Powering Real-time Loan Underwriting at Vontive | Materialize



Vontive uses Materialize to power its loan underwriting engine. With Materialize, Vontive deployed 170 SQL underwriting rules to assess borrower eligibility in real-time, allowing underwriters to process loans faster.

With Materialize, Vontive reduced borrower eligibility checks from 27 seconds to half a second. This allowed underwriters to eliminate idle time, fund more loans, and increase company profit margins. 

About Vontive

Vontive is an embedded platform that connects real estate investors, lenders, and capital providers to deliver mortgages for investment properties. With a no-code, white-label solution, Vontive empowers B2C companies to launch an investment property mortgage operation in weeks.

Vontive serves as a complete system for investment property mortgages, combining the borrower journey, mortgage office, loan origination, and servicing into a single unified platform. By merging trusted brands with pre-built infrastructure, Vontive provides reliable debt capital in an intuitive platform. 

The Challenge 

Wolf Rendall, Director of Data Products, manages analytics and BI, dashboards, pricing, and borrower eligibility at Vontive. Wolf’s team is responsible for the rules and eligibility engine that powers Vontive’s underwriting operation. 

The rules and eligibility engine executes a series of eligibility checks that determine whether borrowers are good candidates for mortgages. This engine is at the core of Vontive’s loan underwriting process.  

When Wolf arrived at Vontive in 2022, the company’s underwriting rules were contained in a 23 page Google Doc. Vontive used Google Sheets to operationalize these rules, with varying levels of success. 

However, this spreadsheet solution was decentralized, not scalable, and too manual to serve as a durable engine for Vontive’s mortgage underwriting operation. The spreadsheet method not only slowed down underwriters, but also led to errors and inaccurate loan decisions. This cut into Vontive’s profit margins. 

To build a scalable rules and eligibility engine, the Vontive team wanted to move from spreadsheets to a cloud database. This would allow them to build a flexible, centralized rules and eligibility engine, defined by SQL, for mortgage underwriting.     

Before Materialize

Before Materialize, Vontive leveraged a main OLTP application backed by PostgreSQL, which replicated to an analytics instance of PostgreSQL with AWS DMS. Their SQL checks ran off of the replica. 

The new SQL-based rules and eligibility engine launched in February 2023. The rules and eligibility engine functioned as its own microservice. Vontive powered the engine with PostgreSQL on Amazon RDS. 

The team converted underwriting protocols into SQL rules to power the engine. During the underwriting process, borrower data was combined into tables, and replicated in PostgreSQL. The data was subsequently queried by the SQL rules to determine if borrowers met eligibility requirements.   

These eligibility determinations were sent to the front end underwriting application, so underwriters could use them to make loan decisions. Underwriters submitted borrower information on the front end, while the SQL checks determined borrower eligibility on the backend. Eligibility determinations were then fed back into the front end for loan decisioning.

The system ran on batch. But it was a batch-on-demand model, one that executed every time an underwriter requested eligibility determinations. Although the system was batch, the on-demand function allowed underwriters to receive eligibility determinations in about 30 seconds. However, in the rapid-fire market of loan underwriting, this speed bordered on unacceptable.  

In the beginning, the engine utilized about 10 SQL rules. The team wanted to convert the entire 23 page underwriting manual into SQL rules. While the initial results were promising, converting the entire manual would result in around 1,700 SQL rules. Unfortunately, the team found that after adding about 100 SQL rules, the load was too heavy for their PostgreSQL RDS. Eligibility determinations began to take upwards of 20 seconds, too slow for the rapid-fire pace of loan underwriting.  

That’s when the team started to investigate materialized views as an option. After discussing build versus buy options for some time, Vontive discovered Materialize.

After Materialize 

After researching Materialize, Vontive found that the data warehouse had both the speed and flexibility needed to power its loan underwriting operation. Materialize combines streaming data with SQL support, enabling lenders to continuously execute SQL underwriting rules and produce eligibility determinations with sub-second latency.

Materialize leverages streaming data to deliver real-time eligibility determinations to lenders. By employing change data capture (CDC), Materialize offers access to the most up-to-date borrower data. And with SQL support, analysts and non-technical users can easily develop SQL underwriting rules. 

Upon first using Materialize, Vontive leveraged the data warehouse’s full support for SQL, PostgreSQL, and dbt. Materialize is PostgreSQL wire compatible, allowing Vontive to easily migrate from its PostgreSQL RDS database. 

With full SQL support, Vontive ported their SQL underwriting rules over to Materialize without any delay, and a few painless modifications. And by harnessing Materialize’s dbt adapter, Vontive immediately productionized their SQL rules. 

The plug-and-play functionality of Materialize was a huge deal for us. We were able to move all our SQL models over to Materialize instantly, without having to rewrite them. It was a seamless process.
— Wolf Rendall, Director of Data Products, Vontive

Making use of indexes in Materialize, Vontive was able to greatly increase the number of SQL underwriting rules it employed. Vontive quickly added 70 new SQL rules — increasing the total load to 170 SQL rules. With Materialize, Vontive could add thousands of SQL rules without slowing down eligibility determinations.   

Within a matter of days, Materialize was powering Vontive’s entire rules and eligibility engine. Vontive saw a dramatic improvement in loan decisioning times as soon as Materialize went live. Before Materialize, eligibility determinations took 27 seconds. After Materialize, eligibility determinations took half a second. 

Our success was pretty head-spinning. It didn’t take long to set up Materialize, and once we did, we saw loan determination times drop to half a second. This was a lot lower than we anticipated. The reduction in latency allowed our underwriters to make decisions much faster.
— Wolf Rendall, Director of Data Products, Vontive

Materialize also offered the accuracy Vontive required to make profitable loan decisions. Other streaming approaches offer eventual consistency, which can lead to incomplete results, and bad loan decisions. That’s why Vontive needed the strong consistency offered by Materialize. With strong consistency, the results always match the input data, so Vontive can always make accurate loan decisions. 

The Results

  • 98.15% decrease in loan eligibility rule calculation time

  • 70% increase in SQL underwriting rules

  • Soft feedback: Underwriters work faster, with full confidence in the underwriting process

After launching with Materialize, Vontive saw several immediate improvements to its loan underwriting operation. 

Loan eligibility rule calculation time decreased by 98.15%, dropping from 27 seconds to half a second. 

The Vontive team increased the number of SQL underwriting rules by 70%, bringing the total from 100 SQL rules to 170 SQL rules. In the future, Vontive plans to deploy over 1,700 SQL rules. 

And the feedback from the underwriting team was enthusiastic. They loved the rapid decrease in loan eligibility rule calculation time, and clamored for more SQL rules to speed up the underwriting process.  

With Materialize, Vontive was able to port over all of its pre-written SQL logic, and start powering its loan underwriting operation with real-time data in a matter of hours. 

This led to a precipitous decrease in loan decisioning times, resulting in more funded loans, and higher profit margins. 

Try Materialize Free