15 Best Claude Prompts for Finance That Actually Work

Whether you are managing personal finances, weighing an investment, or trying to read a balance sheet without an accounting background, what you get out of Claude depends almost entirely on how much of your situation you put into the prompt. Claude prompts for finance work by supplying the context a calculator has no way to ask for.

Below are 15 prompts grouped into budgeting and debt, investing decisions, reading financial reports, and planning major decisions. Each has a short note on when to reach for it. Paste any of them into Claude, or run them in Chat Smith, an AI assistant that keeps Claude, GPT and Gemini together so you can compare how each one reasons about the same numbers.

Finance rewards clarity, precision and context-aware reasoning. Traditional tools hand you data but rarely tell you what it means for your circumstances — a prompt that includes your income, your horizon and your obligations is what closes that gap.

A long-context model such as Claude Sonnet 5 can hold a full statement, your portfolio and three scenarios in the same conversation, which is what makes real comparison possible. Three things separate a prompt that returns something you can act on from one that returns a definition.

Real numbers, not ranges — actual balances, rates and dates. Vague inputs produce advice you could have found in any article.

Ask for the assumptions — every projection rests on a return rate and a time horizon. Make Claude state them explicitly so you can argue with them.

Ask to be argued with — "build the case against this" is worth far more than "is this a good idea", which tends to come back yes.

Four prompts for the part of your finances you control directly. Have your actual figures to hand before you start, and keep an AI math calculator open to check the arithmetic yourself — which you should.

1. The personal budget framework

A budget only works if it starts from your real income and real obligations. Give both and the framework comes out usable rather than generic.

I take home [amount] per month after tax. My fixed expenses are [list them with amounts]. I have [debt balance] at [interest rate] and I want to [savings goal] within [timeframe]. Build me a detailed monthly budget using the 50/30/20 framework, adapted where my fixed costs make that split unrealistic. Then suggest how to accelerate the savings goal, and tell me which trade-off you would make first.

2. The debt payoff strategy

Multiple debts need a method rather than willpower. Avalanche and snowball produce very different totals depending on your balances and rates.

I have the following debts: [list each with balance and APR]. I can put [amount] per month toward repayment above the minimums. Compare the avalanche and snowball methods for my exact situation: show total interest paid and payoff timeline for each, tell me which you recommend and why, and flag whether consolidation or a balance transfer would beat both given these rates.

3. The spending audit

Most budgets fail on the small recurring costs nobody reviews. This turns a bank statement into a list of decisions instead of a list of transactions.

Here are my last three months of transactions: [paste them or summarise by category]. Group my spending into fixed, variable and genuinely discretionary. Identify every recurring charge, flag anything I appear to be paying for twice or not using, and rank the possible cuts by annual saving against how much I would actually notice them. Do not suggest cutting anything under [amount] per month.

4. The emergency fund sizer

"Three to six months of expenses" is a rule of thumb, not an answer. Job stability, income variability and dependants move the real number considerably.

Help me size my emergency fund. Essential monthly expenses: [amount, broken down]. My job situation: [industry, how quickly you could realistically be re-employed, whether income is fixed or variable]. Dependants: [describe]. Other safety nets: [redundancy entitlement, partner's income, credit available]. Tell me how many months I should actually target and why, what should count toward the fund and what should not, and where to hold it given I need it accessible.

Investing prompts are most useful when they argue with you rather than agree with you. Three of these four are built specifically to find the weakness in a position you already hold or are about to take.

5. The investment options comparison

Generic investment commentary is useless because it cannot know your horizon or your tolerance for a bad year. Supply both and the comparison becomes specific to you.

I am [age] with a [X]-year investment horizon and [describe] risk tolerance. I have [amount] to invest. Compare these options: [list two or three]. For each, explain the expected return range, the key risks, liquidity, and tax treatment in [your country]. Then give me a recommendation for my profile, and state the assumptions it depends on so I can check whether they hold for me.

6. The portfolio allocation check

Portfolios drift. This checks whether what you hold still matches what you intended to hold.

Here is my current portfolio: [list each holding with its value and share of the total]. I am [age], my horizon is [X years] and my risk tolerance is [describe]. Tell me what my actual asset allocation is, where I am more concentrated than I probably realise, which holdings overlap with each other, and what the allocation would look like if it matched my stated horizon and tolerance. Explain the gap between the two.

7. The fee drag audit

Fees are the one variable in investing you control with certainty. Seeing them compounded across decades tends to change behaviour.

Here are the fees on my investments: [list each holding with its expense ratio, plus any platform or adviser fee]. Total invested: [amount]. Horizon: [X years]. Calculate what I pay annually in total, and what those fees compound to across my horizon assuming [X%] annual returns. Then tell me which of these fees are normal for what they buy and which are high enough to be worth acting on.

8. The investment thesis challenger

Being talked out of a bad idea is considerably cheaper than discovering it later. Ask for the strongest case against your own position.

I am considering investing in [asset or company] because [state your thesis in full]. Do not agree with me. Build the strongest possible bear case: what would have to be true for me to lose money, which of my assumptions is doing the most work, what a well-informed sceptic would say I am underweighting, and what evidence would tell me the thesis is wrong. Then name the single piece of information that would most change the picture.

