The finance function runs on deadlines that never move: the close, the board pack, the budget, the audit. It is deep, repetitive, judgment-heavy work, and it is exactly where AI is landing hardest. A Stanford and MIT study of 277 accountants across 79 firms found AI cut the month-end close by 7.5 days and returned roughly 240 hours a year per person, and Deloitte's Q4 2025 CFO Signals survey found 87% of CFOs expect AI to be extremely or very important to finance operations in 2026 [1][2]. The teams getting those hours back are not the ones with the biggest tool budget. They are the ones with a reusable prompt library aimed at the close, the reporting, and the analysis that eats their month.
This article collects 50 ready-to-use prompts for finance teams (controllers, FP&A, finance managers, and the CFO's office), organized by workflow: month-end close, audit preparation, management and CFO reporting, budgeting and forecasting, variance and financial analysis, AP/AR and cash flow, accounting policy and process, and stakeholder communication. If you searched for AI prompts for finance, ChatGPT prompts for month-end close, or AI prompts for CFO reports, the relevant set is below. Each uses [bracket placeholders] so you can adapt it. They work with ChatGPT, Claude, and Gemini, with notes where one model clearly fits a task better. (This is corporate, in-house finance; if you run an accounting or tax practice, the sibling 50 AI prompts for accountants and tax advisors is built for that work instead.)
Finance team workflow map: eight areas where prompts give time back
Before you paste: two non-negotiable rules
1. Financial data is confidential, and often material. Unreleased results, forecasts, deal terms, payroll and personal data, and bank or vendor details do not belong in a consumer AI chatbot: the free tiers retain prompts and may train on them. For a public or pre-IPO company, pasting unreleased numbers can also mean handling material non-public information the wrong way, which is a legal line, not a style choice. For real data, use an enterprise or no-training deployment with a contractual commitment and the right controls. Otherwise de-identify before you prompt: strip names and account numbers, replace real figures with placeholders or ratios, and keep only the shape of the problem ("a customer 60 days past due on a [amount] balance"), never the identity.
2. AI is not the accountant of record. Verify every number and every treatment. These prompts draft, structure, and explain. They do not hold the accounting judgment. AI invents figures, misstates standards (US GAAP, IFRS, or local like OIC), applies the wrong treatment with total confidence, and fabricates citations. Never let a model compute, reconcile, or "check" real numbers you have not verified yourself, and verify every figure, calculation, and accounting treatment against the ledger, the source document, and the applicable standard before it reaches a report, a filing, or the board. Treat AI as a fast, tireless analyst whose work you always review, never as the controller who signs off.
Safe-use decision aid: what you can paste, what to de-identify, and what a human must verify
With those two guardrails in place, here are the 50 prompts.
1. Month-End Close & Reconciliations
The recurring grind, and where AI buys back the most days. It structures and explains; you verify every number against the ledger.
1. Build a close checklist
Act as a controller. Build a month-end close checklist for a [industry] company with [team size] in finance. Organize it by day (close day 1 to [N]), with each task, the owner role, the dependency, and the account or system involved. Flag the tasks most likely to cause delays. I will adapt it to our chart of accounts and confirm completeness.
2. Explain a reconciliation difference
I am reconciling [account, e.g. bank / intercompany / AR sub-ledger to GL] and have a difference of [amount]. Here are the de-identified line items on each side: [paste]. List the most likely causes of a difference like this in order of probability (timing, unrecorded items, FX, cutoff, duplicates), and the check that would confirm or rule out each. Do not tell me the answer is any specific line; help me investigate.
3. Draft a reconciliation template
Design a reusable reconciliation template for [account type]. Include the columns and the sign-off fields a reviewer needs, a section for reconciling items with an aging, and a short instruction for the preparer on what "reconciled" means here. Keep it auditable. I will populate the real balances.
4. Write accrual and journal entry logic
Explain the standard month-end accruals for a [industry] company (e.g. payroll, unbilled revenue, prepaid amortization, accrued expenses), and for each give the entry logic in words (what debits, what credits, and the basis for the estimate). Note the ones that most often get missed or misstated. I will book and verify the actual amounts and confirm the treatment.
