The lending industry is undergoing a fundamental shift—from manual document review to real-time risk intelligence. At CBA Live, Jessica Pillado Gonzalez shared how that shift is already playing out across credit, operations, and fraud teams. Not theory. Not hype. What’s changing now—and what comes next.
The Old Rule Still Applies. But It Looks Different Now.
“Trust but verify” still holds—but verification no longer means friction.
For years, the industry equated verification with manual review: stipulations, delays, and back-and-forth. It was seen as a necessary cost of managing risk.
That model is breaking.
Verification today isn’t about slowing things down. It’s about making smarter decisions, faster—replacing guesswork with real signals and moving forward instead of reverting to manual workflows.
Fraud Isn’t Just Growing. It’s Evolving.
What used to take skill now takes minutes.
AI-generated pay stubs, utility bills, bank statements, and phone bills are clean, convincing, and scalable. Fraud is no longer a niche problem—it’s accessible to anyone with the right tools, and it’s accelerating quickly.
We’re seeing sharp increases in document manipulation driven by generative AI. Not incremental—exponential.
The uncomfortable truth: human review alone is no longer enough. The eye can’t reliably detect what AI can now create.
AI vs AI Is No Longer Optional
If fraud is being created by machines, it has to be detected by machines.
This isn’t just about efficiency—it’s about survival. AI can identify patterns that humans cannot see:
- Formatting inconsistencies across documents
- Metadata anomalies invisible to the eye
- Reused templates and known fraud signatures
- Subtle structural misalignments
More importantly, it connects the dots. Because fraud doesn’t happen in isolation.
The Power of Data Collisions
One of the most overlooked fraud signals is repetition.
The same employer ID. The same transaction pattern. The same document structure—appearing across different applications and lenders. These aren’t coincidences. They’re signals.
AI turns those signals into what Jessica described as “data collisions,” or moments where patterns intersect and expose organized fraud.
This is where verification becomes intelligence. Not just checking a document, but understanding behavior across a network.
The End of Static Risk Models
Here’s what many lenders are underestimating: fraud is evolving faster than risk models can keep up.
If your credit risk model updates annually, you’re already behind. Fraud isn’t annual anymore—it’s real time.
That gap shows up in early payment defaults, in losses that surface months later, and quietly in the P&L.
The institutions that win won’t be the ones with the best models last year. They’ll be the ones that can adapt continuously.
Network Intelligence Is the Unlock
For years, data stayed siloed, with each lender fighting fraud alone. That approach no longer works.
Fraud rings don’t operate within a single institution—they move across lenders, channels, and time. Detection has to do the same.
Network intelligence changes the equation. Patterns become visible across institutions, fraud can be stopped earlier, and signals compound instead of staying isolated.
This is the shift from reactive to proactive—from individual decisions to ecosystem awareness.
Compliance Is Catching Up to Reality
With AI comes responsibility—and scrutiny.
Lenders are asking tougher questions: How is the model making decisions? Is it fair? Is it explainable? Can it stand up to regulatory review?
This is where vendor expectations are evolving. It’s no longer enough to provide technology—you have to provide accountability.
AI policies, fair lending testing, and third-party validation are becoming table stakes. If AI is making decisions, it must also withstand examination.
Underwriting Is Becoming Real-Time Intelligence
The biggest shift may not be fraud detection—it’s what happens after.
AI is transforming underwriting from a static checkpoint into a dynamic system. Decisions happen in seconds, not hours. Straight-through processing becomes the standard, not the exception. Insights are delivered in natural language, not just pass or fail.
This isn’t just faster underwriting. It’s better underwriting—with more context, more confidence, and less noise.
The Competitive Line Is Being Drawn Now
Two paths are emerging.
One continues with manual review, legacy checks, and slower decisions. The other embraces AI to move faster, reduce friction, and increase confidence.
The gap between them is widening.
Because speed matters. Experience matters. And risk, left unchecked, compounds. Early adopters aren’t just improving operations—they’re changing the economics of lending.
Final Thoughts from CBA Live 2026
Fraud doesn’t announce itself at origination. It shows up later—in defaults, losses, and questions no one can answer after the fact.
The shift Jessica outlined is simple, but not easy: see risk earlier, act on it faster, and do it without slowing down the customer.
That’s the new standard. And it’s already here.
