Thank you for Subscribing to CIO Applications Weekly Brief
Thank you for Subscribing to CIO Applications Weekly Brief
In an interview with the editorial team of CIO Applications, Tim Grace, CEO of Point Predictive, discusses at length how his company’s robust technology solutions help automotive, mortgage, retail, personal lending, and student loan finance companies curb loan fraud effectively by identifying consumer loan applications that have truthful and reliable information.
Could you provide a brief overview of Point Predictive?
Point Predictive takes a unique approach to Ai development for financial services. Our experience in risk management ensures that our Ai is usable, effective, and safe for application in consumer lending. Combining Artificial + Natural Intelligence [Ai+Ni], Point Predictive carefully develops and tests machine learning (ML) models that help lenders know with confidence if the information on a loan application is true and accurate or the loan application contains misrepresented or materially inaccurate information. Human experience is the “natural intelligence” edge of the sword and is uniquely integrated into these ML models, resulting in unrivaled identification of truthful and accurate loan applications. This results in more and faster loan approvals, with less documentation and verification required. Point Predictive’s technology helps lenders profitably grow their loan portfolios while reducing losses related to misrepresentation and fraud and the costs associated with preventing them.
Point Predictive has collected and maintained a growing number of unequaled data assets. This data is the result of powerful orchestration of nearly 100 million historical loan applications (with lender confirmed outcomes and performance) from a cross-section of all lender types. This data drives billions of proprietary derivative attributes from lender-provided Consortium data. There is no other collection of application data remotely similar or as robust. The additional benefit comes from adding carefully sourced third-party data tested to provide complementary performance lift to our predictive models. Finally, the continuous feedback of application and loan outcomes from lenders is used to keep our scores and alerts fresh and relevant since that performance data grows by millions of applications and loans each month. This unique combination of industry data and derived features and attributes powers our ability to precisely identify truthful information from misrepresented information on loan applications.
Our team has spent almost three decades building and helping lenders and credit card issuers successfully implement predictive machine learning scores built on data contributed through industry consortia. This team is considered the authority on fraud and risk analytics solutions across industries and around the globe. Drawing upon these decades of ML model creation and technology solution deployment across credit card issuers, retail, mortgage, and auto lenders, Point Predictive scientists and technologists know what works operationally and how to create innovative, reliable solutions for lenders.
Our team builds and delivers predictive models that enable critical real-time production decisioning.
Our team has built and delivered hundreds of predictive models supported by the most comprehensive data consortiums that enable critical real-time production decisioning across hundreds of financial organizations worldwide
What are some of the major market pain points your company addresses?
Most lenders are facing pressure to safely and profitably grow their loan portfolios. With the pandemic in perspective, it is still a historic time for lenders and consumers as interest rates are still low and demand for cars, homes, higher education, and alternative financing remains robust and growing.
However, hot economic conditions inevitably invite fraud, compelling lenders to strike the appropriate balance between growth and risk. On the risk side of this balancing act are a large and growing number of nefarious third parties – hackers, fraud rings, and even unscrupulous third parties that can misrepresent borrowers or defraud lenders with incorrect or defective information on the loan application. With these schemes getting easier to perpetrate, lenders need a technological edge in this fight. Our Ai+Ni technology gives them predictive horsepower to accelerate and reduce the cost of loan origination while preventing substantially more fraud from being funded. Our lending industry Consortium data, securely contributed by member lenders, is a living data science laboratory that constantly adjusts to mutations and variants in fraud patterns, uncovering suspicious connections across risky loans. Member finance organizations are firmly convinced that predictive technologies and data science like ours have a large role to play in the future of lending.
Could you shed light on the features of your Predictive Solutions?
There are two core features of our predictive platform. Firstly, the Consortium data set, which is enhanced with a wide range of specialized, publicly available data, including Census Bureau, Small Business Administration, Social Security Administration, vehicle, broker/dealer, and consumer-provided contact information.
Secondly, Ai+Ni is the patented machine learning approach. It predicts the trustworthiness of loan application information on the basis of its links, connections, and patterns to prior fraud, incorporating human intuition to overcome variants and mutations in fraud patterns.
Together, this technology platform powers a range of scoring, alerting, and analysis products that provide a new level of visibility and transparency into the data submitted on loan applications. For example, lenders who wish to determine the likelihood of overstated income rely on IncomePass to assess that risk, while those who wish to protect themselves against fraudulent assertions of employment use EmployerCheck to ensure that the borrower’s employment can be assumed true. On the other hand, our comprehensive Auto Fraud Manager powers decision improvement and automation with a score and alert report that auto lenders use to cost-effectively clear applications of burdensome stipulations.
Could you share a customer success story?
One of our clients, MidAtlantic Finance Company, has seen several profitability gains using our Auto Fraud Manager solution. High scoring applications above the lender’s desired risk threshold are routed to review with a human analyst, protecting the organization against the costliest and hardest-to-prevent fraud. Low-scoring applications are routed for faster funding, ensuring profitable growth of the auto loan portfolio. MidAtlantic submitted 106 alerts on almost $874,000 in loan amounts in just over six months by implementing Auto Fraud Manager.
What, according to you, steers you ahead of the market competition?
Point Predictive has a unique advantage in terms of the nature of the proprietary data contributed by a large number of lenders. The predictive power that lies in Consortium data is not limited by lack of visibility into non-funded loan applications, where many more fraud signals lie. Furthermore, on funded loans, traditional credit data is only sparsely populated with fraud outcomes due to the effort involved in confirming that outcome.
The Consortium approach makes Point Predictive an equally potent partner in the fraudtech ecosystem. Consortium data could complement a host of document authentication technologies, identity verification solutions, or device authentication technology to round out the trust/security/privacy/loss prevention regime that is imperative for modern, digital lenders. Naturally, our fraudtech platform integrates directly to a loan origination process with a proprietary API and standard integrations through popular loan origination systems such as defi SOLUTIONS, First American, and MeridianLink, among others.
What does the future hold for your company?
Currently, the Point Predictive Consortium addresses the most common fraud and misrepresentation risks facing consumer lenders: identity fraud, income and employment misrepresentation, collateral misrepresentation, broker/dealer/correspondent risk, and co-borrower/co-signer schemes. However, as we learn more from our customers, it is becoming clearer that Consortium data, along with Ai+Ni and ML technology, has potency outside of a traditional indirect home, auto loan, or direct online loan application.
The growth we are experiencing has convinced the leadership team that the platform can be extended into adjacent financial vertical markets. Beyond that, the volume and mobility of applicant-provided data lead to all sorts of trust issues that murky the waters for any company extending credit to consumers. While the vast majority of borrowers tell the truth, the minority of fraudsters consumes far too many resources prior to funding and customer acquisition. Point Predictive’s data science allows organizations to spend less to get a more profitable portfolio.