5. Summarize the close status
Turn this raw close-tracker status into a short update for the finance lead: [paste de-identified task list with statuses]. Give me what is done, what is blocked and why, what is at risk for the deadline, and the one or two items that need a decision. Keep it to a few lines, no filler.
6. Draft a flux/close review commentary
Here are the preliminary close numbers vs prior period, de-identified: [paste]. Draft a brief close-review commentary for the controller: the two or three movements worth a second look before we lock, framed as questions to investigate, not conclusions. Flag anything that looks like a possible cutoff or classification issue. I verify every figure.
7. Build a close post-mortem
We just closed in [X] days and want to get to [Y]. From these de-identified notes on what slowed us down: [paste], identify the recurring bottlenecks, separate the ones we control from the ones we do not, and propose three specific process changes for next close, each with the day it would save. Keep it realistic for our team size.
2. Audit Preparation & Controls
High-stakes, documentation-heavy, and safe for AI as long as you verify. It drafts the narrative and the request list; you own the evidence.
8. Build a PBC (prepared-by-client) list
Act as an audit manager. Draft a PBC request list for the [year-end / interim] audit of a [industry] company, organized by area (revenue, cash, AR, inventory, fixed assets, payroll, debt, equity). For each item, note the document or report that satisfies it and the owner. Keep it standard; I will tailor it to our auditor's specific requests.
Want to know how effective your prompts are? Prompt Score analyzes them on 6 criteria.
Write a control narrative for the [process, e.g. procure-to-pay / order-to-cash / month-end close] process. Describe the flow end to end, and at each step note the control, whether it is preventive or detective, who performs it, and the evidence it produces. Keep it factual and specific enough to walk an auditor through. I will confirm it matches how we actually operate.
10. Explain a control-gap remediation
We have this control gap: [describe, de-identified]. Explain the risk it creates in plain terms, propose two or three practical remediation options that fit a [team size] finance team, and note the trade-offs of each (cost, effort, segregation of duties). Do not assume tools we may not have.
11. Prepare for an audit question
The auditor is likely to ask about [area, e.g. revenue recognition on a specific contract type / an unusual accrual]. Help me prepare: the standard that applies (I will confirm the exact reference), the questions they will probably ask, the evidence I should have ready, and the clearest way to explain our position. Flag where our treatment might warrant a second opinion.
12. Draft a walkthrough document
Create a walkthrough document template for the [process] control walkthrough: the steps, the sample selection approach, the attributes to test, and the space to record results and exceptions. Keep it aligned to how auditors document walkthroughs. I will run the actual walkthrough and record real results.
13. Summarize an accounting standard for a specific case
Summarize how [standard, e.g. ASC 606 / IFRS 15 revenue] applies to this specific fact pattern: [describe the transaction, de-identified]. Give me the recognition steps, the judgment areas, and the disclosures it triggers. Cite the standard sections so I can verify each against the actual text. Flag any area where the treatment is genuinely a judgment call.
3. Management & CFO Reporting
Where the numbers become a story for the board. AI drafts the narrative from your verified figures; you own the message.
14. Draft board-pack commentary
Here are the verified results for [period] vs budget and prior year, de-identified: [paste key figures]. Draft the management commentary for the board pack: what happened, why, and what it means, in the CFO's voice. Lead with the headline, keep it to [length], tie movements to business drivers, and separate one-offs from trend. Leave every number as I gave it; do not compute new ones.
Claude tends to hold a measured, executive tone across a long commentary; give it one past board note as a voice sample first.
15. Write an executive summary of the month
Turn this month's verified P&L and KPI results into a one-page executive summary for the leadership team: [paste de-identified figures]. Structure it as headline, what drove it, what to watch, and actions needed. Translate finance metrics into business impact. Skip the vanity metrics. I will confirm every figure before it goes out.
