June 12, 2025
Introducing Plaid Protect – the next chapter in fraud prevention
In the world of financial fraud, there's what you see within your own systems, and then there's everything else. That everything else spans across banks, fintech apps and other financial services, where sophisticated fraudsters exploit blind spots that no single company can catch on its own. This is why today’s biggest fraud challenges – like first-party fraud, synthetic identities, and account takeovers – remain so difficult to catch.
Today we’re launching Plaid Protect: a real-time fraud intelligence system that helps detect and prevent fraud from the moment a user first interacts with your app or service. By drawing on fraud signals across a billion devices in the Plaid network, Protect goes beyond what any single company can see – surfacing fraud patterns that exist between linked bank accounts, connections to financial apps and services and more. Protect brings together two powerful components: ML-powered risk scores and fraud attributes with an intuitive dashboard built for fraud operations.
With this new lens on fraud, companies can reduce fraud losses, dramatically improve conversion, and make smarter decisions from the very first user interaction and every step thereafter.
First-of-its-kind fraud prevention you can’t get anywhere else
Protect is built on a powerful fraud model called Trust Index (Ti). It is an adaptive, ML-powered risk engine that generates real-time risk scores and attributes that evolve as user context changes—starting from the first interaction through to onboarding, account linking, and ongoing activity.
Our first production model, Ti1, is built on top of Plaid’s extensive network with the ability to tap into 10,000 high-signal attributes including cross-app patterns, device history, bank account risk signals and much more. It is highly predictive in detecting synthetic identities, first-party fraud, and account takeovers because we are leveraging fraud signals that are difficult for fraudsters to manipulate or fake.
The Trust Index is built on:
Network intelligence: Behavioral insights and attributes from interactions across a billion devices with thousands of apps, capturing behaviors like high-velocity account linking, frequent login failures, and deleted financial connections.
Bank account risk: With hundreds of millions of accounts that have connected to Plaid, we can see key changes to the connected financial account, whether they were recently opened or closed, or if contact details like phone numbers and emails show suspicious alterations.
Consortium feedback: We surface real-time fraud alerts contributed by other businesses and consumers - for example, account takeover flags from banks or synthetic identity detections from lenders.
Advanced Identity Intelligence: Any and all fraud signals captured during the ID verification stage are included as part of the risk analysis and help inform other data categories.
Drive greater conversion with dynamic risk-based onboarding
Protect is activated when a user begins to onboard your app or service - generating an early fraud signal without collecting more personal information. Because more than 1 in 2 banked U.S. individuals have used Plaid to link their bank accounts with apps and services, we can immediately compare a user’s device against billions of historical sessions. This goes beyond device fingerprinting, giving you more insights with fewer inputs from your users.
As users share more information – phone number, email, SSN via the IDV process – the Trust Index updates in real time, continuously refining the risk profile and helping you apply the right level of friction (for example, step‐up verification) at the right time. This helps boost both security and improves conversion. Every decision feeds back into the model, so our signals grow smarter as the network expands.
With one of our early fintech customers, we tested Protect against a set of suspicious user accounts that had slipped through their existing defenses. We found that just by stepping up 5% of users for additional verification, we could’ve helped catch nearly 40% of first-party fraud.
Investigative toolkit built with AI-powered tools that anyone can use
Protect helps streamline your fraud operations with intuitive tooling, making it easy to search, investigate and take action - all within the Protect dashboard. You can get a birds eye view of every user interaction across devices, identity checks, and bank links scored across your company in real time by the Trust Index to help spot fraud patterns. That means you get a complete, up-to-the-minute view of how fraud might be evolving across your platform.
The Protect dashboard offers semantic search powered by natural language. You can use plain English to ask questions of your data—no SQL or custom queries required. For example:
“Events this week with a TI score under 50”
“Users who opened 5+ new accounts in 30 days”
“Users who linked multiple accounts last week and have a new phone number”
Protect understands what you mean, surfaces matching sessions instantly, and lets you take action. With just a few clicks, you can turn any search into an automated rule - so you can flag similar behavior going forward, stop repeat fraud, and stay ahead of new patterns without writing a line of code.
Get started with Protect
Plaid Protect is now available in beta. Early adopters are already using Protect to fight fraud and improve user conversion. Reach out to your Plaid account team to get access, or sign up for the waitlist here to stay tuned for more updates, including general availability and our next model release.