PROJECT 01 · FINANCE / AUTOMATION

AI Financial Reporting Agent

A lightweight workflow that consolidates personal financial data, applies business logic, generates contextual AI commentary, and archives a monthly PDF report automatically.

OUTCOME ~92%
less reporting time~1 hour → ~5 minutes, including data preparation
PRODUCT AREAFinancial reporting
FOCUSAutomation and insight
TECHNOLOGYApps Script · Gemini · Google Workspace

THE PROBLEM

Tracking data was not the same as understanding it.

Personal financial information was spread across spreadsheets covering net worth, portfolio allocation, planned investments, and historical reports. Every monthly review repeated the same consolidation and interpretation work.

HOW THE MVP WORKS

One complete loop, from scattered sheets to an archived report.

The MVP connects tools already used for tracking finances rather than introducing another application. Its four capabilities cover the complete journey from source data to an archived report.

01

Collect and aggregate the data

Reads net worth history, portfolio allocation, DCA plans, and previous reports from multiple Google Sheets, then consolidates them into one reporting dataset.

02

Apply financial logic

Finds the latest valid values, detects the relevant tabs, and identifies upcoming contributions before any AI commentary is generated.

03

Generate contextual commentary

Uses Gemini to produce observations on allocation, concentration, investment discipline, progress, and possible next actions.

04

Create and archive the report

Builds a Google Doc, generates a timestamped PDF, and stores the completed monthly report in Google Drive.

Current portfolio allocation chart produced for the financial report
Portfolio-allocation visual included in the report. The data shown is illustrative.
AI-generated commentary section from the financial report
Example AI commentary covering observations, risks, discipline, and possible next actions.

PRODUCT DECISIONS

Practical automation before product complexity.

01

Build around the existing workflow

Google Sheets remained the source of truth. The solution added interpretation and reporting without asking the user to migrate their financial system.

02

Keep calculations deterministic

Aggregation and financial logic happen before the model is called. Gemini contributes commentary; it does not replace the structured financial calculations.

03

Position it as reporting, not advice

The purpose is to surface context and support reflection. It is explicitly not intended to provide financial advice.

OUTCOME & LEARNING

A complete monthly report at negligible operating cost.

The workflow successfully orchestrates multiple financial sources, creates contextual observations, generates a professional report, and archives the output. According to the project testing, a monthly execution costs only a fraction of a cent per report.

KEY LEARNING
AI was most useful after the reporting logic was made reliable: structured data creates trust; commentary adds context.

Boundaries

  • The workflow depends on fixed spreadsheet structures and expected tab names.
  • The quality of commentary depends on the completeness and accuracy of the source data.
  • The output supports personal reporting and reflection; it is not financial advice.
Explore the implementation

See the workflow, sample data, setup instructions, and project documentation on GitHub.

View GitHub ↗