How It Works

Step 1: Upload & Select Pages

Upload one or more PDFs through our secure interface.

For single large PDFs (like CIMs), the system analyzes all pages and highlights those most likely to contain financial statements. Select specific pages to extract — perfect when FY2023 is on page 5, FY2024 on page 6, and FY2025 on page 7.

Each file is scanned, indexed, and preprocessed. If your statements are scanned images, the system automatically applies OCR to make them machine-readable.

Supports:

• Single or bulk uploads
• Page-level selection from large PDFs
• Multi-period consolidation from one file or many
• Combined reports with multiple years or entities
• Password-protected PDFs (if authorized)

Step 2: Statement Detection

Our engine reviews each page and scores its likelihood of being an Income Statement, Balance Sheet, or Cash Flow Statement.

Page scoring combines layout patterns, keyword context, and numerical consistency to locate relevant tables even in multi-section filings.

Step 3: Table Extraction

Each detected table is extracted using multiple parsing methods to ensure accuracy.

We combine outputs from pdfplumber and Camelot in both lattice and stream modes, then compare results and choose the highest-confidence version.

Common accounting structures are recognized automatically, including:

• Multi-line headers
• Subtotals and sections
• Negative values in parentheses
• Continuations across pages

Step 4: Unit and Scale Normalization

Financial statements often report figures in thousands or millions.

DealSheets.ai detects the declared units, rescales all values to a consistent standard, and ensures subtotals and totals align correctly.

Example:
If a Balance Sheet shows "($ in millions)," the export converts every figure accordingly.

Step 5: Statement Classification

Each extracted table is classified and labeled by statement type using rule-based logic and learned patterns.

Income, Balance, and Cash Flow statements are assigned to separate sheets in Excel.

Additional tables, such as KPIs or segment data, are stored in supplemental sheets for reference.

Step 6: Validation and Cross-Checks

Before export, the system performs multiple consistency checks:

• Balance Sheet assets = liabilities + equity
• Cash Flow net change = difference in cash between periods
• Income Statement net income matches Cash Flow inputs

Results are displayed in a Validation Sheet with color-coded indicators and links back to the original data.

Step 7: Excel Generation

Your final Excel workbook includes:

• Separate sheets for each statement type
• A validation sheet with cross-check results
• A provenance sheet showing page numbers, parser method, and confidence scores
• Clean number formatting, accounting negatives, and grouped subtotals

Step 8: Download and Review

Once processing is complete, you can download your Excel workbook or a ZIP containing all files from a bulk run.

Each workbook is traceable, versioned, and ready for analysis, integration, or model updates.