Hi Reader ,

Big news: Claude Sonnet 5 was released, and if you work in Finance or FP&A, this is one of the most exciting AI model updates to test.

The reason is simple: finance teams do not need AI that only writes emails faster.

They need AI that can help build analysis, dashboards, forecasts, commentary, and decision-support tools.

I created a full video guide for you: I Tested Claude Sonnet 5 on Real Finance and FP&A Use Cases​

That is what makes Sonnet 5 interesting.

In my test, I used it across several real Finance and FP&A use cases:

• Scenario modeling
• Budget vs actual variance analysis
• CFO-ready executive commentary
• Forecasting simulators
• Monte Carlo analysis

I also created an interactive guide with 100 use cases: https://claude.ai/public/artifacts/13fbc0e4-4bf0-463d-a4c8-cd0c5e2dab20​

And the results were genuinely impressive.

Below is a breakdown of each use case, how it works, and why it matters for finance teams.

But first, if you're looking to go deeper, my friend Nicolas Boucher will be demonstrating practical AI use cases, including some with Claude, in his free AI CFO Masterclass this Tuesday.

You'll learn how to master AI with just one hour per week, automate reporting, forecasting, and analysis without writing code, reduce days of manual work to minutes, adopt AI securely, and stay up to date with the latest AI tools and when to use them.

Register here: https://nicolasboucher.online/op/ai-cfo-live-masterclass/?utm_medium=partners&utm_source=christian-martinez&utm_campaign=wky-become-an-ai-cfo​

In case you cannot attend live, email me back and I can send the recording after the masterclass!

Now to the use cases:

Claude Sonnet 5 for Finance and FP&A: Practical Use Cases

1. Scenario Modeling Dashboard

One of the first use cases I tested was a scenario modeling dashboard.

The idea was simple:

Give Claude Sonnet 5 a realistic finance dataset and ask it to build an interactive dashboard with three scenarios: Base case, Best case, Worst case

Instead of just giving a written answer, Claude created a full dashboard-style output with KPIs, assumptions, sensitivity ranges, scenario values, and an executive summary.

Why this matters for FP&A:​
Scenario modeling is not just about calculations. It is about helping leaders understand trade-offs. Claude Sonnet 5 can help turn raw assumptions into a visual decision-support tool much faster.

This was the prompt used:

You are an FP&A analyst. Using the dataset below (or the attached sheet), build a super aesthetic
html dashboard with a 3-scenario model (Base, Best Case, Worst Case) for FY26 showing Revenue, COGS, Gross Margin, Operating Expenses, and EBITDA.
Apply each driver's Best/Worst Case delta to the base value using formulas so the model is fully auditable.
Then write a short executive summary (under 150 words) identifying which 2 drivers have the biggest impact on EBITDA, and recommend one action
management could take to protect the downside case.
​
Make it the most impressive dashboard for a CFO
​
Driver Unit FY26 Base Value Best Case Δ Worst Case Δ Notes
Units Sold (000s) units 1,200.00 +12% -15% New product launch upside vs. demand softness
Average Selling Price $ 42.50 +3% -6% Pricing power vs. competitive discounting
COGS per Unit $ 24.10 -4% +8% Supplier contract renegotiation vs. input cost inflation
Marketing Spend $000s 8,500.00 +10% -20% Growth investment vs. cost-cutting
Headcount (FTE) FTE 340.00 +5% -10% Planned hiring vs. hiring freeze
FX Rate (USD/EUR) rate 1.08 +5% -8% EUR revenue translation sensitivity
Churn Rate % 0.09 -2pp +3pp Customer success investment vs. competitive losses
CapEx $000s 4,200.00 -15% +25% Deferred projects vs. accelerated infrastructure spend

And got this amazing output on the first try:

Then you can go top right and download it or share it:

I demonstrate step by step here: I Tested Claude Sonnet 5 on Real Finance and FP&A Use Cases​

2. Use Claude Sonnet 5 for Budget vs Actual Variance Analysis with CFO-Ready Executive Commentary

The second use case was budget vs actual variance analysis.

