SBA Loan Audit System
Automated document classification, gap analysis, and audit report generation for SBA loan applications. Powered by Lamatic AI for Locality Bank.
How It Works
Four-step pipeline from document upload to final audit report.
Document Categories
7 primary categories covering all SBA loan documentation. Click any card to explore its subcategories.
The applicant business entity. One per loan package.
Affiliated business entities. Repeats per entity. Triggered by 1919 Q#3 or discovered during audit.
Corporate guarantor entities. More docs than affiliate, fewer than borrower. Repeats per entity.
Individual owners and personal guarantors. Repeats per person.
Conditional documents based on loan structure. Not all apply to every deal.
Documents created by the lender. Not borrower-submitted.
Catch-all for documents that don't fit other categories.
What AI Catches
Cross-document inconsistencies and SBA credit box violations detected automatically.
| Finding | Severity |
|---|---|
| Liabilities mismatch between debt schedule and balance sheet | CRITICAL |
| Bankruptcy/criminal disclosure conflicts across forms | CRITICAL |
| Owner names or percentages differ across documents | HIGH |
| Revenue in tax returns doesn't match P&L or projections | HIGH |
| DSCR below 1.15x minimum threshold | CRITICAL |
| Insufficient equity injection (10-30% required) | HIGH |
3-Flow Architecture
Deep dive →Three automated workflows with clean separation of concerns.
job_queue. Flow 3 polls and executes. Clean decoupling.