Financial reporting is written to be accurate, not necessarily to be clear. These four help you read what is actually there, and they pair well with an AI PDF summarizer for annual reports or an AI article summarizer for the coverage around them.

9. The financial statement analyser

Income statements and balance sheets are dense by design. Paste the numbers and ask for the ratios plus what they signal, rather than a description of the document.

Here is the income statement and balance sheet for [company]: [paste them]. Analyse the key ratios — gross margin, operating margin, current ratio, debt-to-equity — and tell me in plain language what they indicate about financial health. Flag any red flags, and separately list the figures you would want to see that are missing from what I have given you.

10. The earnings call decoder

Earnings calls are carefully written. The useful output is not a summary but a note on which claims are load-bearing and which are hedging.

Here is the management commentary from [company]'s latest earnings call: [paste it]. Identify the three most important facts or figures, any statement that reads as overly optimistic or that hedges around bad news, and the questions an analyst should ask to get more clarity. Then give me a plain-language summary of what management is actually saying, separated from what they are implying.

11. The peer benchmarking check

A ratio in isolation means very little. The comparisons that matter are against the industry and against the company's own history.

Here are the key financials for [company]: [paste the figures or ratios]. Benchmark them against typical values for its industry and against its own numbers from [an earlier period, if you have them]. Tell me which metrics are genuinely strong, which are weak, which look alarming but are normal for this industry, and which I should treat sceptically without more detail. Say explicitly where you are uncertain about the industry benchmark.

12. The jargon decoder

Finance jargon makes simple ideas feel inaccessible. Asking for the practical consequence alongside the definition is what makes the explanation stick.

Explain [financial term or concept] — what it is, why it happens, and why people treat it as significant. Use one simple analogy. Then explain what it means practically for someone deciding whether to [the specific decision you are facing], and tell me what it does not tell me, since these signals get over-read.

Big decisions deserve a model rather than a feeling. These three run the numbers on both paths before you commit — whether the decision is a house, a career change, or something an AI business plan generator and an AI math solver would help you cost out properly.

13. The financial plan stress test

A plan that only works in the good case is not a plan. Downside scenarios show you the weak point while it is still cheap to fix.

Here is my current financial position: [income, savings, investments, debt, monthly expenses]. Stress-test it against three scenarios: (1) I lose my income and it takes six months to replace, (2) markets fall [X%] next year, (3) I face an unexpected [amount] expense. For each, tell me how long my finances hold, exactly where the plan breaks, and the single change now that would most improve resilience.

14. The decision simulation

Major decisions are far easier to judge when both paths are modelled to the same horizon under stated assumptions.

I am deciding between [option A] and [option B] over the next [X years]. Option A: [all the relevant figures]. Option B: [all the relevant figures]. Assuming [X%] annual returns on anything invested, model both over the period and tell me which comes out ahead financially, under what assumptions the answer flips, and which single assumption the conclusion is most sensitive to. Show the arithmetic so I can check it.

15. The adviser meeting preparer

An hour with an adviser goes much further when you arrive organised and know which questions matter. It also makes it easier to tell a good adviser from a salesperson.

I am meeting a financial adviser for the first time. My situation: [age, household, income, retirement savings, other investments, debt, insurance, dependants]. Help me organise this into a clear one-page picture, generate 10 sharp questions I should ask — including about how they are paid — and explain the concepts I am likely to encounter so I am not lost. Flag anything in my situation an adviser will probably raise that I have not thought about.

Specificity is the whole game. Real balances, real rates, real dates — a prompt left full of placeholders returns a template, which is the thing you were trying to escape.

Check the arithmetic. Language models are considerably better at framing a financial question than at computing one, so treat any figure that comes back as something to verify rather than something to act on. Asking for the formula and the assumptions alongside the number lets you check both.

Match the model to the task. Batch work — categorising transactions, drafting questions, explaining terms — runs fine on a light model such as Claude Haiku 4.5. Stress-testing a plan or arguing against your own investment thesis wants the strongest reasoning you have access to.

And the point worth stating plainly: none of this is financial advice, and Claude does not know your full circumstances, your tax position or the rules in your jurisdiction. These prompts are for arriving at better questions and a clearer view of the trade-offs. Decisions with tax, legal or long-term consequences still warrant a qualified professional.

Each prompt targets a specific failure. Budgeting from an estimate of your spending rather than the statement builds the plan on a number you invented. Paying down the smallest balance first feels satisfying and often costs more than clearing the highest rate would have. Holding six funds that track the same index is concentration dressed up as diversification. Ignoring fees because they are quoted in fractions of a percent hides the largest controllable drag on a long-term return. And judging a plan only against the case where nothing goes wrong is the mistake that turns an ordinary setback into a crisis.

Budgeting, investing, statement analysis and scenario planning each want a different prompt, and the ones that match how you actually manage money are worth keeping. Chat Smith lets you save any prompt here as a reusable template, group them by purpose, and launch one in a click — which matters most for the ones you run monthly rather than once.

It is a multi-model assistant, so the same question can run across Claude, GPT, Gemini, Grok and DeepSeek from one model library and the answers compared. For financial questions that is more than convenience: when two models model the same rent-versus-buy scenario and reach opposite conclusions, the assumption they disagree about is usually the one your decision actually turns on.

Start with whichever prompt sits closest to a decision you are facing this month — the stress test if you already have a plan, the spending audit if you do not.