Create a copy-paste prompt pack for a financial analyst working on [BUSINESS CONTEXT]. The available materials are [DATA SOURCES AND PERIOD], and the intended audience is [AUDIENCE]. Focus on [ANALYSIS PRIORITIES] and follow these constraints: [ACCOUNTING, CONFIDENTIALITY, AND MODEL CONSTRAINTS]. Produce exactly six distinct prompts, each 90–130 words, for: (1) monthly actual-versus-budget variance analysis, (2) rolling forecast update, (3) management commentary, (4) investment or project evaluation, (5) data-quality and reconciliation review, and (6) executive meeting preparation. Each prompt must contain bracketed input fields, ask for a specified output format, and require the model to distinguish facts, calculations, assumptions, and management interpretation. Build appropriate checks into the prompts: reconcile totals to a stated source, show formulas or calculation logic, identify materiality thresholds, separate one-time effects from run-rate effects, and label unknowns rather than filling them in.
For forecast and investment prompts, require scenario assumptions, base/upside/downside cases, and sensitivity drivers. For executive commentary, require a concise headline, key drivers, risks, actions, and decisions needed. Do not ask the model to provide investment advice, approve accounting treatment, or fabricate benchmark data. Before answering, confirm that each prompt serves a different job and that no prompt encourages unsupported numbers or conclusions. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [BUSINESS CONTEXT] | Enter the company type, business model, and relevant finance function. | A 220-person B2B software company with subscription revenue and a finance planning team |
| [DATA SOURCES AND PERIOD] | Enter the reports, model tabs, source systems, and reporting period available. | August actual P&L, FY26 budget, July rolling forecast, CRM bookings export, and headcount model |
| [AUDIENCE] | Enter the intended readers, such as a CFO, business-unit leader, or investment committee. | CFO, VP of Sales, and monthly operating review attendees |
| [ANALYSIS PRIORITIES] | Enter the metrics, business questions, or decisions the analysis should address. | ARR growth, gross margin, sales hiring, customer churn, and cash runway |
| [ACCOUNTING, CONFIDENTIALITY, AND MODEL CONSTRAINTS] | Enter accounting policy limits, data restrictions, materiality rules, and model conventions. | Use management-reporting definitions; do not share customer names; treat variances above $50,000 or 5% as material; preserve existing forecast formulas. |
How to use
- Describe the data sources and definitions before asking AI to interpret a variance or forecast.
- Run the generated prompts one at a time with a clean table or export, not a screenshot with unreadable labels.
- Reconcile every output total to your model or source report and trace narrative claims to a number or named assumption.
- Follow up with: "Convert the variance output into a CFO-ready one-page brief and label every inference separately from the reported data."
Variations
Earnings brief
Use to prepare a factual internal readout of an earnings release or call.
Analyze [COMPANY]'s [QUARTER] earnings materials: [RELEASE, DECK, AND TRANSCRIPT]. Produce a two-page internal brief for [AUDIENCE] with reported results versus stated guidance, segment drivers, cash-flow and balance-sheet observations, management outlook, and open questions. Cite the page or source for every reported figure, distinguish management claims from your inference, and do not invent consensus estimates or investment recommendations. Flag non-GAAP measures needing reconciliation.
Three-statement review
Use to inspect a model before sharing it with stakeholders.
Review this three-statement model description or export: [MODEL DETAILS]. The key assumptions are [ASSUMPTIONS], and the review period is [PERIOD]. Produce a model-audit checklist covering balance-sheet balancing, cash-flow links, working-capital signs, debt and interest circularity, scenario consistency, and formula breaks. For each issue, show the affected line, likely cause, test to run, and severity. Do not alter assumptions or claim formulas are correct without evidence.
Board KPI pack
Use to design concise management reporting for a board meeting.
Create a board KPI pack outline for [BUSINESS] for [PERIOD]. Board priorities are [PRIORITIES], and available metrics are [METRICS AND DEFINITIONS]. Produce 8–10 pages with each page's question, headline metric, comparison period, visual recommendation, required narrative, and decision or risk signal. Include definitions and data-source ownership for nonstandard KPIs. Do not create results or targets; mark every unavailable metric as a data request.
Tips
- Define the denominator before asking about margin, retention, conversion, or growth; otherwise AI may compare incompatible measures.
- Make the variance threshold explicit and require both dollar and percentage movement, since a large percentage on a small base can be immaterial.
- For forecasts, keep input assumptions in a separate block and require the narrative to distinguish model output from judgment.
- Never paste customer-level, payroll, or material nonpublic data into an unapproved AI environment; follow your employer's data-handling policy.
FAQ
Can AI build a financial model for me?
It can help structure assumptions, formulas, and review tests, but you should validate every formula, sign convention, and link in the actual workbook. It should not replace the finance owner or accounting review.
How do I prevent hallucinated financial insights?
Provide a controlled source table, require citations to rows or tabs, and instruct the model to label unknowns. Reconcile its totals and recompute critical metrics independently.
Can I use these prompts for investment decisions?
They can organize research and scenarios, but they do not provide investment advice. Use verified primary sources and your organization’s investment process.