The client is a US-based mortgage lender with decades of experience in lending. They specialize in wholesale and correspondent lending. Mortgage brokers and loan originators trust them nationwide. With over 70 product offerings and exceptional service, they are often known as one of the top lending businesses in the country.
As their customer base grew, an increasing number of documents arrived. The client required a more efficient method for organizing borrower documents before entering them into their Loan Origination System (LOS). These documents are large PDF files with multiple forms uploaded at various stages of the loan cycle.
To minimize operational friction and maintain SLA compliance, the lender sought Flatworld.ai to implement a smart indexing system capable of automatically segregating, classifying, and uploading documents in real-time.
The manual document indexing was taking a long time, and thus, making the turnaround time longer too. This was slowing down the loan closure process. The borrower's documents came in a single PDF file with all the information merged together. Then, the internal team had to review each file manually, split it, and classify it. This process was not only time-consuming but also prone to error.
The business needed a solution that could do the following :
The first step was to work closely with the lending company's business and operations team and understand the challenges within the present document management flow. Flatworld.ai identified the loan lifecycle stages and the main documents required. Automating high-frequency file tasks saves time and money. Additionally, it was crucial to define clear project milestones at this stage and establish feedback loops. They also planned for phased UAT and ongoing support after deployment.
Once the planning phase was complete, Flatworld’s team collected real-world loan files with diverse situations, which were used to train the MSuite’s classification engine. MSuite is our AI-powered document classification and indexing engine.
Training the classification engine enabled the creation of a robust, accurately labeled dataset for model development. It also helped tune classification limitations to minimize false positives. Edge cases, such as handwritten forms and low-resolution scans, were simulated. Finally, the Flatworld.ai team tested the model in a controlled sandbox to ensure the output quality.
MSuite configuration was aligned with LOS data fields. A LOS is a mortgage workflow platform designed for processing and compliance. Collaboration with the client's LOS necessitated careful consideration of both system compatibility and data integrity.
The team performed extensive UAT alongside the client, tracking accuracy and turnaround across different file scenarios.
MSuite was launched in production with complete monitoring and fallback safeguards. The client’s live loan packages were processed in real-time, with metrics reported daily.
Despite the successful rollout, the team faced several critical hurdles -
Manual indexing is a lengthy process and can’t always be dependable. The client was looking for a solution to end this, and Flatworld.ai delivered. Now they have a reliable system that works smoothly 24/7.
Flatworld’s MSuite delivered a measurable business impact within just weeks of launch
MSuite is helping them meet SLA deadlines, and the indexing process is consistent, monitored, and has a higher accuracy level. These factors are assisting the mortgage lending company in staying ahead of its competitors in a market where turnaround time and compliance are crucial.
Ready to Speed up Your Back Office with AI?
Talk to Flatworld.ai solution experts about how MSuite can help you organize your workflows, without losing accuracy or control.
Mortgage firm cut response time by 80% and tripled conversions using Telena AI for program queries.
Know More
Mortgage lender automated document indexing with MSuite, achieving 80% time savings and 99% classification accuracy.
Know More
3D printer brand automated 90% of support emails with Telena AI, achieving over 85% CSAT score.
Know More