This course gives you a ready-to-use Claude prompt that turns a raw financial dataset — a P&L export, a set of period-over-period figures, or a simple revenue/expense breakdown — into calculated ratios, trend commentary, and flagged anomalies. You'll get a clean summary table plus a written narrative you can drop straight into a report. The whole process takes under five minutes once your data is prepared.
What Your Finished Financial Analysis Looks Like
Below is the kind of output Claude produces from a quarterly dataset. Each period is compared to the one before it, key ratios are calculated automatically, and any line item moving outside a normal range is flagged for review.
A
B
C
D
E
F
1
Period
Revenue
Gross Margin %
OpEx
Net Margin %
Flag
2
Q1 2025
412,000
58.2%
145,000
22.9%
Normal
3
Q2 2025
438,500
57.6%
151,000
21.8%
Normal
4
Q3 2025
401,200
52.1%
149,500
17.4%
Watch
5
Q4 2025
395,800
41.3%
162,000
6.2%
Anomaly
Claude flags gross margin or net margin swings beyond a normal range so you can investigate the underlying cause before reporting the numbers.
Calculates gross margin, net margin, and other core ratios from raw revenue and cost figures
Compares each period to the previous one and reports growth or decline in plain language
Flags any line item whose movement falls outside a normal range — no silent averaging
Produces a written summary you can paste directly into a report or send to stakeholders
Where this fits: Financial data analysis is one of the core tasks covered in our AI for Excel course hub. If you want to track these ratios on an ongoing basis rather than one-off, see the Financial KPI dashboard course next.
Prepare Your Financial Dataset for Analysis
Claude works best with a single, clean table rather than a full formatted P&L export. Pull the core line items into Excel or CSV with one row per period, and keep the column headers consistent. Here is the minimum required layout:
Financial Summary Layout
A
B
C
D
1
Period
Revenue
COGS
OpEx
2
Q1 2025
412,000
172,200
145,000
3
Q2 2025
438,500
186,000
151,000
4
Q3 2025
401,200
192,100
149,500
01
Use one row per period, not per transaction
Aggregate transaction-level data into period totals first. Claude analyzes trends and ratios far more reliably from a compact summary table than from thousands of raw rows.
02
Keep every period in the same currency and units
Mixing currencies or reporting some periods in thousands and others in full figures will throw off every ratio Claude calculates. Normalize first.
03
Label periods clearly and consistently
Use a consistent format like Q1 2025 or Jan 2025 across every row so Claude can order periods correctly and calculate period-over-period change.
Tip: If your export includes subtotal or "Total" rows mixed in with the period rows, remove them before pasting into the prompt — otherwise Claude may double-count them as an extra period.
The Exact Claude Prompt for Financial Data Analysis
Copy this prompt into Claude and paste your dataset where indicated. The prompt includes guardrails so Claude calculates from your actual figures rather than estimating or filling gaps.
Prompt — paste into Claude
I need you to analyze a financial dataset in Excel. Below is the data, one row per period.
FINANCIAL DATA (columns: Period, Revenue, COGS, OpEx):
[Paste your rows here]
Rules:
- Calculate Gross Margin % as (Revenue - COGS) / Revenue for each period.
- Calculate Net Margin % as (Revenue - COGS - OpEx) / Revenue for each period.
- Compare each period to the immediately preceding period and calculate the % change in Revenue, Gross Margin %, and Net Margin %.
- Flag any period where Gross Margin % or Net Margin % moves more than 5 percentage points versus the prior period as "Watch", and more than 10 percentage points as "Anomaly". Otherwise mark it "Normal".
- Do NOT estimate or fill in missing figures. If a period is missing a value, state that explicitly instead of guessing.
- Base every number strictly on the data provided — do not reference external benchmarks unless I ask for them.
Output:
1. A summary table with columns: Period | Revenue | Gross Margin % | OpEx | Net Margin % | Flag
2. A short written summary (3-5 sentences) describing the overall trend and calling out any period flagged as "Watch" or "Anomaly"
3. One sentence suggesting what might be worth investigating for each flagged period
Format the table as Excel-ready so I can paste it directly into a spreadsheet.
Why Financial Data Analysis Still Needs Human Judgment
Claude is fast at arithmetic and pattern spotting, but it doesn't know why a number moved — it only knows that it did. A margin drop could mean a pricing problem, a one-off cost, or a data entry error, and only someone with business context can tell the difference. Treat every "Watch" or "Anomaly" flag as a starting point for a question, not a finished conclusion.
Claude handles the repetitive calculation and comparison work across periods. Your job is to investigate each flagged item, confirm the numbers against source data, and decide what — if anything — needs to change. If you want to track these same ratios on a rolling basis rather than analyzing a static export, the Financial KPI dashboard course shows how to keep this analysis continuously updated.
Common Analysis Errors
A
B
C
1
Error
Likely Cause
Fix
2
Every period shows "Anomaly"
Mixed currencies or units between periods
Normalize every row to the same currency and scale before pasting
3
Ratios look implausibly high or low
A "Total" row was included alongside period rows
Remove subtotal/total rows so each row is a single period
4
Periods analyzed out of order
Inconsistent period labels (e.g. "Q1" vs "2025 Q1")
Use one consistent label format across every row
Frequently Asked Questions
Can Claude analyze my financial data directly from Excel?
Claude cannot open .xlsx files natively in the free chat. Copy the relevant columns from Excel and paste them as text into the prompt, or use Claude Pro, which supports direct file attachments.
What kind of financial insights can Claude generate from raw data?
Claude can calculate margins and ratios, identify period-over-period trends, flag unusual swings in any line item, and summarize the overall financial picture in plain language, all from the numbers you provide.
Will Claude's calculations always be accurate?
Claude performs arithmetic reliably when given clean, well-labeled data, but you should always spot-check key totals and ratios against your source figures before using them in a report.
Can I use this for multi-year or multi-period financial data?
Yes. Include a Period column identifying each row (e.g. Q1 2024, Q2 2024) and Claude will compare across periods to calculate growth rates and highlight trends.
Does this replace a financial analyst?
No. Claude accelerates the data-crunching and first-pass insight generation, but interpreting business context, validating assumptions, and making decisions still requires a human analyst.