use case

Purchase Advice Automation for Residential Mortgage Operations

Residential Mortgage Lending

Mid-sized, privately held mortgage lender based in the U.S., offering retail, wholesale, and correspondent home loan services across all states.

Post-Closing & Accounting

Challenge

Mortgage lenders must process hundreds of Purchase Advice (PA) documents each week from multiple investors, each with its own format, structure, and naming conventions.

These PAs confirm secondary market transactions, outlining the terms, wire details, and purchase adjustments.
Historically, post-closing teams manually keyed critical fields—such as Investor Loan Number, Purchase Date, and First Payment Due Date—into loan systems, creating bottlenecks, error risks, and reconciliation delays between the lender’s core system(s) and investor reporting.

Key challenges included:

  • High document variability across investors and loan types
  • Time-consuming manual review of PDF and Excel-based PAs
  • Inconsistent field naming conventions (e.g., 30+ aliases for “Investor Loan Number”)
  • Limited integration between extracted data and accounting systems
  • Increased operational cost and turnaround pressure during investor delivery windows

Fisent BizAI automates the end-to-end processing of investor Purchase Advice documents using Applied GenAI Process Automation.

Fisent BizAI Solution

Representative BizAI Actions

Classify incoming documents and identify investor-specific templates and aliases

Split multi-loan PAs into individual transactions for field-level processing

Extract required data fields (e.g., Investor Loan Number, Purchase Date, Payment Due Date, Escrow & Fee amounts)

Verify extracted values against expected formats, aliases, and business rules

Analyze output for completeness and confidence thresholds, flagging any missing or low-confidence values for manual review

BizAI adapts to each institution’s workflow and data policies — returning structured data for system ingestion without storing, training, or retaining any client content.

Business Outcomes

80%+ automation of post-closing Purchase Advice data extraction

>95% field-level accuracy across high-variance investor templates

Reduction in manual entry and review time by over 60%

Accelerated reconciliation and accounting updates through structured data delivery

Improved standardization across investor templates, creating consistent downstream data quality

Broader Impact

By introducing automation at this key stage of the secondary market workflow, BizAI standardizes an otherwise fragmented process — creating a unified, reliable data model for all Purchase Advices regardless of format or investor. This standardization not only improves operational speed and accuracy but also strengthens audit trails, investor relations, and regulatory reporting.