In today’s digital economy, data has become the essential resource that creates the engine powering financial institutions (FIs), credit unions and FinTechs.
Yet in an environment where fraud is escalating, government data pipelines are narrowing and artificial intelligence is reshaping decision-making, the urgent question is no longer whether data matters but whether firms can trust the signals they receive.
That’s the challenge explored in PYMNTS’ “Searching for Reliable Signals in Banking’s New Data Reality” series, part of the What’s Next in Payments coverage.
The series draws insights from leaders across four firms in the financial services ecosystem: David Durovy, senior vice president of transformation at i2c; Jeremiah Lotz, senior vice president of enterprise data and experience design at Velera; Pradheep Sampath, chief product officer at Entersekt; and Kyle Becker, chief financial and risk officer at Concora Credit. Together, their perspectives offer a window into how data is being reshaped as the foundation and frontier of banking.
Durovy said financial providers must ensure AI does not take the “51% seat” in decision-making, warning that without human-led oversight, firms risk losing control of data integrity. Credit unions are compensating for shrinking government datasets by leaning on first-party transaction data and consortium partnerships.
Sampath said financial crime requires data to be “a team sport,” blending traditional sources with faster, alternative signals.
Meanwhile, Becker described how alternative data, particularly cash flow underwriting, is expanding credit access while strengthening fraud defenses.
What emerges is a common theme. Data remains indispensable, but it is evolving. Institutions that strike the right balance between traditional and real-time sources, which share intelligence responsibly, and that deploy AI with governance will be best positioned to compete. Reliable signals are not given; they are built, safeguarded and continuously validated.
Fraud Defense Requires Data, Speed and Trust
Fraud has become an ongoing challenge for financial institutions and FinTechs, and executives across the series underscored data’s central role in the fight. As Sampath put it: “It takes a village to fight fraud.”
Traditional data sources, from the Federal Reserve’s payment fraud reports to suspicious activity filings at the Financial Crimes Enforcement Network, have long formed the backbone of information used in risk monitoring. However, Sampath cautioned against relying on these sources alone.
“We need to detect suspicious trends faster by looking beyond just government data feeds,” he said.
The solution is to blend historical signals with real-time behavioral insights, device fingerprints and geolocation markers, he said.
Velera’s Lotz agreed that fraud prevention requires immediacy and perspective.
“Real-time data is certainly going to ‘win the moment’ in many cases, but historical data is going to ‘win’ the strategy,’” he said.
This allows AI models to spot anomalies in the moment while grounding those alerts in broader patterns.
Durovy stressed the importance of human-led “data intelligence” in fraud defense.
“We all need to have that data intelligence capability to see the first-party data, to make sense of it, to ensure our third-party data is not driving aberration in our portfolios…,” he said.
Fraud defense mechanisms increasingly must be intertwined with credit underwriting. Alternative means of bolstering defenses, such as cash flow verification, “not only helps lenders gauge repayment ability but also doubles as a fraud defense,” Becker said.
Across the perspectives of the four executives that weighed in for the What’s Next in Payments series, the consensus was that fraud cannot be stopped with old data alone. By fusing legacy signals with faster, alternative sources and maintaining governance, FIs can keep pace with criminals while protecting consumers.
Data Collaboration Without Collusion
While FIs compete in the marketplace to forge new customer relationships and keep existing ones intact, executives said fraud prevention and data reliability are areas where cooperation can benefit all. Durovy called fraud risk management “a rare, competitive-neutral space where institutions can share intelligence without compromising proprietary advantage.”
Velera has put this into practice, aggregating shared data across 4,000 credit unions.
“My advice is not to be afraid of consortium data,” Lotz said.
Consumers already expect their financial providers to use their data for safety and experience, so when it comes to transforming that expectation into coordinated defense, the opportunity is significant, he said.
Sampath described consortium models as essential for combating fraud.
“Where the industry needs to go forward is to share data in a responsible manner across consortia, while preserving fair competition,” he said.
Privacy-enhancing techniques, including encryption and federated data models, allow institutions to exchange risk intelligence without exposing sensitive details, he said.
Becker underscored the benefits of layered intelligence. By continuously evaluating and integrating about a dozen new alternative data sources each year, the company strengthens underwriting and fraud detection.
“We often find one or two [new data sources] per year that we add, and we just keep layering that into this stack … that makes us a little bit better over and over and over again,” he said.
No single dataset, institution or tool can address fraud and risk in isolation. Collaboration is not collusion; it is resilience. By contributing anonymized insights into shared pools, firms collectively raise the bar against fraudsters.
“And the thread that binds us all together is data that’s actionable, shared in good faith, and governed responsibly,” Sampath said.
Building Layers of Data Intelligence
The series also explored how institutions are balancing traditional data with alternative sources of information previously unavailable to enterprises to strengthen decision-making.
For Durovy, legacy data such as historical bureau files and first-party transaction records remain “indispensable” for underwriting and compliance.
“When you’re doing things like underwriting, you can’t use unproven data,” he said.
Yet traditional sources are under pressure, with government statistics shrinking and consumer behaviors evolving. That’s where alternative data is gaining traction.
“Alternative data is super useful because it allows you to maintain or reduce risk while also providing access to credit to more people,” Becker said.
Cash flow underwriting, for instance, offers a real-time lens into repayment ability while serving as a fraud check, he said.
Lotz said the strongest strategies combine historical and real-time data.
“We want to use real-time data and real-time transactions to be able to understand what’s happening in that moment for that consumer,” he said, while also looking back at “retro” data to spot long-term trends.
Sampath echoed that dual perspective with a nod to the balance between what went before and the information that’s being used in real time.
“Historical data is the foundation; any model you have today needs that history,” he said. “But we also need real-time risk radars because threats evolve every week.”
These “radars” include behavioral analytics, device fingerprints and geolocation trails, he said.
Ultimately, executives described data not as a single pipeline but as a layered system. Traditional bureau files, proprietary transaction data, commercial datasets and alternative signals each play a role. The opportunity is in integration, or knowing when and how to apply each source responsibly.
“The credit bureau data will always matter,” Becker said. “But I think alternative data is going to play a big role.”
