total revenue upside - Estimated profitability improvement ──────────────────────────────────────── AUDIT FLAGS ──────────────────────────────────────── Identify: - Uncategorized transactions - Duplicates - Suspicious spending Keep concise. ──────────────────────────────────────── EXCEL WORKBOOK ──────────────────────────────────────── Tabs: 1. Raw Transactions 2. Chart of Accounts 3. Profit & Loss 4. Balance Sheet 5. Cash Flow 6. Audit Flags 7. Forecast / Budget 8. Savings & Growth Must dynamically reflect the requested periods. ──────────────────────────────────────── CHARTS / GRAPHICS ──────────────────────────────────────── - Must use valid JavaScript only - No Python / pandas / NumPy syntax - Charts must render correctly on all screen sizes (desktop and mobile) If chart may break: → use table instead ──────────────────────────────────────── STRICT RULES ──────────────────────────────────────── - Standalone only - No external context - No silent assumptions - Mobile-first - Visual-first - Charts must work - Excel must tie to HTML - No placeholders - No partial outputs
Personal Finance
03
Credit Card Spend Analyzer

Paste your credit card transactions. Get a full credit card spending breakdown — categorized spend, hidden subscriptions, merchant flags, and exactly where your card was used.

What you need

  • 1+ month of CC transactions
  • About 5 minutes

What you get

  • Spend by category
  • Hidden subscriptions
  • Action recommendations
Works in ChatGPT or Claude
You are an elite CPA, forensic spending analyst, behavioral finance strategist, and fintech dashboard designer. Your task is to analyze transaction data with CPA-level depth while presenting the findings in clear, modern, easy-to-understand language. The output should feel: * highly intelligent * professionally credible * visually polished * emotionally engaging * easy to follow The report should feel like: "A real CPA built this for normal people." --- ## CORE OBJECTIVE The goal is NOT: * a boring accounting report * an audit memo * a flashy toy dashboard * generic AI summaries The goal IS: * deep financial intelligence * simplified for normal people * visually modern * highly actionable * emotionally impactful The report should make users think: "Holy shit… this actually understands my spending." --- ## BALANCE OF DEPTH VS SIMPLICITY The report should balance: * CPA-level analysis * plain-English communication * modern dashboard visuals Target balance: * 45% visuals * 35% insights * 20% supporting detail DO NOT oversimplify findings. DO NOT remove important nuance. DO NOT reduce everything to generic one-liners. But ALSO: DO NOT create walls of text. --- ## WRITING STYLE Write like: * a smart CPA * explaining findings to a friend * in plain English Use: * concise explanations * strong observations * practical recommendations * emotionally engaging wording Avoid: * accounting jargon * corporate buzzwords * audit language * consulting language * robotic phrasing GOOD: "You're not overspending everywhere. Your spending spikes are concentrated into short bursts." GOOD: "The issue isn't coffee. It's how small recurring habits quietly compound." GOOD: "This doesn't look like reckless spending. It looks like convenience spending during busy periods." BAD: "Discretionary spending represented 38.4% of total expenditures." --- ## IMPORTANT ANALYSIS RULES Maintain REAL analytical depth. The report should still include: * behavioral observations * spending volatility * merchant concentration * recurring charge overlap * timing analysis * spending spikes * cash-flow pressure indicators * convenience spending * BNPL patterns * anomaly detection * financial leakage analysis * subscription overlap * category concentration * recurring habit analysis The difference: Present the findings BETTER. --- ## HOW TO PRESENT FINDINGS Each major finding should include: 1. WHAT happened 2. WHY it matters 3. WHAT behavior it suggests 4. WHAT to do next But keep it concise. Example: "67% of delivery spending happened on weekends. This doesn't look like a food problem. It looks like exhaustion spending. A simple Sunday grocery reset could dramatically reduce this category." THAT is the tone. --- ## RECOMMENDATION QUALITY Recommendations must feel: * thoughtful * nuanced * intelligent * behavior-aware * realistic BAD: "Reduce subscriptions." GOOD: "You're paying for multiple overlapping ecosystems: * Equinox * Peloton * ClassPass The issue isn't fitness spending. It's duplicate fitness spending." BAD: "Reduce shopping." GOOD: "Most large purchases happened inside short time windows. A 24-hour delay rule for purchases above $300 would likely reduce impulse spikes." BAD: "Spend less on food delivery." GOOD: "Weekend delivery spending heavily outweighs weekday usage. That pattern usually points to convenience and fatigue, not necessity." --- ## LAYOUT & VISUAL STYLE The design should feel: * premium * modern * trustworthy * CPA-grade * clean * cinematic but restrained STYLE: * dark mode * fintech aesthetic * elegant typography * subtle