FP&A Command Center

FP&A Command Center AI Ultra

Elevator Pitch

FP&A Command Center AI Ultra is a local-first AI-enabled Financial Planning & Analysis command center for planning, forecasting, variance analysis, financial statements, CSV uploads, templates, exports, and live data workflows — no login or backend required.

App Link

https://app-bjcv8bavb401.appmedo.com/

Inspiration

The project was inspired by the gap between spreadsheets and enterprise FP&A platforms.

Spreadsheets are flexible, but they become fragile as planning models grow. It becomes difficult to manage data quality, assumptions, versions, templates, calculations, and outputs reliably.

Enterprise FP&A tools such as Anaplan, Oracle EPM, OneStream, and Workday Adaptive Planning are powerful, but they are expensive, implementation-heavy, and usually require enterprise infrastructure.

I wanted to build something between these two worlds: a lightweight FP&A command center that feels structured like an enterprise planning tool but runs locally in the browser.

The goal was not to replace enterprise EPM systems. The goal was to make core FP&A workflows easier to learn, test, and use.

Built With MeDo

This project was built using MeDo, which helped turn a complex FP&A product idea into a working browser-based application.

MeDo was used to rapidly generate, test, refine, and improve the app across multiple development cycles. The project went through many iterations, including building the dashboard, planning modules, editable grids, upload system, template library, Demo Mode, Live Mode, financial statement logic, exports, and AI Finance Copilot support.

One of the biggest advantages of using MeDo was the ability to move quickly from idea to working prototype, then continue improving the app through testing and issue fixing. Instead of only describing the product, I could build and refine actual workflows.

MeDo helped with:

• Generating the initial app structure
• Building the FP&A module layout
• Creating dashboards and planning pages
• Adding Demo Mode and Live Mode
• Creating editable planning grids
• Implementing CSV templates and upload flows
• Adding local browser persistence
• Building financial statement pages
• Improving exports and validation
• Iterating through bugs and fixes quickly

The project also showed that even with AI-assisted building, serious testing is still necessary. Page loading is not enough. A finance app must prove that inputs produce correct outputs.

What It Does

FP&A Command Center AI Ultra helps users move from financial inputs to planning outputs.

The app includes:

• Executive Dashboard
• Demo Mode with sample FP&A data
• Live Mode for user-entered or uploaded data
• CSV upload and mapping
• Template Library
• Revenue Planning
• COGS Planning
• OPEX Planning
• Headcount Planning
• CAPEX Planning
• Working Capital Planning
• Budgeting
• P&L Statement
• Balance Sheet
• 3-Statement Model
• Forecasting
• Variance Analysis
• Data Quality and Anomaly Center
• AI Finance Copilot support
• Export functionality
• Browser-based local persistence

Demo Mode and Live Mode

One important design decision was separating Demo Mode and Live Mode.

Demo Mode allows users to explore the platform with sample financial data. It helps users understand dashboards, statements, forecasts, and analysis before using their own data.

Live Mode allows users to manually enter planning data or upload CSV files. Live Mode is designed to stay separate from Demo Mode so users know whether they are viewing sample data or their own data.

This matters because finance tools must be honest about data source.

Upload and Template System

The app includes templates for structured FP&A inputs.

Templates include:

• Company Setup
• Planning Calendar
• Chart of Accounts
• Dimensions
• Assumptions
• Driver Library
• Revenue Planning
• COGS Planning
• OPEX Planning
• Headcount Planning
• CAPEX Planning
• Working Capital
• Cash Flow
• Balance Sheet
• Budget vs Actual
• Forecast
• Trial Balance
• General Ledger
• KPI
• Product Profitability
• Customer Profitability
• Scenario Assumptions
• Workflow / Approval
• Audit Trail Import
• Report Input
• AI Finance Copilot Prompt Template

The intended workflow is:

1. Choose the relevant template
2. Download the CSV
3. Fill or modify the sample data
4. Upload the file
5. Map and validate fields
6. Import into Live Mode
7. Review dashboard and statement outputs

Financial Logic

The app is built around core FP&A calculations.

Revenue Planning

Revenue = Units × Price × (1 - Discount %)

Example:

Units = 100
Price = 1,000
Discount = 0%

Revenue = 100 × 1,000 × (1 - 0%)
Revenue = 100,000

Gross Profit and Gross Margin

Gross Profit = Revenue - COGS

Gross Margin % = Gross Profit / Revenue × 100

Example:

Revenue = 100,000
COGS = 60,000

Gross Profit = 100,000 - 60,000
Gross Profit = 40,000

Gross Margin % = 40,000 / 100,000 × 100
Gross Margin % = 40%

EBITDA

EBITDA = Gross Profit - OPEX

Example:

Gross Profit = 40,000
OPEX = 25,000

EBITDA = 40,000 - 25,000
EBITDA = 15,000

EBIT and Net Profit

EBIT = EBITDA - Depreciation

Profit Before Tax = EBIT - Interest Expense

Tax = max(0, Profit Before Tax × Tax Rate)

Net Profit = Profit Before Tax - Tax

Working Capital

Working Capital = Receivables + Inventory - Payables

Example:

