Organizations increasingly use efficient technologies to harness and harvest data to deliver new and efficient service bundles to customers. In addition, regulatory standards such as PSD2 and Dodd-Frank have forced organizations to manage and use data more efficiently.
Large amounts of data are handled by all financial service sectors—payments, lending (unsecured and secured), insurance and wealth, and banking. Consumers of this data are both internal and external to financial institutions. Big Data demonstrated its ability to harvest large data volumes, especially unstructured data (also known as dark data). Therefore, it is a widely preferred choice of technology at present.
Account aggregation or financial data aggregation solutions were initially used by wealth management and personal finance management advisors. However, they are now used across the global financial sector. Data aggregation is one such component offered exclusively by some FinTech players, as well as managed internally by larger industry players.
Mortgage lenders recognize the need to operate in this fast-changing landscape. Mortgage aggregators, i.e., entities that purchase mortgages from financial institutions and securitize them into mortgage-backed securities (MBS) for sale, such as Fannie Mae and Freddie Mac, earn a profit by purchasing individual mortgages at low prices and subsequently selling the pooled MBS at high prices. Both these players in the mortgage landscape have an underlying need for aggregated data. Most of them rely on FinTechs to receive or harness this data.
FinTechs are innovating in multiple areas of data aggregation with various technologies and solutions. Here are some examples:
Yodlee: Yodlee, owned by Envestnet, is a data aggregation and data analytics platform powering cloud-based innovation for digital financial services.
Plaid (Quovo, Mint): Plaid, owned by Visa, provides technical infrastructure APIs that connect consumers, traditional financial institutions, and developers. The company provides critical insights into the data access that it provides with a suite of analytics products. Plaid owns Quovo, a data platform providing insights and connectivity for millions of financial accounts across thousands of institutions. With industry-leading APIs, modular applications, and enterprise solutions, Quovo makes it easier for enterprises to connect with their clients’ financial life.
Finicity: Finicity is a cloud-based financial API platform that offers financial application developers worldwide services to create next-generation financial applications.
Other key players in this category are Fiserv, Signzy, Google, CMOTS, Tink, 3i Data Scraping, Perfios, and Ionixx Technologies.
On the one hand, aggregators such as Envestnet’s Yodlee provide secure and compliant data on banks, credit cards, investments, loans, rewards, and financial accounts from thousands of global data sources. On the other hand, there are point technology solutions. For example, Google’s LendingDocAI helps mortgage companies accelerate the process of evaluating a borrower’s income and asset documents using specialized machine learning models to automate routine document reviews.
Fiserv is one of the FinTech incumbents that offer AI-enabled, automated mortgage workflow. In addition, Signzy is among those boutique FinTechs that offer AI-powered technology to enable Optical Character Recognition (OCR) and image forensics for KYC processes.
Increasing Use of Artificial Intelligence (AI) in Mortgage Business
AI provides highly sophisticated tools and solutions for data aggregation. Loan origination is still not optimized; paper-based, operator-dependent processes create significant time delays and mistakes, leading to inconsistent and inaccurate data.
Checking financial wellness and adhering to KYC compliance guidelines require capturing the unstructured data of new customers and extracting structured information from sources. In this scenario, AI plays a more prominent role than Big Data. The use of AI for document data capture, image forensics, and automation of routine document reviews is beneficial to aggregators.
Data needs to be captured at scale, where its capture is streamlined and the processing time is substantially reduced.
Optical Character Recognition (OCR) and Intelligent Character Recognition (ICR) are the common deployments of AI in lending processes. Electronically obtaining necessary information from non-digitized documents reduces processing time and errors substantially.
Google’s LendingDocAI: Launched in October 2020, this solution automates mortgage document processing. It is offered as a dedicated service to the mortgage industry with a well-optimized use of AI. Currently, the solution focuses on the US market. It reads and understands documents using Google Cloud’s OCR and Natural Language.
Here are the critical components of LendingDocAI:
Document OCR: This component has an AI-enabled optical character recognition feature that reads from digitally filled or even handwritten forms and documents.
Natural Language: This component derives insights from unstructured text using Google’s Machine Learning. Natural Language APIs enable sentiment analysis, entity analysis, entity sentiment analysis, content classification, and syntax analysis.
Invoice and Form Parser: This component extracts data from invoices and general documents. The mortgage lending industry uses specialized form parsers to extract text and spatial structures from Form W-2 and Form 1099 using OCR.
- W2 Parser: Extracts data such as employee name, employer name, and wages from Form W-2
- 1099 MISC Parser: Extracts data such as payer name, recipient name, and amount from Form 1099-MISC
- 1003 Parser: Extracts over 50 fields from Fannie Mae Form 1003
Lending Document Splitter and Classifier: This component performs automated document classification and validation. It identifies documents in a large file and classifies known lending document types.
Document AI API: A single API that can be used to access OCR, Natural Language products, form parsers, and invoice parsers.
An average mortgage application can run into 300+ pages. An AI-enabled, automated data extraction method provides an immediate and substantial reduction in processing time.
M&A and partnerships in this space strongly reflect the growing interest in data aggregation. Here are some examples:
- Visa acquired financial aggregator FinTech Plaid (which had earlier acquired Mint and Quovo) for a sizable sum of $5.3 billion in January 2020.
- Recently, Google invested heavily in creating and launching LendingDocAI, a lending industry-specific AI-enabled solution.
- Mastercard partnered with FinTech Signzy (hailed as a no-code AI platform) for the global rollout of its biometric video KYC solution. It co-invested $5.4 million in Signzy.
- Wells Fargo, a leading US mortgage player, and Yodlee recently signed a data exchange agreement.
Users of data aggregation services are looking for solutions that offer high-quality data, a better experience for borrowers and lenders, lower loan origination costs, faster workflows, and unstructured data capture. Point solutions that offer these benefits and additional features, such as insightful text analyses, have a promising road ahead.
To learn about Prove’s identity solutions and how to accelerate revenue while mitigating fraud, schedule a demo today.
Keep reading
Developers know identity verification is an essential element of effective digital onboarding and the customer lifecycle. Choosing the right one can feel like navigating a maze of features and complexity.
In an age where our smartphones have become almost like extensions of ourselves, the identity assurance achieved through smartphone possession and data is a natural evolution.
Rodger Desai, CEO of Prove, a leading identity verification solution provider, offers a unique perspective on the rising fraud in the gig economy, advocating for robust digital identity verification as a key defense mechanism.
