BankTech

TrackStar aims to improve credit scoring data with AI

Company believes removing inaccuracies in consumer credit lines can benefit lenders

Credit scoring data that lenders and banks receive for consumer loan products often contain inaccuracies, making it more difficult for financial institutions to determine which individuals qualify for mortgages and credit cards.

The artificial intelligence developed by TrackStar, a Chandler, Arizona-based company, will create millions of new data points originating from over 15 years of credit data from over 30,000 lenders in the U.S. The company will run its historical data against current scoring data from the three credit agencies to look for errors and inconsistencies. 

The impact of COVID-19 has resulted in millions of Americans seeking forbearances on mortgage and student loan payments and will result in “a lot of incorrect reporting and inaccuracies,” said TrackStar CEO Mark Miller, who was the former president and chief operating officer at Equifax.

The metadata created by TrackStar’s AI means that lenders can also provide consumers with more customized lending options, payment plans and deferment actions, according to Miller.

“The tool has never been available before,” he said. 

TrackStar collected the data over a 15-plus year period during which it was helping consumers with their credit. The company used that data to create a proprietary database on thousands of lenders and creditors, including credit card companies, personal loan lenders, mortgage companies, student loan lenders, banks and credit unions. The platform’s machine learning can track trends that can lead to sniffing out inconsistent and often incorrect data. 

One in every five Americans has a mistake on their credit report, which can prevent consumers from being approved for loans, according to the Federal Trade Commission. Many of these errors are the result of data entry by employees, but these mistakes are material and could stop individuals from being approved for loans, Miller said.

“Everything from a simple 30 day late to medical billing discrepancies, bankruptcies and foreclosures have been proven inaccurate,” said Clint Lotz, president and founder of TrackStar. “And with the fallout from COVID-19, we expect to see a lot more forbearance and loan modifications issues pop up on people’s credit reports for some time to come.”

With TrackStar’s data, lenders can see a 5% to 7% increase in approvals from applicants in their marketing funnel when they implement the company’s new API into their decision flow, Lotz said. The API is driven from TrackStar’s internal data.

Better applicant qualification can help reduce customer acquisition costs, as viable leads from lead aggregator sites average as much as $90 cost-per-click for people searching for loans or mortgages and can be costly for lenders.

“We believe in more transparency and more loans could be approved,” Miller said.

With over 200 million active credit files in the US today and 1 in 5 having inaccuracies, that’s about 40 million consumers as of 2019, Miller estimates.

“With the year we have been having so far, we expect that number to drastically increase, along with a significant number of people who modified their auto loan and mortgages six months ago when the pandemic first hit the U.S. widespread,” he added. “Many of them have balloon payments coming due in the next 30-60 days and given the overall state of the economy, expect to see missed and late credit card payments, which will ultimately impact credit reports.”

TrackStar’s API can be added to a financial institution’s existing infrastructure such as lending platforms, consumer finance applications, risk models and point-of-sale financing and aims to predict a consumer’s future lending qualifications. The AI can help determine which negative credit items could be removed from a customer’s credit history, including hacking and fraud incidents.  

The practice of deploying artificial intelligence or machine learning for credit risk modeling applications has grown rapidly in recent years. For commercial credit, Moody’s Analytics’ RickCalc product is helping lenders and creditors assess default and recovery risk for private firms, financial institutions and project finance transactions.

For consumer finance applications – like mortgage, automobiles, credit cards and personal lending – software companies like Zest AI are also addressing the opportunity to deploy machine learning to help lenders “…expand portfolios, reduce losses, optimize pricing, and drive greater automation across the lending cycle.”

The API entered beta on April 2, and after that testing period, the API recently officially launched.

Founded in 2004, TrackStar is a privately-held company. 

Latest Articles

Flagstar Bank to back trio of mortgage-tech startups
Dec 01, 2020 By

Flagstar Bank and the FinTech Consortium revealed that fintech startups Home Lending Pal, Stavvy and Real Key are the latest initiates to the companies’ ongoing MortgageTech Accelerator program.

Content from our partners

3d rendering of a row of luxury townhouses along a street

Log In

Forgot Password?

Don't have an account? Please