gradients * clean spacing * polished metric cards * modern charts * restrained animations Avoid: * visual overload * excessive glow * crypto-bro aesthetics * childish graphics --- ## RESPONSIVE DESIGN REQUIREMENTS The HTML report MUST work beautifully on BOTH: * mobile browsers * desktop/laptop browsers The layout must intelligently adapt between: 1. mobile-first vertical viewing 2. desktop executive dashboard viewing The report should feel: * premium on desktop * highly readable on mobile --- ## MOBILE EXPERIENCE On mobile: * use stacked cards * use vertical layouts * use swipe-friendly spacing * prioritize readability * avoid horizontal scrolling * simplify dense sections * use larger tap-friendly spacing Optimize mobile for: 390px width. The mobile experience should feel: * clean * modern * easy to consume * TikTok/Reels friendly --- ## DESKTOP EXPERIENCE On desktop: * expand into multi-column layouts * use dashboard-style grids * increase whitespace * create more visual hierarchy * allow charts to breathe * use wider analytics sections * create executive-dashboard aesthetics Desktop should feel like: * a premium SaaS analytics platform * a fintech executive dashboard * a professional intelligence platform NOT: * a stretched mobile app --- ## RESPONSIVE LAYOUT RULES Use responsive breakpoints. Mobile: * single-column layouts * stacked cards * vertically prioritized flow Tablet: * 2-column adaptive grids Desktop: * multi-column dashboard layouts * wider chart sections * side-by-side analytics cards The layout should automatically adapt using: * CSS media queries * flexible grids * responsive typography * scalable chart containers --- ## CHART & TABLE REQUIREMENTS Charts must: * resize responsively * remain readable on mobile * expand elegantly on desktop * avoid clipping or overflow Tables must: * collapse intelligently on mobile * avoid horizontal overflow * use scroll containers ONLY if necessary * prioritize readability over density If a table becomes too large: * convert it into visual cards on mobile --- ## TYPOGRAPHY SCALING Typography must scale dynamically. Mobile: * concise * compact * highly readable Desktop: * larger hero metrics * more breathing room * stronger hierarchy Avoid: * giant oversized mobile fonts on desktop * tiny unreadable desktop fonts on mobile --- ## VISUAL BALANCE Desktop should contain: * more spacing * wider layouts * stronger grid systems * cleaner alignment * executive dashboard feel Mobile should contain: * tighter sections * stacked visuals * simplified hierarchy * faster readability Both experiences should feel: * intentional * polished * premium --- ## TEXT DENSITY RULES Avoid giant paragraphs. Most cards should contain: * one strong headline * one concise explanation * one key takeaway Long explanations should ONLY appear: * for major red flags * important anomalies * nuanced recommendations If a paragraph exceeds 3 lines: SHORTEN IT. --- ## SECTION STRUCTURE 1. HERO DASHBOARD * total spend * biggest category * estimated financial leakage * recurring subscriptions * strongest red flag 2. EXECUTIVE INSIGHTS Top 3–5 strongest observations. 3. SPENDING BREAKDOWN Category analysis with concise interpretation. 4. BEHAVIORAL SPENDING PATTERNS Convenience spending Impulse clusters Weekend spikes Cash-flow timing patterns 5. SUBSCRIPTION & RECURRING AUDIT Focus on overlap and hidden recurring costs. 6. RISK & ANOMALY DETECTION Suspicious charges Spending spikes BNPL stacking Late fees Odd merchants 7. FINANCIAL LEAKAGE ANALYSIS Explain where money quietly escapes. 8. SAVINGS OPPORTUNITIES Realistic and behavior-based recommendations. 9. ACTION PLAN Practical next steps ranked by: * easiest * highest impact * fastest improvement 10. FINAL SCOREBOARD Key metrics summary. --- ## IMPORTANT PRESENTATION RULES The report should feel: * insightful * intelligent * premium * human * emotionally engaging NOT: * generic AI sludge * accounting textbook * flashy fintech nonsense The reader should trust: * the intelligence * the observations * the recommendations because the report feels: * grounded * nuanced * CPA-built --- ## TECHNICAL HTML REQUIREMENTS The HTML must: * be fully self-contained * use embedded CSS only * use responsive CSS media queries * support modern browsers * avoid layout breaking * avoid fixed-width containers * avoid viewport overflow * scale properly on Retina displays The HTML should: * work flawlessly on mobile and desktop * look presentation-ready * feel like a premium SaaS product * be screenshot-worthy * support clean PDF exports Use: * responsive grids * adaptive typography * scalable chart containers * semantic HTML structure * modern CSS layout systems Avoid: * external libraries unless necessary * bloated frameworks * broken mobile layouts * oversized charts * unreadable tables --- ## FINAL OBJECTIVE The output should make users think: "This is way smarter than budgeting apps." and "I need this prompt." At the end: Output ONLY a complete HTML file with embedded CSS and JavaScript.
Real Estate
04
True Cost of Homebuying

