Plaid’s Fall 2025 Product Release

Most risk and fraud systems operate on lagging signals. Credit decisions are made based on whether someone missed a payment 30 or 60 days ago, not on what’s in their bank account today. Fraud detection often defaulted to blocking new users, because there wasn’t enough signal to separate the legitimate users from fraudsters.

With real-time data from across the Plaid Network, we’re launching a new generation of predictive risk and fraud scores. These scores provide real-time intelligence so businesses can make faster, more confident decisions across all stages of the user journey from onboarding to underwriting and payments, without sacrificing conversion. 

New scores, powered by the Plaid network

Every day, nearly a million connections happen across the Plaid Network, spanning real-time activity from over 150 million consumers and more than 7,000 apps. From budgeting tools to brokerages. From neobanks to auto lenders. This reach gives us an unparalleled view of how consumers move money, spend, and behave across digital finance.

That scale matters because it means our scores aren’t just based on static history. They’re powered by live, predictive signals:

• Trained on billions of transactions

• Informed by one billion linked devices

These signals unlock a more complete picture of financial behavior—from sign-up to funding an account, from applying for a loan to moving money. Plaid-powered scores deliver a smarter foundation for decisioning at every step.

LendScore: credit risk score

Traditional credit scores still matter. But they miss what’s happening in a borrower’s financial life right now. LendScore looks at real-time cash flow, income patterns, and most importantly, account activity across the Plaid Network to understand how a borrower is managing their finances, and spot risk behaviors that could impact their ability to repay. 

Our first model, LS1, built on top of hundreds of attributes, predicts 12-month default risk and delivers a 25% lift in predictive performance compared to traditional credit data alone. 

Trust Index: fraud risk score

Five months ago, we launched Protect powered by our anti-fraud model the Trust Index (Ti). With Ti, we provide a fraud risk score based on account-level and device-level behavior across our network.

Today, we released Ti2, which includes:

• Behavioral signals, based on patterns in transaction history. This gives you stronger signals around whether a potential user is acting like a real person or a fraudster.

• Live, network-wide graph analysis to detect patterns of shared risk across users, devices, accounts, and apps to surface threats that are invisible when looking at one app in isolation.

With 2X the training data, Ti2 is already catching 30% more fraud.

Signal: transaction risk scores

Signal started with a simple idea: use machine learning to reduce ACH risk. But in the past year, it’s grown into a full platform for managing transaction risk at scale, with more predictive scores, new features, and new tools to make enterprise-grade risk management accessible to all.

Here’s what’s new:

• A powerful dashboard and rules engine: pinpoint risk and easily take action with improved explainability 

• Model upgrades and improvements: 140% better performance on our core model

• Risk scores or real-time balance checks: choose lightweight balance checks or full ML-powered scores to match the risk profile of every transaction.

Whether you're a startup managing returns or a large platform scaling bank payments, Signal now gives you sharper control, without needing to code.

Beyond our scores, we also have a number of new updates and features that help improve customer engagement, cut costs, and drive conversion:

• Transfer for platforms: Transfer now supports platforms, helping you onboard customers to bank payments in minutes, and cut processing costs by up to 40% compared to cards.

• Bank intelligence suite: A new suite of tools to help financial institutions better understand customers, deepen engagement, and fight fraud.

• Layer: We’ve supercharged our instant onboarding experience Layer with two new features: an improved UI that’s driving up to 5% conversion lift, and extended Autofill expanding access to Layer for users who are not already in the Plaid Network. 

• Plaid for Mortgage: With industry-leading account linking conversion, our new asset verification suite is now GSE-certified and includes built-in analytics designed for the unique needs of mortgage lenders.

• Payments dashboard (EU): A new payments dashboard for European customers that gives customers clearer insights, real-time tracking, and streamlined reporting—all in one place.

Our goal has always been to help our customers get more value from being part of the Plaid Network so they can better serve their users, reduce risk, and grow their business. The scores we launched today are a direct extension of that. By turning real-time, permissioned data into predictive signals, we’re making it easier for risk teams to operate with more clarity, less friction, and stronger outcomes. And we’re just getting started.