Short answer: Use this workflow to turn defined metric terms, raw numbers, periods, formulas, currency, and assumptions into a reproducible model specification and QA checklist. It is not financial advice and should not silently fill missing numbers.
This page is a practical prompt workflow. Treat every bracketed field as required input; do not ask an AI system to invent facts, citations, customer results, credentials, compliance conclusions, or financial outcomes.
When not to use this workflow
- Metric definitions, periods, currency, or source numbers are ambiguous.
- The output will be used as investment, lending, tax, accounting, or financial advice without qualified review.
- The model depends on forecasts that are presented as actual results or guaranteed outcomes.
Variables to provide
| Variable | What to provide | Quality check |
|---|---|---|
[Metric definitions] |
Exact definitions for MRR, ARR, churn, retention, CAC, LTV, or other metrics | Definitions match the business reporting convention. |
[Raw numbers] |
Values, source systems, period, currency, and inclusions/exclusions | Numbers are traceable and not silently normalized. |
[Formulas] |
Formula, denominator, time basis, cohort rule, and rounding | Another reviewer can reproduce the result. |
[Assumptions] |
Forecast inputs, scenarios, confidence, and sensitivity range | Actuals and assumptions are labelled separately. |
[Reviewer and decision] |
Qualified reviewer, intended decision, and deadline | The model is reviewed for the decision context. |
Copy-ready prompt
You are a financial-modeling documentation assistant, not a financial adviser. Metric definitions: [DEFINITIONS]. Raw numbers with source, period, currency, and exclusions: [DATA]. Formulas: [FORMULAS OR REQUESTED DEFINITIONS]. Assumptions and scenarios: [ASSUMPTIONS]. Intended decision and reviewer: [CONTEXT].
Return: (1) a model specification, (2) an input table, (3) formula definitions, (4) a reproducible calculation checklist, (5) scenario and sensitivity questions, and (6) limitations. Never invent numbers, present forecasts as actuals, give investment/tax/accounting advice, or label the business healthy/unhealthy without an explicit framework. Mark missing fields [SOURCE NEEDED].Example prompt failure
Illustrative example only — not a customer run and not evidence.
Build a SaaS model from these numbers and tell me whether the business is financially healthy.
The illustrative prompt asks for a health judgement without metric definitions, source numbers, periods, formulas, assumptions, or decision context.
Why the controlled version is safer
The controlled prompt produces a model specification and reproducibility checks. It separates reported data from assumptions and leaves judgement to the qualified reviewer.
- Confirm the source set, version, audience, and owner.
- Run the prompt with the required fields; leave unknowns labelled rather than filling them.
- Compare each factual sentence with its source and record any judgement call.
- Obtain the named reviewer approval before publication or external use.
Proof-of-execution record
Empty proof template — no execution claim. Complete this section only after a real, permissioned run. Redact confidential customer, employee, supplier, personal, and commercially sensitive data.
- Model family and version
- [Record the actual model and version]
- Run date
- [YYYY-MM-DD]
- Input excerpt
- [Paste a short approved excerpt]
- Observed output
- [Paste a short approved excerpt or write “not recorded”]
- Human reviewer
- [Name or role]
- Changes made
- [Record edits and why]
- Limitations
- [Record failure cases, missing evidence, and unresolved questions]
Limitations and reviewer responsibility
Financial outputs can be materially wrong when definitions, cohorts, periods, or currency are inconsistent. This workflow is documentation support, not financial, investment, tax, legal, or accounting advice.