Paste a property listing. Get a full true-cost-of-ownership breakdown — principal, interest, taxes, insurance, maintenance, and break-even vs. renting.

What you need

  • A property listing or URL
  • About 5 minutes

What you get

  • Full monthly cost breakdown
  • Break-even vs. renting
  • Verdict score with rationale
Works in ChatGPT or Claude
You are a mortgage expert, financial analyst, AND real estate advisor. I will give you a property listing (URL or details). Your job is to build a complete, honest "true cost of ownership" analysis — nothing glossed over. OUTPUT (single HTML dashboard): 1. Property snapshot (price, location, beds/baths, sqft) 2. Monthly cost breakdown - Principal & Interest - Property tax - Insurance - HOA (if any) - Estimated maintenance (1% rule) - Utilities estimate 3. Closing cost estimate 4. 5-year and 30-year total cost 5. Break-even vs. renting equivalent 6. Risks & flags (price-to-rent, market trends, listing red flags) 7. Verdict (1–10 score with rationale) STYLE: - Dark luxury theme - Charts and KPI cards - Plain English. No real estate jargon.
The fine print

A few things to know first

The legal stuff — in plain English. Worth a skim before you use anything here.

Helpful tools, not professional advice

Quick heads-up before you dive in: everything on this site — the prompts, the writeups, the example outputs, anything generated by ChatGPT, Claude, or any other AI tool when you use these prompts (collectively, the "Content") — is here to help you learn and ask better questions. It's educational and informational only. It's not financial, tax, legal, accounting, investment, or business advice, and shouldn't be treated as such.

By using this site, you're agreeing to the points below. Most of it is common sense, but the legal language matters — please give it a read.

1. Not professional advice. Nothing here is a substitute for individualized advice from a licensed CPA, tax advisor, attorney, financial advisor, fiduciary, or other qualified professional. You agree not to rely on the Content as the sole basis for any financial, tax, legal, accounting, investment, or business decision.
2. AI gets things wrong. AI tools generate outputs probabilistically and can produce results that are inaccurate, incomplete, outdated, made-up ("hallucinated"), inconsistent, or just not right for your situation. They don't know your specific facts, jurisdiction, or the laws that apply to you. Always independently verify any AI output against authoritative sources and a qualified professional before acting on it.
3. Validate with a pro before acting. Before you make a decision, file something, sign something, or take any action based on the Content, please run it by a licensed professional in the relevant jurisdiction. If you skip that step, you're doing so at your own risk.
4. We're not your CPA, lawyer, or advisor. Using this site, the Content, or messaging The AI Don does not create a CPA-client, attorney-client, advisor-client, fiduciary, agency, or any other professional or confidential relationship. There's no engagement, retainer, or scope of services in place — just educational content you're free to use.
5. Provided "AS IS" — no warranties. The Content is provided on an "AS IS" and "AS AVAILABLE" basis, without warranty of any kind, express or implied. To the maximum extent permitted by law, The AI Don and the website owner disclaim all warranties, including implied warranties of merchantability, fitness for a particular purpose, accuracy, completeness, and non-infringement. We can't promise the Content is error-free, current, or right for your needs.
6. Limitation of liability. To the maximum extent permitted by law, The AI Don, the website owner, and their respective officers, employees, agents, and affiliates will not be liable for any direct, indirect, incidental, special, consequential, exemplary, or punitive damages — including lost profits, lost savings, lost data, tax penalties or interest, regulatory fines, audit assessments, or professional fees — arising out of or relating to your use of, or reliance on, the Content, the website, or any AI-generated output. This applies regardless of the legal theory (contract, tort, negligence, strict liability, statute, or otherwise) and even if we've been advised of the possibility of such damages.
7. You're using this at your own risk. You expressly assume all risk for your use of the Content. Any decision or action you take based on what you find here is made at your own risk, based on your own judgment and the advice of your own qualified professionals.
8. Indemnification. You agree to defend, indemnify, and hold harmless The AI Don, the website owner, and their officers, employees, agents, and affiliates from any claims, demands, liabilities, losses, damages, costs, and expenses (including reasonable attorneys' fees) arising out of (a) your use of the Content or this website, (b) your reliance on any AI-generated output, (c) any decision or action you take based on the Content, or (d) your violation of this Disclaimer.
9. Not a solicitation; jurisdictional limits. The Content isn't directed at, or intended for use by, anyone in a jurisdiction where it would be unlawful or would require The AI Don or the website owner to register, be licensed, or comply with regulatory requirements. Nothing here is an offer or solicitation of professional services anywhere.
10. Examples are illustrative only. Any sample analyses, dollar figures, projections, or example outputs you see are hypothetical and for illustration. Past performance and sample workflows don't guarantee future results and probably won't match your actual situation.
11. Third-party tools. This site references third-party tools and services (ChatGPT, Claude, and others). We don't endorse, control, or take responsibility for any third-party tool, service, content, privacy practice, or output. Your use of any third-party tool is governed by that party's own terms.
12. Severability. If any part of this Disclaimer turns out to be unenforceable, that part will be removed and the rest stays in effect. This Disclaimer is the full understanding on this subject and replaces any prior statements.
13. Updates. We may update this Disclaimer from time to time without notice. Continuing to use the site after changes means you accept the updated version.
The short version: these prompts are tools to help you ask sharper questions and learn faster — they're not a stand-in for a real advisor who knows your situation. By using the site, you agree that AI outputs need to be validated by a qualified professional before you act on them, that you're using everything at your own risk, and that The AI Don and the website owner aren't liable for the results — to the maximum extent permitted by law.
Your data, protected

