Introducing fraud calibration and Trust Index for Plaid Identity Verification

Alain Meier
Head of Identity, Plaid

Alain spent his career building identity and anti-fraud products. Now he’s on a path to lower fraud at scale for the digital finance ecosystem.

Fighting fraud is a priority for every company, but keeping up with evolving tactics can feel like a never-ending challenge. That’s why we’re working hard to arm you with solutions to help you onboard more prospective users and fight fraud effectively across the entire customer journey.

Today, we’ve added three new fraud-fighting capabilities within Plaid Identity Verification: fraud calibration, granular machine learning-powered scores for evaluating every aspect of your verification funnel, and an overall user score that we call the Plaid Trust Index to give you a holistic picture of each of your users.

Simple, powerful fraud labeling and fraud calibration

Our redesigned fraud labeling system allows you to easily mark sessions as fraud or not fraud in one click. This not only helps you keep track of your risk decisions but also helps fine-tune our fraud algorithms specifically for your business.

The benefits include:

• Higher pass rates: Delivers protection tailored to your business’s unique needs

• Adaptive learning: Continuously refines algorithms to counter evolving fraud trends

• Fewer false positives: Increases precision with customized, granular risk thresholds

• Better fraud detection: Enhances detection of fraudulent activity

Improve risk model precision with machine learning-powered Plaid Trust Index

We’re also making it easier to refine your risk models by helping you understand why a user was flagged through our new machine learning-driven scores. Rather than sifting through a long list of match flags and reason codes, you quickly determine where a user’s hot spots are. Not only does this detailed visibility empower you to make the right decision based on your specific risk requirements, but it also helps you reduce the time it takes to review sessions and improve your own internal machine-learning models.

The latest update introduces 11 new machine learning-driven scores to give you a deeper understanding of the risk factors for each user. Additionally, our new Trust Index consolidates all verification data into a single number, offering a comprehensive view of the user at a glance. We’ve also added a quick snapshot at the top of every IDV session, so you don’t have to scroll through lengthy reports. A new sidebar feature displays the user's overall Trust Index to help you prioritize reviews more efficiently. Finally, new risk percentiles help contextualize a user’s risk against the broader verification population.

Since releasing the new labeling system last month, we’ve already seen a 300% month-over-month increase in customer-submitted labels, helping to secure the entire ecosystem. Both the fraud calibration and machine learning-powered Trust Index features are available today in the dashboard for all customers. We’ll also be releasing additional features in the coming months, such as dynamic thresholding based on your fraud calibration. Check the dashboard for the latest updates.

Interested in learning more about Identity Verification and Plaid’s fraud solutions?