FP&A
Key financial KPIs every business owner should track
Five or six numbers show whether the business is healthy and what to fix first. Here is each KPI's formula, the question it answers, and how often to check it.
How has AI changed FP&A for a growing business?
AI tools can take over parts of the preparation work in planning and forecasting. Here is where they help, where they fall short, and how to try one safely.
By Surfside Capital AdvisorsPublished Updated 4 min read
AI in FP&A can speed up the preparation work: cleaning data, drafting variance commentary, and running more scenarios in less time. It does not change who answers for the numbers. A forecast an AI tool drafts is only as good as the data and assumptions behind it, and a person has to check it before you act on it.
A vendor, an employee, or your feed says AI will run your forecasting. You want to know what is real, what is safe to try next week, and what a person still has to check. FP&A, or financial planning and analysis, is the budgeting, forecasting, and reporting work that turns your records into decisions.
You are not behind. As of May 3, 2026, 19.8% of U.S. businesses told the U.S. Census Bureau they used AI in any business function, and less than 20% of firms with four or fewer employees did. Among finance and insurance firms the rate was 33.9%.
Where does AI in FP&A help today?
The strongest use is the preparation layer of the work described in what FP&A is. Each task below still ends with a person's check.
| Task | What an AI tool can do | What a person checks |
|---|---|---|
| Data cleanup | Sorts and tags transactions, and flags duplicates and odd entries | That the categories match how you run the business |
| Invoice and expense capture | Reads documents, matches them to purchase orders, and flags exceptions | The exceptions, and who approves them |
| Variance commentary | Drafts a first explanation of why results differ from budget | Each reason against what you know, such as a delayed order |
| Forecast drafts | Proposes a baseline from your history and seasonality | The assumptions: price changes, hires, and lost customers |
| Scenarios | Runs more what-if cases in less time | Which cases matter, and what you decide |
| Reporting | Turns the numbers into a plain-language summary | Every figure against the source |
Some accounting and planning tools you already pay for include features like these. Ask your provider which ones exist and what data they use.
Where does AI fall short in FP&A?
It falls short in four places:
- Data quality. A tool learns from your records, so inconsistent categories produce inconsistent output.
- Context. It does not know about a customer's delayed order or a planned price change unless you say so.
- Confident errors. Generated text can state a wrong number or an invented reason in fluent sentences, so every claim needs a source.
- Change. A tool built on past patterns can struggle when the business changes shape, for example after a new product or the loss of a major customer.
Can AI forecast your revenue?
Not reliably on its own. A tool can fit patterns in your history, but it cannot know what is about to change in your business. Treat its output as a draft baseline, adjust it for what you know, and measure its error the way you would any forecast. How to do financial forecasting shows how to track that error month by month.
How do you protect your financial data?
Before you paste financial data into any AI tool, find out what the provider does with it: whether it stores your inputs, whether it uses them to train its models, and who at your company can reach them. Customer, employee, and bank data need extra care, so check what your customer contracts and privacy commitments allow before any of that data leaves your systems.
The National Institute of Standards and Technology publishes a voluntary AI Risk Management Framework, with a profile for generative AI released on July 26, 2024, to help organizations identify risks like these. You can borrow its core questions without adopting the whole framework: what could go wrong, and who checks?
What does not change?
Someone still answers for the numbers. A lender, a partner, or a buyer will ask you, and "the tool said so" will not satisfy them. Good FP&A has always depended on clean records, written assumptions, and a monthly review that ends in decisions, and none of that moves because a tool drafts the first version faster.
What to do next
- List where your month-end and planning hours go, and pick one repetitive task, such as first-draft variance commentary, as a pilot.
- Decide what data the tool may see, and read the provider's terms on storing inputs and training models.
- Run the pilot for one month beside your normal process. Compare the two outputs and record every error.
- Set one rule: a person checks every number before it leaves the company.
To talk through what to automate, see how Surfside Capital Advisors approaches forecasting, KPI dashboards, and scenario modeling within its FP&A work.
Questions owners ask
Will AI replace your finance person?
It takes over some of the repetitive tasks. Judgment about what the numbers mean, the questions to ask, and accountability for the answer stay with people. The work shifts toward reviewing and deciding, and that favors someone who understands your business.
Do you need an AI tool to do FP&A well?
No. A clean spreadsheet, written assumptions, and a monthly review work well without one. Add a tool for a specific task once you can name the hours it will save and the check a person will run on its output.
General information, not legal, tax, or investment advice.
Sources
- Large firms with at least 20 employees biggest AI users (Business Trends and Outlook Survey), U.S. Census Bureau,
- AI Risk Management Framework, National Institute of Standards and Technology,
Figures and rules as published by these sources on the dates shown.
Free consultation
Facing this decision now?
Talk it through with our team in a free 30-minute consultation, in English or Spanish, with no obligation.
Keep reading
FP&A
How to do financial forecasting for your business
A forecast is your best current estimate of revenue, costs, and cash. Here is how to build one from your history, test it against reality, and keep it useful.
FP&A
What is FP&A, and does your business need it?
FP&A is the planning and analysis built on your accounting records: budgets, forecasts, scenarios, and the monthly review. Here is what it produces and who needs it.
A