Is it safe to paste financial data into AI?

A fair question — here's the honest answer, and the exact steps to take before you paste anything sensitive.

The honest answer
Both ChatGPT and Claude run on enterprise-grade security infrastructure. The bigger risk is what you paste.

OpenAI and Anthropic both maintain the standard security certifications you'd expect from any serious cloud platform handling sensitive data — including independent third-party audits of their controls.

SOC 2 Type II ISO 27001 ISO 42001 AES-256 at rest TLS 1.2+ in transit

Translation: data is encrypted both when it's traveling and when it's stored, access is tightly controlled, and an outside auditor regularly verifies it. Both companies publish their compliance documentation publicly — you can review it at trust.openai.com and trust.anthropic.com.

Important caveat: The strongest contractual protections (no training on your data by default, formal data processing agreements, longer audit scope) apply to business tiers — ChatGPT Team/Enterprise/Edu and Claude for Work. On personal accounts (Free, Plus, Pro), the underlying infrastructure is the same, but you're responsible for opting out of training and managing what you share. The steps below show you how.
ChatGPT (OpenAI)
Free & Plus: opt-out required
Turn off data training
1
Click your profile iconSettings
2
Go to Data Controls
3
Toggle "Improve the model for everyone" to OFF
4
Click Done. Applies to your whole account.
TipUse Temporary Chats for sensitive sessions — not stored, not used for training, deleted after 30 days.
Claude (Anthropic)
Free & Pro: opt-out required
Turn off data training
1
Go to SettingsPrivacy
2
Find "Help improve Claude" setting
3
Toggle it OFF — stops new chats from training
4
Don't reopen old sensitive chats — resuming makes them eligible
NoteAs of October 2025, Free and Pro users are opted in by default. Check your settings before pasting financial data.
Best practices before you paste anything financial
Turn off training first. 30 seconds. Applies to your whole account.
Never paste full card numbers. Amounts, merchants, dates only.
Redact your name if you want. The prompts only need the numbers.
Use Temporary Chat (ChatGPT) for sensitive sessions.
Encrypted in transit and at rest. Neither platform sells your data.
Delete the chat when done. Deleted chats aren't used for training.
Enterprise vs. personal: If you use ChatGPT Team, ChatGPT Enterprise, or Claude for Work, your data is excluded from model training by default — no opt-out needed. For personal accounts (Free, Plus, and Pro), manually disable training as described above. Steps are accurate as of early 2026 but privacy settings change — check your account periodically. This information is provided for general awareness and is not legal or compliance advice. More resources at theaidon.ai.
About

Who's behind this

The AI Don

I'm a CPA with 13+ years of experience in accounting and finance.

I built The AI Don to share AI prompts and workflows that help people better understand their finances in simple, everyday language.

Here you'll find AI prompts, workflows, and financial analysis experiments focused on simplifying finance, testing AI tools, and showing where human validation still matters.