CUSTOMER Q&A: Credit Genie

How Credit Genie uses network insights to support financial inclusion

For many consumers living paycheck to paycheck, a cash advance can be a valuable bridge. But these consumers can also be difficult to evaluate with traditional risk decisioning, especially when they are thin file, new to the financial system, or earning income in less traditional ways.

In this conversation from Plaid Effects, Niki Taylor, Product Marketing Lead at Plaid, joined Rohan Sriram, Product Lead at Plaid, and Shan Xu, Chief Revenue Officer of Credit Genie, to discuss how Plaid’s Cash Advance Index helps providers make smarter approval, sizing, and payment decisions using broader network signals.

This conversation has been edited for clarity and length.

Taylor: To start, Shan, can you tell us about Credit Genie and how you think about your customers?

Xu: At Credit Genie, our mission is to drive financial inclusion. A significant portion of our members fall into the underserved category. They’re hardworking individuals facing a rising cost of living who are often overlooked by traditional financial systems. The traditional risk decisioning process often fails these customers because they might be considered thin file, new to the financial system, or have non-traditional income streams. 

To solve this, we moved beyond the traditional credit report and pioneered our own cash flow-based risk decisioning process. By looking at real-time financial behavior, we can see the true risk profile as well as the true potential of the members we serve and make smarter decisions from there.

Taylor: Rohan, what does a typical cash advance journey look like?

Sriram: A first-time user might download a cash advance app, start onboarding, input their personal information, securely link their bank account via Plaid, and request a cash advance. The application then needs to make three decisions. First, do they onboard this user? Second, do they approve the first cash advance request? And third, if yes, for how much? And that’s just for the first cash advance. 

Often, users take the first cash advance, pay it back, and come back to request a new one. So the decision to approve and size the advance needs to be made again and again every time they come back.

Taylor: Shan, why is that so challenging for a provider?

Xu: Cash flow data was a huge step forward, but looking at just one bank account only gives you part of the picture. We wanted to understand the customer’s full financial story: how they interact with the entire ecosystem, not just us. That’s why we partnered with Plaid. The Cash Advance Index gives us the big-picture perspective, helping us spot healthy patterns that are hard to capture with traditional models.

Taylor: Rohan, what was Plaid’s starting point for solving this problem?

Sriram: We’ve heard the same thing Shan described from many cash advance customers. One in two U.S. adults with a bank account have connected to Plaid, so a new user to a cash advance customer is often not a new user to Plaid. That means Plaid can understand a user’s behavior across the financial ecosystem, which is really valuable in risk assessment. 

For example, a user connecting to multiple apps—maybe cash advance, buy now, pay later, or consumer finance apps—in a very short window can be a meaningful risk signal.

We also benefit from having strong relationships with much of the cash advance provider ecosystem. That gives us the opportunity to learn directly from the best, like Credit Genie, to inform our product strategy and design.

Taylor: Shan, what did Credit Genie need from this partnership?

Xu: We were looking for two things: market-wide perspective and performance uplift.

First, we needed insights that extended beyond our own platform—data that provided a truly differentiated view of our customers’ activity across the market. Plaid was the natural partner to provide that scale.

Second, we needed a solution specifically engineered for a cash advance product. Most traditional tools are built for long-term loans, but our products are high velocity and short duration. We needed data that matched that velocity, so we could optimize risk decisions and drive sustainable growth with rapid iterations.

Taylor: What made the partnership successful?

Xu: The high level of collaborative chemistry between the two teams. It wasn’t just a data handoff. The two teams worked as a single unit from development through implementation.

That helped us drive alignment with our core mission and allowed us to move from concept to real-world impact quickly.

Taylor: Rohan, how does the Cash Advance Index work in the decisioning flow?

Sriram: The first decision a cash advance app makes is at onboarding and approving the first cash advance. The app calls the Cash Advance Index API and specifies the onboarding model.

In the background, we draw user insights from our network—like concurrent advances across multiple providers, repayment patterns across the ecosystem, or signs of financial stress in account activity—and bring that together using ML and AI to generate scores representing the likelihood of the advance being repaid in 30 days.

The app receives a rank-ordered score from one to 99 for various advance amount ranges. The higher the score, the higher the likelihood the cash advance will be repaid in 30 days. Armed with that information, the app makes an approval decision, determines the right advance size, and sets a repayment date. For repeat users, the app can call the API again and specify the ongoing model. Repeat users often have a different risk profile than first-time users, and the ongoing score is designed to capture that nuance.

Taylor: Shan, how did Credit Genie get started, and where is the Cash Advance Index used today?

Xu: We followed a disciplined, multi-phase rollout process. We began with a retro study to validate the score’s predictive power against our historical performance data. Once we confirmed the signal strength, we moved to controlled production experimentation to quantify the performance uplift. 

From a technical standpoint, integration was efficient because we already had a deep foundation with Plaid. Enabling the Cash Advance Index in production was a low-friction, low-effort process. Today, the Cash Advance Index is fully integrated into our real-time decisioning engine, informing critical risk decisions such as approval logic and amount assignment.

Taylor: What results are you seeing in production?

Xu: The Cash Advance Index has truly upleveled our risk assessment capabilities. By synthesizing our performance data with Plaid’s broader network signals, we’ve achieved a new level of precision in predicting repayments.

We’re now making much more informed decisions on eligibility and the sizing of advances. This allows us to better support our members’ liquidity needs while maintaining the risk discipline that enables further growth. During testing, we observed about an 8% relative reduction in missed payment rate while maintaining our target approval rate. From a risk perspective, that’s a pretty significant improvement from our pre-existing risk decisioning process.

Sriram: Separately, Plaid looked at our model performance against similar models in the market. In a live production test, the Cash Advance Index was 17% more predictive in distinguishing between low- and high-risk users than other models used by the provider. That level of precision is meaningful, whether it’s hitting growth targets or keeping missed payments in check.

Taylor: What advice would you give other risk leaders considering a solution like this?

Xu: The real value of network signals is that they give you the big picture of a customer’s activity across financial institutions. That’s something no single cash advance provider can see on their own.

It opens up a fundamentally different dimension of information—an independent layer of risk insight on top of our existing models.