This is one of the most practical applications for any finance team.

Every month, finance teams need to explain what changed:

• Revenue vs budget
• Gross margin vs budget
• Operating expenses vs budget
• What drove the variance
• What management should care about

In the test, I gave Claude Sonnet 5 a budget-vs-actual dataset and asked it to prepare monthly variance commentary for a CFO pack.

The output was:

Why this matters for FP&A:​
Most finance teams spend too much time preparing commentary manually. Claude Sonnet 5 can help accelerate the first draft, highlight key drivers, and structure the story for management.

This was the prompt I used:

You are preparing the monthly variance commentary for the CFO pack.
Using the budget-vs-actual data below, calculate variances in $ and %, then
decompose the Revenue variance into price vs. volume effects (assume Budget volume 420,000 units
at $44.05 ASP; Actual volume 390,000 units at $44.10 ASP).
Write variance commentary (2-3 sentences per line item) for Revenue, Gross Margin, and EBITDA, ranking the three
lines by which one management should be most concerned about and why.
​
Make it as a super impressive artifact to share with CFO
​
Line Item Budget ($000s) Actual ($000s) Variance ($000s) Variance (%) Driver Notes
Revenue 18,500 17,200 (1,300) (7.0%) Volume miss in EMEA region; pricing held
COGS 9,250 9,050 (200) (2.2%) Lower volume reduced variable costs
Gross Margin 9,250 8,150 (1,100) (11.9%)
Sales & Marketing 3,200 3,650 450 14.1% Unplanned trade show + agency overage
R&D 2,100 2,080 (20) (1.0%) On plan
G&A 1,450 1,610 160 11.0% One-off legal fees
Total OpEx 6,750 7,340 590 8.7%
EBITDA 2,500 810 (1,690) (67.6%)

3. Forecasting Simulator

The third major test was a forecasting simulator.

I gave Claude Sonnet 5 six months of actuals and a set of assumptions, then asked it to build a forecasting tool.

The result was an interactive simulator where you could adjust assumptions like: Monthly revenue growth, December seasonal uplift, Cash collection rate andVolatility assumptions

As the assumptions changed, the forecast updated automatically.

This is one of the most exciting use cases because forecasting is usually not a static process.

In real FP&A work, stakeholders constantly ask:

What happens if growth slows?
What happens if collections improve?
What happens if December is stronger than expected?

A forecasting simulator lets you answer those questions interactively.

Instead of building one static forecast, you create a tool where management can test different assumptions.

This is a big shift.

The forecasting simulator also included Monte Carlo analysis.

This is an advanced but very useful FP&A technique.

Instead of showing only one forecast, Monte Carlo analysis runs many simulations based on different possible outcomes.

For example, rather than saying:

“Revenue will be $25M.”

You can say:

“Based on the current assumptions and volatility range, revenue is likely to fall between $23M and $27M, with a base case around $25M.”

That is much more useful for decision-making.

In finance, the future is uncertain.

Revenue growth can change. Margins can move. Cash collections can be delayed. Seasonality can be stronger or weaker than expected.

Monte Carlo analysis helps show a range of outcomes instead of pretending that one number is perfectly certain.

Claude Sonnet 5 can help create a simulator where the user can adjust assumptions and run simulations directly.

Why Claude Sonnet 5 Matters for Finance Teams

The biggest takeaway from my test is this:

Claude Sonnet 5 is not just useful for writing.

It is useful for building.

That distinction matters.

Many finance professionals still think of AI as something you use to draft emails, summarize documents, or rewrite commentary.

Those use cases are helpful, but they are only the beginning.

The more interesting use cases are:

• Building dashboards
• Creating simulators
• Generating analysis tools
• Writing first-draft commentary
• Designing scenario models

I also created a guide for Fable 5! You can look at the guides for the latest Claude models here: https://drive.google.com/drive/folders/1pKoWsBw_C0oz7N51w3KJ1sFuZOBVIjZW?usp=drive_link​

Thanks,

Christian Martinez