16. Build a KPI commentary
For each of these KPIs and their movement vs target, write one tight line of commentary explaining the driver and whether it needs attention: [paste de-identified KPI table]. Do not restate the number, explain it. Flag the one or two that most need a decision or a follow-up.
17. Prepare board Q&A
Based on these results and the story we are telling: [paste de-identified summary], anticipate the ten hardest questions the board will ask, and for each draft a clear, honest two-to-three sentence answer. Include the questions we would rather avoid. Note where the honest answer is "we do not know yet."
18. Draft an MD&A section
Draft the [revenue / margin / cash / liquidity] section of a management discussion and analysis for [period], from these verified figures: [paste]. Explain the drivers and the year-over-year change in plain, defensible language. Avoid boilerplate and unsupported optimism. I will verify every figure and the compliance framing.
19. Rewrite a report for a non-finance audience
Rewrite this finance update so a non-finance executive audience gets it fast, without dumbing it down: [paste]. Lead with the "so what", replace jargon with plain terms, and keep the two or three numbers that actually matter. Same facts, clearer delivery.
20. Build a reporting narrative template
Create a reusable narrative template for our monthly reporting: the fixed sections (results, drivers, risks, outlook, asks) with a one-line instruction under each for what belongs there. Make it consistent enough that the board sees the same structure every month and can find things fast.
4. Budgeting & Forecasting (FP&A)
Where finance looks forward. AI structures the model and the assumptions; you own the numbers and the judgment.
21. Build a budget-build checklist
Act as an FP&A lead. Build a checklist for the annual budget process for a [industry] company: the sequence from assumptions to consolidation to board approval, the inputs needed from each department, the deadlines, and the review gates. Flag the steps where budgets usually slip. I will fit it to our calendar.
22. Design a driver-based forecast structure
Help me structure a driver-based forecast for [revenue / a cost line]. Identify the underlying drivers (e.g. units, price, headcount, ramp), how they connect to the output, and the assumptions I need to set for each. Describe the model logic in words so I can build it; do not invent our actual numbers.
23. Pressure-test assumptions
Here are the key assumptions in our [forecast/budget], de-identified: [paste]. Play skeptical CFO: which assumptions are most aggressive or most fragile, what would break them, and what evidence I should have to defend each. Rank them by how much the output moves if they are wrong.
24. Draft scenario definitions
Help me define three planning scenarios (base, downside, upside) for [the business / a specific line]. For each, describe the assumption changes, the trigger that would put us in that scenario, and the management action it implies. Keep them distinct and decision-useful, not just plus-or-minus 10%.
25. Write a rolling-forecast cadence
Design a rolling-forecast process for a [team size] finance team: the cadence, what gets re-forecast each cycle vs held, the inputs, and how it connects to the annual budget. Keep it light enough to actually sustain monthly. Note the discipline that makes rolling forecasts fail if skipped.
The techniques you're reading about work. Test your prompts now with Prompt Score and see your score in real time.
Our forecast moved from [prior] to [current] for [metric], de-identified: [paste the bridge]. Draft a short explanation for leadership: what changed, why, and what it means for the year. Distinguish the parts within our control from the parts that are not. I confirm the numbers.
27. Build a budget-vs-actual review structure
Design a monthly budget-vs-actual review: the format, the threshold that makes a variance worth discussing, the questions to ask the budget owner, and how to capture actions. Keep it focused on decisions, not a line-by-line read-out. I will set the real thresholds.
5. Variance & Financial Analysis
Where "what happened" becomes "why, and what now". AI reasons over the figures you give it; you decide the story.
28. Explain a variance
Here is a variance: [line item] came in at [actual] vs [budget], de-identified: [paste supporting detail]. Walk through the likely drivers, separate volume from price/rate from mix, and tell me which explanation the data supports versus which is a hypothesis I still need to check. Do not invent a cause the data does not show.
29. Build a bridge analysis
Help me structure a bridge (waterfall) from [starting metric] to [ending metric] for [period], de-identified: [paste the components]. Order the bridge steps logically, label each driver, and note which steps are the ones leadership will ask about. I will confirm each amount ties out.
