Calls
MISMO 3.4 Export

MISMO 3.4 and FNMA Loan Export

Aila includes native support for mortgage lending teams to extract Fannie Mae / MISMO 3.4 structured loan application files (Form 1003) directly from borrower calls.


Overview

During mortgage intake and pre-qualification calls, Aila AI models extract borrower profile data, employment history, liabilities, asset declarations, and loan terms into a standardized loan-file context package.

Lending teams can export this data as:

  • MISMO 3.4 XML: Standard mortgage industry data format.
  • FNMA 3.2: Legacy Fannie Mae 1003 format for older Loan Origination Systems (LOS).

Exporting a Loan File from a Call

Step 1: Open the MISMO Surface

  1. Open a completed borrower intake call.
  2. Click the MISMO / Loan File tab in the call workbench.

Step 2: Review Extracted Loan Data

Review the structured fields extracted from the conversation:

  • Borrower Information: Full legal name, SSN mentions (redacted in UI), contact details, marital status, dependent count.
  • Current & Target Residence: Address, housing status (Own/Rent), monthly payment.
  • Employment & Income: Employer name, job title, years on job, base monthly salary, bonus/commission income.
  • Assets & Liabilities: Bank balances, real estate owned, monthly debt payments.
  • Loan Parameters: Target loan amount, purchase price, property type, occupancy intent (Primary, Secondary, Investment).

Step 3: Edit or Validate Fields

  1. Click any field to edit or supplement data with borrower documentation.
  2. The system flags missing mandatory MISMO 3.4 schema elements before export.

Step 4: Export to LOS

  1. Click Export MISMO 3.4 or Export FNMA 3.2.
  2. Select your export destination:
    • Direct Download: Download the .xml or .fnm file to your computer.
    • Email to LOS: Send the file directly to your LOS ingestion email address (Encompass, Byte, Calyx Point, LendingPad).

Technical Architecture

  • Extraction Model: Powered by the LOA v2 extraction pipeline (src/backend/loan-file/).
  • Context Freshness: Enforces deterministic validation against raw transcript segments before file generation.
  • PII Protection: Social Security numbers and account numbers are tokenized according to enterprise security standards.

Related Documentation