A future-proof solution for faster mortgage loan origination

Meta-description: Harnessing AI solutions in the right way is the key to speeding up the loan origination process. Read more.

Introduction:

90% of mortgage origination in the UK and Ireland is non-intelligent. Around 40% of origination requires manual intervention, resulting in a higher time to credit.

In the post-pandemic world, borrowers have become more digital-savvy and prefer getting applications processed faster. This creates a need for an AI mortgage lending solution that facilitates first-time-right applications. 

This whitepaper discusses how mortgage AI can be leveraged in multiple forms to tackle the problems arising at various stages in lending, such as application health check, underwriting, case packaging, and more, all through first-time-right applications. 

The Problem:

Mortgage lending UK is a vast and diverse industry, requiring a variety of tools at various stages of the application. This creates a need for a platform that harnesses the power of top-notch AI technologies in the market and is easy to integrate with existing LOS systems.  

The loan origination process is lengthy, manual, and expensive.

Right now, there are major hurdles in the loan origination process due to a lack of coordination between stakeholders, delays in manual document verifications, and lengthy underwriting processes. 

The problem arises from the fact that outdated methods and technologies are being used in the process. Apart from that, there is no viable way of harnessing superior AI tech as per custom requirements.

The Solution: 

A modular platform that harnesses the latest AI mortgage lending technologies is the key to creating a future-proof solution.

Digilytics RevEl: The right combination of AI tools, in the right form

Digilytics offers a cloud-based loan origination system that is targeted towards driving down costs at every stage of the loan origination process, helping businesses increase competitiveness and quality of service. 

Let’s explore the various aspects involved and see how RevEl offers the right solution to the problem at hand.

Saving on SLAs 

Increased errors and human intervention create delays. There are thousands of types of documents that are submitted. Getting accustomed to all of them and validating the right ones is bound to create inefficiencies in the application health check process.

RevEl harnesses the power of computer vision and machine learning, along with NLP based on Google’s BERT algorithm to facilitate an intelligent document verification process. This feature is available in the RevEl Application Component. 

RevEl Application

RevEl leverages computer vision to understand images, which uses deep neural networks based on the highly successful convolutional neural network algorithm. Along with that, RevEl leverages machine learning to make sense of all the textual data captured by OCR, like estimating income and recognizing loan-specific documents (pay slips, bank statements, etc.).  

The textual data is then further enriched by Natural Language Processing (Based on Google’s BERT algorithm) to make contextual sense of all the words and phrases. 

The case is then auto-packaged and made ready for underwriting. With instant visibility of errors, reduced human-led admin assistance, and auto-packaging of the case, SLA penalties can be reduced by a huge margin.

 

Real-Time Application Health Check

The first-time-right application delivers machine-driven assistance prior to submission. Let’s explore how it is made possible.

One-Shot Learning: 

Normal machine learning models are trained on multiple instances of just one class. One-shot learning deals with using a fixed reference instance for all classes, which is then compared with the unknown input given to the model. 

A similarity score is then generated between the known reference and the unknown input, allowing the model to predict its validity. For, eg.: Recognizing an unknown payslip by comparing it with a known reference of a payslip. 

Cross-Validation:

In this technique, machine learning models are trained on sub-sets of available input data and are tested on the complementary subset of the input data. This allows machine learning models to improve accuracy when it comes to understanding independent data. 

With automatic validations for correctness, consistency, and completeness in the application, a real-time application health check is made possible.

Speeding up Underwriting: Uberification of underwriting 

Lengthy and detail-intensive underwriting processes delay the delivery timelines and often cause a mess for the lenders. With the uberification of underwriting, the process can be streamlined and made more efficient. 

RevEl Underwriting

The RevEl Underwriting module classifies application cases into various categories, like standard and accelerated (multispeed). It then algorithmically allocates loan processors and underwriters to application cases as per their availability.

Depending on the category, the document scrutiny level is decided. The case is then auto-packaged for underwriting. The auto-validations and computerized data entry, along with intelligent document processing, helps in reducing the underwriting time by up to 30%

Furthermore, it enables machine-learning based classification and allocation of cases based on the type of products. It also offers automatic case-load balancing to avoid hampering the efficiency of underwriters.

Document Submission Flexibility 

Borrowers and brokers want to have complete flexibility in the way they are submitting the documents. RevEl facilitates both online & offline document submission.

Real-Time Application Tracking

RevEl Tracker:

This module facilitates near real-time status tracking of key metrics for loan processor and underwriter. The data is consolidated across POS, LOS and back-office systems.

It enables an artificial intelligence dashboard to highlight key metrics of the application, along with bringing real-time alerts and notifications for any pending actions to be taken or any SLA breaches. There is a conversational chat interface to address ad hoc queries from customers. 

Using a product like Digilytics RevEl 

Let’s explore the technical value additions RevEl brings to the table for lenders.

• AI technology: State of the art technology to facilitate document processing reduces manual processing time by 60 % 

• Modular but integrated: Solutions for individual business lines, such as consumer finance and mortgages, can be deployed independently of each other 

• Product flexibility: Digilytics RevEl is a self-service platform with configurable capabilities. This means the flexibility to launch new products with ease.

• Technological superiority with easy bolt-on: The solution is based on a Modern technology stack, which can be easily integrated with existing LOS systems without upgrading business infrastructure.

Functional benefits of RevEl

• Enhances efficiency: This product is targeted towards reducing cost at every stage. Along with that, the time to offer is accelerated, and human efforts are drastically reduced.

• Processing Time Reduced: RevEl effectively brings down the Time to Fund by 40% conservatively.

• Consumer experience: Digilytics understands that borrowers care about access to credit. The Digilytics manufactured journey to that is fast, easy, and seamless.

  
  
  
  

• First-time right application: Submitting an acceptable application with correct documentation can be strenuous. Digilytics assisted applications are error-free and convenient.

  
  
  
  

• Lower cost to originate: With Digilytics Solutions, the costs are at least 30% lower

Conclusion

The post-pandemic has created strong competition amongst lenders, which makes it imperative for lenders to adopt faster solutions. This is the only way to deliver on the promises made to borrowers, increase the business reputation and gain more market share. That means more than 90% of loans need to be First Time Right.

About Digilytics

Digilytics™ RevEl for Financial Services is an AI-enabled revenue elevation product built on the Digilytics™ platform. It is available as a SaaS-based solution with an artificial intelligence dashboard and can be plugged into any loan origination system (LOS). For more information about Digilytics RevEl, visit our dedicated website here.