Receivables = 50,000
Inventory = 30,000
Payables = 20,000

Working Capital = 50,000 + 30,000 - 20,000
Working Capital = 60,000

Free Cash Flow

Free Cash Flow = Operating Cash Flow - CAPEX

Example:

Operating Cash Flow = 40,000
CAPEX = 10,000

Free Cash Flow = 40,000 - 10,000
Free Cash Flow = 30,000

Closing Cash

Closing Cash = Opening Cash + Operating Cash Flow - CAPEX + Financing Cash Flow

Example:

Opening Cash = 100,000
Operating Cash Flow = 40,000
CAPEX = 10,000
Financing Cash Flow = 0

Closing Cash = 100,000 + 40,000 - 10,000 + 0
Closing Cash = 130,000

Variance Analysis

Variance = Actual - Budget

Variance % = (Actual - Budget) / Budget × 100

Example:

Actual Revenue = 100,000
Budget Revenue = 120,000

Variance = 100,000 - 120,000
Variance = -20,000

Variance % = -20,000 / 120,000 × 100
Variance % = -16.67%

For revenue, a negative variance is unfavorable. For costs such as OPEX or COGS, a lower actual value can be favorable.

Forecasting

Naive Forecast = Last Actual Value

Moving Average = Average of recent periods

Weighted Forecast = Recent Month × 50% + Previous Month × 30% + Earlier Month × 20%

Example:

March Revenue = 140,000
February Revenue = 120,000
January Revenue = 100,000

Weighted Forecast = 140,000 × 50% + 120,000 × 30% + 100,000 × 20%
Weighted Forecast = 126,000

Anomaly Detection

The app flags unusual financial movements using threshold rules.

Examples:

Revenue drop greater than 20% = anomaly
OPEX increase greater than 15% = anomaly
CAPEX increase greater than 25% = anomaly
Gross margin drop greater than 5 percentage points = anomaly

AI Finance Copilot Support

The AI Finance Copilot section is designed to help users structure finance analysis and management commentary.

It supports use cases such as:

• Variance explanation
• CFO summary
• Forecast commentary
• Cash risk review
• OPEX challenge
• Revenue driver analysis
• Board-pack narrative
• Anomaly explanation
• Scenario recommendation
• Management action planning

The app also includes an AI Finance Copilot Prompt Template so users can structure finance questions clearly.

How I Built It

The project was built as a local-first browser app using MeDo.

The main design principles were:

• Keep data local to the browser
• Avoid login and backend complexity
• Separate Demo Mode and Live Mode
• Use templates for structured data input
• Convert planning inputs into financial facts
• Preserve dimensions such as product, customer, channel, entity, department, and account
• Show outputs through dashboards, statements, forecasts, and exports
• Make financial logic traceable

The core workflow is:

Input Data → Mapping → Validation → Financial Facts → Calculations → Dashboard / Reports / Exports

Challenges I Faced

The hardest part was not creating screens. The hardest part was making the workflows connected and honest.

Some of the key challenges were:

• Preventing Demo Mode and Live Mode from mixing data
• Making sure Live Mode does not silently show demo data
• Ensuring uploaded dimensions such as product, customer, channel, department, and entity are preserved
• Avoiding duplicate facts when users save or upload the same data repeatedly
• Making validation specific to the selected upload type
• Ensuring financial calculations match expected results
• Making exports clearly show Demo or Live mode
• Adding AI Finance Copilot support without overclaiming fake AI capability

This forced the project to move beyond page-load testing. A finance app should not only load pages. It must prove that inputs produce correct outputs.

Input → Processing → Output → Expected Result

What I Learned

This project made one thing very clear: FP&A software is mostly about data integrity, not dashboards.

A dashboard is useful only if the input data is mapped correctly, the calculations are traceable, and the output can be trusted.

I also learned that local-first apps can be powerful for finance learning and prototyping. Users can explore planning workflows without needing cloud infrastructure, login systems, or enterprise setup.

The most important lesson was that testing must be based on expected versus actual outputs, not only screenshots.

I also learned that MeDo can accelerate app creation significantly, but the builder still needs strong product thinking, detailed prompts, ruthless testing, and clear acceptance criteria.

Current Status

FP&A Command Center AI Ultra is being submitted as a Hackathon Release Candidate.

It is suitable for:

• Hackathon demonstration
• FP&A education
• Lightweight financial analysis
• Small-business planning
• Consulting demos
• Workflow prototyping

It is not positioned as a full enterprise EPM replacement.

Known Limitations

Some areas still need future improvement:

• Some configuration templates may be download/manual-reference only
• Some upload workflows may need additional parser hardening
• Users should manually validate outputs before using them for high-stakes financial decisions
• Enterprise features such as multi-user roles, approval workflows, ERP integration, and cloud database persistence are outside the current scope

Final Thought

FP&A Command Center AI Ultra is an attempt to make financial planning and analysis more accessible.

It combines templates, uploads, manual planning, financial statements, forecasting, variance analysis, exports, and AI-assisted finance thinking into one browser-based app.

The project is not trying to replace enterprise finance systems. It is trying to make the FP&A workflow easier to learn, test, and use.

Try the app here: FP&A Command Center AI Ultra