30. Interpret financial ratios
Here are our key ratios for [period] vs prior and vs [benchmark if I have one], de-identified: [paste]. Interpret them in plain business terms: what each says about liquidity, efficiency, leverage, or profitability, and which movements are worth a management conversation. Flag any ratio that could be distorted by a one-off.
Gemini's structured output is handy when you want the ratios and interpretations laid out as a clean table you can drop into a deck.
31. Analyze a trend
Here is [metric] over the last [N] periods, de-identified: [paste]. Describe the trend, whether it is accelerating or decelerating, any seasonality, and the inflection points worth explaining. Separate signal from noise. Tell me what additional cut of the data would sharpen the picture.
32. Draft a margin analysis
Here is our gross/operating margin by [product / segment / period], de-identified: [paste]. Analyze where margin is being made and lost, the mix effects, and the two or three levers with the most impact. Frame it for a decision, not just description. I confirm the underlying numbers.
33. Build a profitability question set
We think [product/segment] may be underperforming on profitability. Give me the analysis I should run to confirm or reject that: the cuts, the allocations to be careful about, and the traps (e.g. shared cost allocation, timing). Structure it as a plan I can execute, not a conclusion.
6. AP/AR & Cash Flow
The working-capital engine, and where clear communication moves cash. AI drafts; you keep the relationships and verify the figures.
34. Build a cash-flow forecast structure
Help me structure a [13-week] cash-flow forecast for a [industry] company: the inflow and outflow categories, the timing logic for each, and the reconciliation back to the bank. Describe the model so I can build it; do not fabricate our actual receipts or payments. Flag the categories that are hardest to forecast.
35. Draft a collections message
Draft a professional, firm-but-warm collections message for a customer [N] days past due on a [amount] balance, de-identified. Give me two versions: a first reminder and a firmer follow-up. Keep the relationship intact and include a clear next step and payment path. I will confirm the balance and terms.
36. Analyze AR aging
Here is our AR aging summary, de-identified: [paste]. Tell me what it says about collection health, which buckets or concentrations are the real risk, and the two or three actions that would most improve DSO. Flag any single customer concentration worth escalating.
37. Draft a vendor payment negotiation
Help me draft a message to a key vendor to [extend terms / negotiate a payment plan / discuss a discount], de-identified: [context]. Professional, relationship-preserving, with a specific ask and a rationale. Give me a fallback position. Do not over-commit on our behalf.
38. Explain a working-capital movement
Our working capital moved by [amount] this [period], de-identified: [paste the components]. Explain the drivers (DSO, DPO, inventory), whether the movement is healthy or a warning, and what it implies for cash. Keep it to what the data supports. I verify the figures.
39. Build a cash-position summary
Turn this de-identified cash and liquidity data into a short daily/weekly cash summary for the CFO: [paste]. Position, movement, upcoming large items, and any covenant or minimum-balance flag. Keep it scannable. Leave every figure exactly as given.
7. Accounting Policy, Compliance & Process
The written backbone of the function. AI drafts the document; you own the accuracy and the sign-off.
40. Draft an accounting policy memo
Draft an accounting policy memo for [topic, e.g. capitalization threshold / revenue recognition for a contract type / lease treatment]. Include the scope, the policy, the rationale tied to the standard (I will confirm the reference), examples of in-scope and out-of-scope, and the review date. Keep it clear enough for a non-specialist to apply. I verify the technical treatment.
41. Write a process document (SOP)
Write a standard operating procedure for [process, e.g. the monthly bank reconciliation / the expense approval flow]. Step by step, with the owner, the system, the control point, and what "done" looks like at each step. Written so a new hire could follow it. I will confirm it matches reality.
42. Summarize a regulatory or standard change
Summarize this [standard update / regulatory change] and what it means for a [industry] company like ours: [paste or describe]. What changes, the effective date, the practical steps to comply, and who needs to act. Flag the parts that need professional advice. I will verify against the primary source.
