Trust Multiplied: How Verified Data Platforms Strengthen Business Lending

During the pandemic, the government extended a $1.2 trillion lifeline through the Paycheck Protection Program (PPP). Four years on, the sobering stats show that the Small Business Administration was victimized by fraudsters who made off with hundreds of billions of dollars of loans.

They lied on applications, set up sham businesses and took advantage of a program that relied on manually-entered data that was then sent to lenders, from purportedly true tax filings to payroll data. If they didn’t get accepted by one provider, they could fudge data and move to the next unsuspecting lender and get the cash in hand.

What was missing was a single source of truth at the point of onboarding, where business data offered up a holistic view of a company accessible by all parties.

There are examples of digital “single sources of truth” in some corners of everyday life. For example, electronic medical records offer an immutable trail of medication, providers and patient data — and much of it is under control, permissioned by patients, so that the data are shared in a discreet and safe manner in standardized formats. There are only certain people who can provide, add or update records — and all stakeholders interact across platforms or ecosystems in an environment that melds public and non-public data.

The health of a patient and the health of a business, as Hany Fam, CEO of Markaaz, told PYMNTS CEO Karen Webster, should be verifiable in a setting where no one can hide the ball and distort the data.

There are many companies that aggregate data on businesses and provide that information to banks and firms — pulling in public data from the web, for example, or engaging in screen scraping. There are other firms, Fam said, that endeavor to make sure that no one’s “fiddled” with data such as bank statements or driver’s licenses. FinTechs and some banks have sought to take those data-collection efforts in house, passing them through black box models that result in a fragmented landscape of business intelligence.

In the lending arena, Fam said, data can be notoriously incomplete. Banks and other creditors will provide an enterprise a loan based on the individual owners’ personal credit score (which in turn hurts the individual as they, not the business, take on more debt), hindering the firm’s growth potential.

“None of this gets us anywhere near this notion of a single source of truth,” he told Webster of standard practices. “A lot of organizations [talk] about their large amounts of data, and about how AI [artificial intelligence] is going to change data management. What I don’t hear about is making sense of the data.”

Two-Sided, Permissioned Platforms

Markaaz, which has built a two-sided platform that has complied information about 542 million companies around the globe gleaned from 65,000 data sources (and growing) — has embarked on what Fam termed “data stewardship, which is verifying the best and the most accurate sources of data, and making decisions about what’s more likely to be right than wrong in that data.”

The platform, he added, is secure and unhackable, while remaining entirely under the control of the businesses, their owners and the permissions that are granted.

“We’ve also built a model that allows us to grade the likelihood that non-public information is accurate,” he added, while offering up a gestalt view of those firms, tied to 198 separate data points including financial information and beneficial ownership.

For the firms and banks examining those 198 data points, “you can look at the history [of the firm] and create a predictive view of the future,” he said, as those enterprises apply for credit or other services.

“This is the cleanest set of data that you can get,” he said. There’s also the advantage of letting end users clean up errors in data as they add non-public information that can be offered up to parties on the other side of the platform, which speeds onboarding, lending and other interactions.

“You have the ability to permission how these things get merged, how they are used and they are monitored … whatever I see about your business,” Fam said. “If you’re the business owner, you should be able to see the same things I’m looking at, as well.”

The golden record that takes shape, he said, is dynamic and proprietary to each relationship that exists between the two parties, and can vary according to the use case (such as cash flow-based underwriting).

Battling the Fraudsters

The single source of truth has proven its worth in ferreting out fraud.

Fam told Webster that Markaaz was recently given a list of convicted fraudsters who had stolen money during the PPP loan program — and using its platform (through back testing) found that it would have detected every single one of them, well before funds would have been disbursed. One of the criminals, he noted, had been associated with over 200 separate legal entities that had applied for PPP loans, and could not have been found through conventional methods deployed by Markaaz’s competitors. But Fam said the combination of AI and data stewardship illuminated the malfeasance.

Fam said Markaaz has tuned its response rates to under 2 seconds for normal verification cases, but is able to reduce that to 200 milliseconds if needed, “which means we can provide real-time verification leveraging all of our golden record infrastructure … that’s fast enough to interject that data mid-flow during a point-of-sale transaction.”

Asked by Webster how the Markaaz platform will gain visibility and use over the longer term, Fam maintained that at the point of transaction, the “first point of trust will come from a third party” such as a bank or insurance company that introduces an enterprise to Markaaz. “This needs to happen a couple of times,” he said, “before you say ‘OK, I’ve used this, it works and I trust it … this leads to systemic trust.”

The golden record of truth, he said, “is living and contextual between two parties … and encourages those parties to look at the same data and take an active interest in it.”

The post Trust Multiplied: How Verified Data Platforms Strengthen Business Lending appeared first on PYMNTS.com.