Financial management decisions affect every part of a business, but can only be done effectively if the data used to inform these decisions is complete and accurate. The smallest of mistakes in data integrity can lead to serious legal consequences, which explains the need for stringent compliance policies and rigorous regulatory oversight. Therefore, data integrity auditing is fundamental to ensuring the completeness and accuracy of information.
In the finance sector, a post-closing data integrity audit will be carried out to identify any loan defects or red flags in mortgage quality control, and the impact of the pandemic caused defect rates in 2020 to increase by 21%. Establishing a robust loan completion risk mitigation strategy is challenging yet critical in the mortgage industry. However, still common to many organizations is the vast number of manual processes required of loan auditors, and the various methods or financial management software is used to capture, share and store information.
With this background, this particular organization was keen to launch a single, integrated platform that could be used for the entire client journey, from onboarding to post-closing. With custom app development in mind, the intention was to migrate to an online application where all data would be stored, processed, and accessed, moving away from manual, error-prone, and time-consuming auditing.
Before the fintech app was built, in the absence of a unified digital platform, the organization was experiencing the following challenges:
- No single source of truth:
Without a single source of truth, audits were carried out by collating information contained within various Excel spreadsheets and loan packages, all being shared via email and FTP. With multiple team members involved in updating information, shared spreadsheets quickly became out of sync, or offline data islands created.
- Time-consuming process:
When conducting an audit, the auditor had to manually sift through loan packages, consisting of up to 500 pages, to pick out and identify relevant documentation. On average, an auditor would spend two to three hours to complete an audit, provided all the relevant documents were available in the loan package.
- Inefficient and error-prone audits:
The entire post-closing audit, consisting of 118 checkpoints across 91 documents, had to be done manually, increasing the likelihood of errors. In the case of missing or incorrect documentation, the auditor would have to contact the client and follow up, increasing the audit turnaround time and associated costs.
- Limited oversight:
The lack of complete data oversight created a significant challenge for audit managers trying to track or monitor audit progress. There was a huge dependency on individuals to own the audit process for each loan, manually generating and sharing team utilization and efficiency reports.
- Defect analysis:
Due to stringent financial management compliance policies overseen by regulatory authorities, clients expected the post-closing audit to highlight critical defects in the loan process. Unfortunately, it was not possible to support the calculation of defect rates or view trends, since there was no collated data available on each loan processor.
Having discussed and reviewed the customer’s challenges and current processes, SourceFuse proposed custom-building an advanced post-closing audit optimization solution, leveraging unrivaled AI technology. Working collaboratively with them, SourceFuse took a step-by-step approach through this custom app development project:
- Firstly, SourceFuse took the time to understand the post-closing audit and documentation required for each stage. While doing this, the team was able to appreciate the various audit and financial management nuances and developed in-depth knowledge and appreciation of the domain and audit process.
- Initially, the customer was keen to use RPA for automation but hadn’t appreciated the associated costs implications. Using RPA, each bot would increase operational costs by as much as $15,000 per year. For a new fintech app build, SourceFuse proposed using AI technologies, such as natural language processing (NLP), computer vision, deep learning, and machine learning, to significantly reduce and optimize costs.
- SourceFuse then identified different aspects of the post-closing audit process that could leverage machine learning and architected the solution to identify documents, extract relevant data, and then automatically apply audit-based business rules.
The end result was a bespoke unified, digital fintech app that addressed all of the customer’s operational challenges associated with the post-closing audits. SourceFuse’s team of cloud modernization experts provided consultative support throughout the project. They were able to identify novel ways to further streamline workflows and suggest additional post-closing audit optimization that would truly elevate the end-user experience.
Solution Key Features
The AI component of the platform has the ability to index documents and populate an initial audit report for a loan, enabling higher productivity across auditors. The solution has led to faster turnaround times, from 120-150 mins down to 20-30 mins. It also assigns a ‘confidence’ score to each checklist which enables users to prioritize time spent on a loan.
Human in the Loop
The platform is designed on the ‘human in the loop’ model. Every data point extracted and populated by the machine learning model can be reviewed and overwritten by a user. Any corrective actions are stored and used for retraining the models, continuously improving the accuracy of the machine learning component.
Centralized Platform for Loan Management
The platform now acts as the customer’s structured and centralized repository for all loans across its organizations. The platform enables users to upload, process, and audit loans in a single location. It acts as an automated document management solution, eliminating the need for any indexing effort. Accurate real-time data associated with a loan can now be viewed at any time by an auditor.
Dashboards and Reporting
The platform provides the ability to gain intelligent business insights, enabling informed business decisions. The user now has a complete overview of the audits, their status, and the option to segregate based on upload date, process status, audit in charge, etc. The new fintech app automatically generates various reports, presenting metrics and trends in a meaningful and impactful way, helping the customer to improve efficiency and increase compliance.
User Management and Access Control
The platform caters to various user roles, granting different roles the relevant access permissions or editor rights. For example, the Administrator governs the platform’s usage and onboarding of business users. The Audit Managers are in charge of the loan audit process, with the Auditors reporting to them, while the Loan Processor can view the audit results and intervene when requested.
The platform enables streamlined collaboration between Auditors and Loan Processors. For example, if there are missing documents or identified errors, the Auditor can tag the Loan Processor in the system, to request additional documents or clarifications. The Loan Processor can then respond via the loan comments section, or upload the necessary documents. In addition, an Audit Manager has a clear overview of audits being processed by their auditing team and can reassign loans when required.
Each and every action taken by any user is logged for compliance purposes. Detailed audit reports can then be automatically generated, based on users or by loans. This enables complete transparency across the organization and improves compliance.
In the financial management industry, this new fintech app solution is now one of the most advanced intelligent, post-closing audit platforms for the definition, identification, management, and communication of loan processing risks throughout the loan processing lifecycle. Unparalleled qualitative and quantitative AI techniques reduce human errors and provide a real-time loan profile supported by multiple data sources.
Driven by industry’s best practice, cognitive computing is woven into the fiber of the new platform, extending the solution to automate repetitive manual tasks and assist human decision-making. It enables clients to reduce cycle times, improve quality and provide cost efficiencies while ensuring global and local compliance.
In addressing the customer’s key challenges, the new application achieved:
- 70% reduction, on average, in post-closing audit times
- Elimination of document review negligence, producing an error-free audit
- 50% potential cost savings on TCO
- Access to real-time data, via automated document management
- Defect identification with higher compliance
To view full technical specifications, click below and download the Case Study PDF.