Quantiphi partnered with a prominent wholesale and correspondent lender in the United States to automate their client onboarding process using AI-powered intelligent document processing. By leveraging Quantiphi’s proprietary platform, Dociphi, the client achieved significant improvements in accuracy, efficiency, and turnaround times.

About the Client

The client is a United States-based wholesale and correspondent lender specializing in providing tailored mortgage solutions. They are known for their commitment to delivering superior services to their partners and borrowers, emphasizing efficiency and accuracy.

Problem Statement

The client’s manual onboarding process led to inefficiencies, errors, and scalability challenges. Underwriters spent excessive time sorting documents, resulting in delays and compliance risks. The lack of automation caused inconsistent data extraction and bottlenecks during peak periods, highlighting the need for an AI-driven solution to streamline workflows and improve accuracy.

Challenges

Manual, Repetitive Work: Underwriters spent significant time performing repetitive, low-value tasks like manually sorting and reviewing onboarding documents. This impacted their ability to focus on higher-value activities.
Error-Prone Processes: The manual nature of document handling resulted in a high error rate, with inconsistent and inaccurate data extraction affecting downstream processes.
Complex Document Handling: The client wanted a solution to classify a wide variety of onboarding documents—such as tax forms, pay stubs, and bank statements—and extract only the required and relevant information. This data needed to be formatted correctly for downstream systems, ensuring seamless integration and compliance.
Unstructured Data: Many submission documents contained unstructured or semi-structured data, making it challenging to identify key-value pairs, validate document quality, and extract actionable information for decision-making.
Scalability Issues: The existing process was not scalable to meet the growing volume of onboarding requests, leading to bottlenecks during peak periods.

The Solution

Quantiphi implemented Dociphi, its AI-powered intelligent document processing platform, to automate the end-to-end client onboarding process. The solution utilized proprietary AI/ML models to process and validate key onboarding documents, ensuring a seamless and accurate data flow:

Solution Highlights:

Document Ingestion and Validation:
- User-friendly interface for uploading documents.
- Automated validation to flag unclear or incomplete documents before ingestion
Advanced Document Processing:
- Automated classification of document types, determining type and content.
- High-accuracy data extraction from multiple fields, embedded objects, and tables.
Human-in-the-Loop (HITL):
- Enabled manual review and adjustments where necessary, ensuring accuracy and compliance.
Downstream Integration:
- Output customized in compliant CSV and JSON formats for seamless integration into downstream systems.
Active Learning:
- Continuous model retraining based on HITL actions and new documents to improve accuracy over time.

Document Types Processed

IRS Forms (1040, 1040 Schedule C, W2, 1099, 4506C)
Driving Licenses
Passports
Bank Statements
Paystubs

Results and Impact Created

Improved Accuracy:
- Achieved ~95% classification accuracy and ~90% data extraction accuracy.
Standardized Workflows:
- Streamlined the onboarding process, reducing operational inconsistencies.
Reduced Errors:
- Minimized manual errors, ensuring data accuracy and compliance.
Enhanced Turnaround Time:
- Accelerated onboarding processes, enabling quicker client onboarding.

Looking to automate your workflows with intelligent document processing? Partner with Quantiphi to achieve measurable results and operational excellence. Contact us today.