Financial Analysis Prompts: Budgeting, Forecasting, Reporting

Numbers do not persuade. Stories about numbers do. AI cannot do your maths — but it is excellent at turning the maths you have already done into clean budget narratives, plain-English forecasts, and board-ready commentary. This tutorial covers the prompts that get you there.

1. Introduction

Most finance work is spreadsheet work, and AI is the wrong tool for that. Where AI earns its place is the layer above the spreadsheet: explaining variance, summarising forecasts, drafting the narrative section of a board pack, and stress-testing assumptions in plain language. Used carefully, it converts what used to be a Friday-night writing job into a 30-minute review.

Used carelessly, however, AI invents numbers with frightening confidence. The whole skill of financial prompting is feeding it the data and forbidding it to invent. We will look at three high-value use cases: budget narratives, quarterly forecasts, and variance commentary.

2. The Concept Explained

A financial analysis prompt has a strict shape: you provide the numbers, the AI provides the interpretation. Reverse that and you get hallucination. The reliable pattern is a three-block prompt — context, data table, interpretation rules.

Think of the AI as a finance analyst who can read fast but cannot use a calculator. You hand them a clean P&L; they write a tight memo explaining what changed. They do not recompute anything. Your maths is the ground truth.

3. The Problem Without This Technique

forecast our revenue for next year

4. The Solution

Act as a Director of Finance preparing the narrative
section of a quarterly board pack.

Company: Saffron Treats, a D2C bakery brand selling
online and through 12 retail kiosks (Mumbai, Pune,
Bengaluru). Financial year: April to March.

Below is the actuals data and our model for the next
two quarters. Treat this data as ground truth. Do not
recompute. Do not invent any new numbers.

Q1 (Apr–Jun) actuals:
- Revenue: INR 4.2 crore
- Gross margin: 58%
- Marketing spend: INR 38 lakh
- New customers: 11,400
- Repeat purchase rate: 32%

Q2 (Jul–Sep) actuals:
- Revenue: INR 3.6 crore
- Gross margin: 55%
- Marketing spend: INR 32 lakh
- New customers: 8,900
- Repeat purchase rate: 35%

Forecast assumptions for Q3 (Oct–Dec, festive season):
- Revenue: INR 5.8 crore
- Gross margin: 56% (slight dip on packaging)
- Marketing spend: INR 55 lakh
- New customers: 16,500
- Repeat purchase rate: 33%

Write a board narrative (350–450 words) that:
1. Opens with the headline story of H1 in two sentences.
2. Explains the Q2 dip honestly — likely drivers
   (monsoon footfall in kiosks, fewer wedding orders).
3. Walks through the Q3 forecast and the three
   assumptions it most depends on.
4. Names the two biggest downside risks if Q3 misses.
5. Ends with the one number to watch (you decide which).

Rules:
- Do not introduce any numbers not in the data above.
- Use INR formatting (crore / lakh).
- No buzzwords. No "we are excited to announce".

5. Step-by-Step Breakdown

1. Do the maths first. AI cannot reliably compute. Bring a clean spreadsheet to the prompt — not a half-finished one.
2. State the audience and document. Board memo, internal monthly review, lender update — each one has a different tone.
3. Paste the data as a clean table or labelled list. The cleaner the input, the cleaner the narrative.
4. Forbid new numbers explicitly. "Do not introduce any number not in the data above." This single line prevents most finance hallucinations.
5. Ask for assumptions and risks. A good narrative names what could go wrong, not just what is expected to go right.
6. End with one metric to watch. Forces the AI to pick a leading indicator, which clarifies your own thinking.

Tip: For budgeting season, give the AI last year's actuals plus this year's department asks, and ask it to flag the three biggest line-item increases with one question each you should ask the requester. It becomes a stress-test partner before the budget review meeting.

6. Practice Exercises

7. Key Takeaways