AI-Powered Property Financial Extractions Overview

Purpose

Built's AI-powered extraction feature streamlines the process of spreading rent rolls and cash flows into your system of record. Users can upload Excel workbooks or PDF files, and Built's AI automatically extracts and validates the financial data — summarizing findings, flagging items for review, and remembering your mapping preferences for future extractions on the same deal.

Key Features

1. Document Extraction: Upload Excel workbooks or PDFs and let Built's AI extract rent roll and cash flow data automatically.

2. Automatic Document Splitting: Built identifies page ranges containing rent rolls or cash flow statements and splits documents accordingly, so you can associate each section with the correct deal before extraction begins.

3. Confidence-Based Validation: Each extraction includes an overall confidence level and a full list of validation checks — from header-count matches to row-value consistency and aggregation tests — so you know exactly where to focus your review.

4. Custom Mapping Preferences: Override any line item mapping recommendation directly in the extraction interface. Built remembers your choices per deal, reducing repetitive work on future extractions.

Target Users

Deal management users who need to extract, validate, and summarize financial data from property documents such as rent rolls and cash flow statements.

How It Works

Access the feature by navigating to Portfolio > Document Extractions. The listing shows every extraction along with its status, the associated deal, and processing time.

Uploading documents

1. Click Start Document Extraction and select Property Financials.

2. Upload your files — Excel workbooks or PDFs. You can upload multiple files at once (e.g., a workbook with multifamily and retail rent rolls, or a PDF with cash flow statements).

   

3. Built scans the documents and identifies page ranges containing rent rolls and cash flow statements. Associate each identified split with the appropriate deal and property, then start the extraction.

   

Reviewing extraction results

1. Once extraction is complete, use the sparkle icon to view bounding boxes showing where data was extracted from the original document.

2. Check the Extraction Summary panel on the right for overall confidence level, validation checks (e.g., aggregate totals matching line item sums), and any items flagged for review. Expanding All checks validated shows the full list of policy evaluations Built's AI ran against the extraction.

   

Note:

• For rent rolls, Built may call out specific fields to double-check even when overall confidence is high. Review these prompts before confirming.

Customizing mapping preferences

Each extracted line item has a Rollup Category dropdown. Open it to see all available categories (Revenue, Affordable GPR, Base Rent, CAM, and more) and remap any line item with a single click. Built remembers your choices for future extractions on the same deal.

Confirming and creating statements

1. When satisfied with the extraction results, proceed to the Confirm and Create screen.

2. Review the statements that will be added to the deal along with headline numbers (Net Operating Income, Net Cash Flow) to sanity-check before committing.

3. Confirm to add the statements to Built.

   

Benefits

• Significantly reduces the manual effort required to spread rent rolls and cash flows.

• Confidence levels and validation checks surface data quality issues before they enter your system of record.

• Mapping preferences are learned per deal, so subsequent extractions require less manual intervention.

• Headline numbers (NOI, Net Cash Flow) are visible at the Confirm and Create step for a final sanity check before committing data.

Best Practices

Review all fields flagged for double-check even when overall confidence is marked High — this is especially common with rent rolls where unit-level details may have minor inconsistencies. Before confirming, verify the headline numbers on the Confirm and Create screen against your source document. When overriding a mapping recommendation, confirm the correct Rollup Category before saving so Built learns the right preference for future extractions.