ChatGPT prompts for a financial analyst interview work best when they make the AI quiz you on accounting links, valuation and variance analysis one question at a time, then check your reasoning instead of doing the work for you. The 13 prompts below work in ChatGPT, Gemini, Claude and Microsoft Copilot, and a worked variance example shows how to check the AI's math.
Chatbots still make arithmetic and sign errors, so recompute every number before you rely on it. Replace the [square brackets] with your details.
What to paste, and what to keep confidential
Paste the job description (FP&A, corporate finance, equity research and credit roles get different questions), your résumé and the tools you've really used, such as Excel, a BI tool or an ERP system. Our financial analyst interview questions guide describes how these interviews usually run.
Leave out anything confidential from a current or past employer: budgets, forecasts, unreleased results, client names and deal details. Use rounded or made-up figures. If the employer gives you an Excel test or a take-home model, do it yourself under its rules; these prompts are for the weeks before. The approach is the one in our guide on how to use AI for interview prep: the AI asks and checks, you answer.
Technical drills
1. How changes flow through the three statements
Quiz me on how changes flow through the income statement, balance sheet and cash flow statement. Give me one change at a time (for example depreciation up 10, inventory up 20, or a 50 equipment purchase funded with debt), assume a [25]% tax rate, and wait for my answer. Then show the correct effect on each statement and point to the exact step I missed, if any.
This is one of the most common technical questions. Check the AI's answer too: with depreciation up 10 at a 25% tax rate, net income falls 7.5 and cash rises 2.5.
2. DCF walk-through with follow-ups
Ask me to walk you through a DCF. After my answer, ask four follow-ups one at a time: why unlevered free cash flow, how I'd choose a terminal growth rate, what happens to value if WACC rises one point, and how I get from enterprise value to equity value. Score each answer for accuracy and clarity, not length.
The follow-ups separate memorized outlines from understanding. Practice the core walk-through until it takes under two minutes.
3. Price and volume variance
Create a revenue variance question: give a budget and an actual for units and price, and ask me to split the total variance into volume and price effects. Don't show the answer until I reply. When you check my work, show the formulas and confirm that the volume and price variances add up to the total.
The 'add up' check catches most AI and human errors. See the worked example below.
4. Ratios from a small table
Give me two years of figures for a fictional company in a small table: revenue, gross profit, operating income, receivables, inventory, current assets and current liabilities. Ask me to calculate three ratios of your choice and explain what changed and why it matters. Check my math line by line.
Interviewers care more about the explanation than the formula, so make the AI ask 'So what?' after each ratio.
5. Excel and tools, honestly
Ask me six questions a financial analyst hiring manager might ask about Excel and tools: lookups, pivot tables, scenario analysis, error checking, and the BI or ERP tools I've used: [tools]. For each, tell me whether my answer sounds hands-on or read-about, and suggest one concrete detail I could add if it's true.
Claiming a tool you can't use is risky if there's a live test. Lead with what you've actually built.
Business-partner and behavioral prompts
6. Explain a miss to a non-finance manager
Play a sales director who dislikes finance jargon. I'll explain why [region] missed its revenue budget by [10]%. Interrupt whenever I use jargon, and end by asking 'So what do you want me to do?' Rate me on plain language, the cause and a clear recommendation.
Analysts are hired to make numbers useful to people outside finance, and this drill tests exactly that.
7. The model error story
Help me prepare 'Tell me about a time you found an error in a model or report.' My notes: [what happened, how I found it, what I changed]. Ask follow-ups about how I caught it, who I told and what control I added afterwards. Ask me for missing details instead of inventing them.
A good answer ends with a control you still use, such as a reconciliation check, not just the fix.
8. An analysis that changed a decision
Turn my notes into a 90-second STAR answer: [notes on an analysis I did]. Lead with the business question, keep numbers rounded and non-confidential, and end with the decision and its result. Mark any number you didn't get from me.
Marked numbers are the ones to check or delete; interviewers will ask how you got each figure.
9. 'Why this company?' from its results
Using [company]'s latest public results that I've pasted here: [paste], help me write a 30-second answer to 'Why do you want to work here?' that names one business driver I'd want to analyze and why. Don't use any figure that isn't in my paste.
One specific driver beats general praise, and limiting the AI to your paste keeps the figures accurate.
Cases, mock rounds and final review
10. Revenue-below-forecast case
Run a 15-minute case: revenue came in [8]% below forecast this quarter for a [industry] company. Give me data only when I ask, in small tables. Expect me to break it down by product, region, price and volume. End by asking for a 60-second recommendation, then score structure, math and communication from 1 to 5.