43. Draft a controls self-assessment
Create a controls self-assessment questionnaire for the [process] area: the key controls, a yes/no/partial question for each, and space for evidence and gaps. Keep it practical for a [team size] team to complete quarterly. I will use it on our actual controls.
44. Build an audit-readiness checklist
Build a running audit-readiness checklist we can maintain year-round, not just at year-end: the documentation to keep current, the reconciliations to keep clean, and the sign-offs to keep on file, by area. The goal is no year-end scramble. I will assign owners.
45. Explain a compliance requirement plainly
Explain [compliance requirement, e.g. a filing deadline / a threshold / a disclosure] in plain terms for a non-specialist stakeholder: what it requires, by when, what happens if we miss it, and what we need to do. No jargon. I will confirm the specifics against the source.
8. Stakeholder Communication & Professional Development
The part that never fits in the close calendar, and compounds the most. AI is a drafting and thinking partner; you make the calls.
46. Draft a cross-functional finance update
Draft a short monthly finance update for [department, e.g. sales / product / ops]: how they are tracking to their budget, the two or three things they should know, and what finance needs from them. Plain language, partnership tone, no lecture. Leave the figures as placeholders for me to fill and verify.
47. Prepare for a difficult finance conversation
Help me prepare to tell [a budget owner] that [difficult message, e.g. their spend is over and needs to come down], de-identified. Give me a structure: open with the shared goal, state the facts, the specific ask, and a path forward. Firm, respectful, solution-focused. The judgment is mine.
48. Brief me on a finance topic
I need to get up to speed on [topic, e.g. a new revenue model / a financing instrument / a metric leadership is asking about]. Teach it to me: the core mechanics, why it matters for us, the questions I should be able to answer, and the two or three things people most often get wrong. Flag anything that may be out of date so I verify it.
49. Draft an investor or lender update
Draft an update for [investors / our lender] for [period], from these verified figures and this narrative: [paste de-identified]. Lead with the headline, be straight about what is working and what is not, tie it to the plan, and end with the outlook. Confident but honest, no spin. Leave every figure as given.
50. Turn my week into a finance-lead update
Here are my rough notes from this week: [paste de-identified]. Turn them into a short update for the CFO: what shipped (close, reports, analysis), what the numbers are saying, what is at risk, and what needs a decision. Confident but honest about what is still a draft or unverified. Leave any figure I must confirm as a placeholder.
Turn these into templates
You do not have to rebuild every prompt from scratch. The Keep My Prompts templates library has structured, ready-to-run prompts built for exactly this work:
A prompt you run once and forget is worth little. A prompt that becomes part of your close and reporting routine compounds. Three moves separate finance teams that get marginal value from AI from those that get real days back:
Save them. Every prompt that returns a usable draft goes into a personal prompt library tagged by close-day and workflow. Next time you tune rather than start from a blank cell. If your prompts currently live in a spreadsheet, here is why that breaks and how to migrate.
Score them. When a prompt returns something vague or off, the problem is almost always the prompt, not the model. Use the six criteria that make a good prompt, clarity, specificity, context, structure, examples, role, to rewrite. Giving the model your real report structure and one past example is usually what turns generic output into your format.
Version them. The first time a close or reporting prompt works cleanly, it becomes a template. Keep the version that works, so next month, next quarter, or after a standard changes, you adapt rather than rebuild. That is the difference between using AI and having a system that runs your close on a full workload.
The finance teams reclaiming those days are not doing more with AI. They are using fewer, better, reusable prompts for the close, the reporting, and the analysis that drains their month, they protect confidential and material data without exception, and they keep every number, every treatment, and every sign-off firmly in human hands.
Sources
[1] Stanford University and MIT, study of 277 accountants across 79 firms on AI in the accounting workflow: AI reduced the month-end close by approximately 7.5 days and returned roughly 240 hours per person per year (about $19,000 of recovered capacity per person). Reported via industry coverage of the study, 2026. https://stealthagents.com/research/ai-in-accounting-and-finance-statistics-2026
