The materials available in this article are for informational purposes only and not for the purpose of providing legal advice. You should contact your own advisors with questions regarding the fraud risk content herein. The opinions expressed in this article are the opinions of the individual authors and may not reflect the opinions of MeridianLink, Inc. The opinions expressed in this article are the opinions of the individual authors and may not reflect the opinions of MeridianLink, Inc.
Approving more borrowers and managing fraud risk sound like opposing goals. But they don’t have to be.
When fraud appears, many financial institutions respond by adding more verification steps. When losses spike, they tighten rules. It feels like risk management. But here’s what’s actually happening: you’re treating the symptom, not the disease. And every layer of friction you add chips away at your approval of good borrowers.
This tension came up recently during a MeridianLink® webinar on lending approval strategies, and it comes down to one insight: adding friction isn’t the same as reducing fraud risk. More often, it just hides the real problem. Here’s what that means for your lending operation.
Why confidence is getting harder to hold
Several pressures have converged at once, and they compound each other. Borrower expectations are climbing. Fraud is getting more sophisticated. Margins are tight. And lean teams are being asked to do more. Each is manageable on its own. Together, they make every decision harder to trust.
The digital journey is where these pressures meet. Consumers overwhelmingly prefer digital channels for their banking, and they want the same instant experience they get from every other service on their device. But the wider that digital door opens, the more room it creates for bad actors—which brings the fraud pressure right back into view.
The controls many institutions built over the last fifteen or twenty years—credit checks, knowledge-based authentication—were designed to catch stolen identities. They still have value, but they no longer cover the full range of what institutions face today.
Synthetic fraud is the clearest case. It’s a manufactured identity built from fragments of real consumer data and cultivated for months or years until it looks completely legitimate by the time it reaches your queue. Controls built for stolen identities were never meant to catch a fraudster who, on paper, looks like a real person.
US synthetic fraud losses reached nearly $3 billion last year, up from $1.8 billion five years earlier. AI continues to accelerate the trend, driving down the cost of producing convincing government IDs, pay stubs, and bank statements
For institutions feeling the impact of this, the first instinct is to check everyone harder. But that approach comes at a price.
The friction trap
When fraud does get through, the losses show up in the loan portfolio—and the natural response is to bolt on safeguards like extra stipulations and additional income documentation. Each step is meant to protect against early payment defaults. Each one also slows down every clean borrower in the funnel.
The cost of that friction is measurable. In account opening, if the process takes more than three to five minutes, roughly half of applicants abandon—and research shows many never come back. Long funding timelines do the same damage. Every minute between application and decision erodes the experience for a borrower who fully intended to say yes.
So reacting to back-end losses with front-end friction punishes the wrong people. The synthetic identity built to pass your checks sails through. The qualified borrower who hits a wall of stipulations walks away. And because institutions often misclassify synthetic fraud as first-party fraud or a credit loss, they go looking for the wrong fix entirely—adding more of the friction that drove the good borrower off in the first place.
Know your funnel before you fix it
A better starting point isn’t adding a new tool or stipulation to your workflow. It’s an honest read on your own numbers. You don’t want to shadow box against a problem you don’t understand, and most institutions can’t see their friction until they measure it.
Two numbers are worth watching closely:
- Application-to-approve time: how long it takes to say yes.
- Approved-to-fund time: how long it takes to turn that yes into a funded loan, and how many approvals actually convert.
The relationship between the two tells a story. A strong decisioning ratio paired with weak pull-through usually means friction is bleeding good applicants somewhere in the process. Application length is another quiet culprit. Every extra question asked upfront is another reason to abandon.
MeridianLink can supply benchmarks around these metrics, so an institution can see how it compares to peers. That outside view often surfaces the gap faster than internal review alone.
Here’s an example
In a full road-mapping exercise, MeridianLink customer—Credit Human—outlined its current state first, identified bottlenecks and friction points, and then looked at how to optimize its existing MeridianLink configuration before layering in anything new.
That process revealed a pattern: Signal problems and workflow problems are intertwined.
Credit Human was manually reviewing nearly every application because it lacked the right signals to screen borrowers earlier. The missing signals created the workflow burden. Once the signal is fixed, the workflow burden eases on its own.
The thought of reworking your fraud risk processes can feel overwhelming. But in our experience, there’s a more manageable path: treat it as a series of drills. Work each stage of the workflow on its own, identify what can be improved, and prioritize from there.
The right signals, matched to the right questions
Strong decisioning and fraud risk management rest on the same logic: verify that a person is who they say they are, and that their documentation supports what they’ve told you. MeridianLink’s partner marketplace helps financial institutions bring those signals directly into the lending and deposit account opening process.
Identity partners like Socure and LexisNexis®, along with the bureaus, deliver the early fraud signals—third-party, first-party, and synthetic—that also support KYC and OFAC requirements.
Income and employment solutions such as Experian Verify™, The Work Number®, and Plaid bank-data verification give teams confidence in stated income without forcing borrowers to dig up more paperwork.
AI-based underwriting and fraud risk solutions from partners like Zest and Point Predictive can read alternative signals—utility and rental payment history, for example—that help approve thin- and no-file borrowers a traditional credit model would pass over.
MeridianLink’s advanced decisioning engine pulls those signals together through a decision-tree approach, handling the kind of nuance a single rule set can’t. Take an auto loan: the size of the down payment changes how you read everything else on the application, because equity and collateral value shape the credit profile and income you need to see. Rather than defaulting to manual review the moment something falls outside a neatly defined box, a decision tree routes each application down the right path based on those inputs.
Fraud orchestration takes this is a step further. Instead of running every signal at once, orchestration evaluates them in sequence, so a clear decline can stop the process early without burning through every check. It’s a more prescriptive way to manage risk in stages.
The Prove integration in MeridianLink® Access shows the approach in action. An applicant enters a phone number, receives a text, and confirms a couple of data points. Prove’s risk algorithms assess the legitimacy of the person behind the scenes and return an authenticated data payload that prefills the application. Good applicants get a fast, near-frictionless start. Anything suspicious gets flagged for closer scrutiny before it goes any further.
The payoff
Salal Credit Union put this into practice and the numbers moved fast. After bringing MeridianLink Access to its members, the credit union saw application processing speed climb 50% and its NPS jump 27 points. Approvals got faster, and members liked the experience enough to recommend it.
That’s the goal most institutions are reaching for: digital applications that run end to end, with the clean ones never touching a manual queue. As straight-through processing rises and time to decision shrinks, community and regional institutions get the speed to go head-to-head with fintechs and the biggest banks.
Where you start
It comes down to two moves.
The first is practical: map where you are against where you want to be, then stage the work and execute one point of friction at a time, with partners supporting each stage.
The second is a mindset shift: deciding the status quo is no longer acceptable. There’s usually a reason an application grew to 30 questions and inertia keeps it in place long after that reason stops applying. Institutions that go in with an open mind regularly surprise themselves with how sophisticated they can become.
The tools to approve more of the right borrowers while keeping the wrong ones out already exist, and MeridianLink brings them together in one place. The institutions getting it right start with their own data, treat signal and workflow as one, and use what they learn to refine their credit decisioning and fraud protection over time.
The next step is seeing what this could look like for your institution. From our digital lending platform to our vast network of integration partners, we can help you put these pieces to work.