Time it yourself; the chatbot won't stop the clock, and real cases are short.
11. Hiring-manager mock
Interview me as an FP&A manager hiring a financial analyst. Ask seven questions one at a time: three technical, two behavioral, one forecasting judgment question and 'Why this role?' Wait for each answer. At the end, list any technical answer with an error and give the correct explanation.
Compare its corrections with a textbook or our FP&A analyst interview questions before you accept them.
12. Stock pitch for investment roles
I'm interviewing for a financial analyst role in [equity research / asset management]. Ask me to pitch a stock I follow in two minutes, then challenge my valuation, the key risks and what would prove me wrong. Don't supply facts about the company; I'll bring my own. Tell me where my numbers need a source.
Pitches fall apart on sourcing, and the rule keeps the chatbot from filling gaps with figures you can't defend.
13. Final accuracy pass
Here are my prepared answers: [paste]. Check every calculation and finance term for accuracy, flag any number that sounds confidential, and mark any answer longer than about 90 seconds spoken. Give me a list of fixes, not rewritten answers.
A list of fixes keeps the answers in your own words.
Worked example: checking the AI's variance math
Say the AI gives you this question for prompt 3:
| Budget | Actual | |
|---|---|---|
| Units | 12,000 | 11,250 |
| Price per unit | 200 | 192 |
| Revenue | 2,400,000 | 2,160,000 |
The correct split, using a common convention (volume at budget price, price at actual volume):
- Total variance: 2,160,000 − 2,400,000 = −240,000, or −10% of budget.
- Volume variance: (11,250 − 12,000) × 200 = −150,000.
- Price variance: (192 − 200) × 11,250 = −90,000.
- Check: −150,000 + (−90,000) = −240,000, which matches the total.
If the AI's two pieces don't add up to the total, ask it to show its formulas. Some teams use a different convention that creates a separate mix or joint line; say which one you're using. The same habit works for valuation: a terminal value with next year's free cash flow of 10 million, a 9% WACC and 2.5% growth is 10 ÷ (0.09 − 0.025), about 153.8 million. If the chatbot's figure differs, recompute before you quote it.
Red flags in the AI's finance answers
- Arithmetic and signs. Recompute every figure, and watch for favorable and unfavorable variances flipped.
- Market data. Interest rates, multiples and company figures from a chatbot may be outdated or wrong. Use a source you can name.
- Definitions. Companies adjust EBITDA and free cash flow differently. Say which definition you mean.
- Confidential figures. Strip anything from a current employer before pasting, and before saying it in an interview.
- Length. Technical answers should be crisp: the method, the result and what it means.
For how finance teams now ask about AI on the job, see AI interview questions for financial analysts. Then practice out loud in an AI mock interview, where every spoken answer gets a score.
Where CoPilot Interview fits
When the math is solid, practice saying it: CoPilot Interview is an AI interview helper for rehearsing technical and behavioral finance answers out loud. The free plan gives 15 minutes a day of live help in sessions of up to 15 minutes, with AI mock interviews included. Excel tests, modeling exercises and online assessments check your own skills, so do them without outside help; for live conversations, follow your employer's rules. Our finance interview help page covers banking and finance rounds more broadly.
Disclosure: CoPilot Interview is our product, so weigh what we say about it accordingly.
Frequently asked questions
Can ChatGPT help me prepare for a financial analyst interview?
Yes, for technical drills, case practice and tightening behavioral stories. Make it ask one question at a time and check your reasoning, and recompute any number it gives you.
Can I trust ChatGPT's finance math?
Not without checking. Chatbots still make arithmetic and sign errors. Ask for formulas, confirm that parts add up to totals, and redo key calculations yourself.
What technical questions come up in financial analyst interviews?
Commonly the three statements and how they link, a DCF walk-through, WACC, variance analysis, ratios and Excel. Investment roles add a stock pitch; FP&A roles add forecasting and budgeting.
Can I use AI on a financial modeling test?
Not unless the employer says so. Treat tests and take-home models as your own work and follow their instructions; use AI to practice similar exercises beforehand.
How do I practice explaining numbers out loud?
Use prompt 6 to explain a variance to a non-finance manager, then run a spoken mock where each answer is scored. Aim for the cause, the number and a recommendation in under a minute.
Rehearse it out loud before the real one
The AI mock interview builds questions from your resume and the job description, reads them aloud and scores each spoken answer out of 10. Included on the free plan.
Start a mock interview