Plaid Launches AI Models for Credit and Fraud Detection Using Cash Flow Data

Plaid Introduces AI Models to Transform Credit and Fraud Assessment

Plaid unveiled a suite of AI models designed to address critical gaps in credit risk prediction, fraud detection, and payment risk assessment. The new tools leverage cash flow data from millions of transactions across the Plaid Network, offering lenders and financial institutions deeper insights into consumer financial behavior patterns1

1

2

The launch comes as millions of US adults fall outside traditional credit scoring systems, making it difficult for lenders to evaluate their loan repayment ability using conventional methods alone. Plaid's approach taps into real-time cash flow data that reveals how people earn, spend, and pay their bills, providing a fuller picture of creditworthiness1

1

LendScore 2 Delivers 42% Stronger Credit Risk Prediction

At the core of Plaid's credit offering is LendScore 2, an AI-powered financial model that analyzes cash flow data to predict a borrower's ability to repay loans. The model demonstrates 42% stronger predictive power compared to traditional credit data alone, according to Plaid1

1

2

Plaid has also developed specialized models tailored for specific lending sectors, including auto loans, home loans, and short-term loans. These industry-specific tools extend the predictive gains of the core LendScore 2 model, enabling lenders to make more informed decisions across different loan types2

2

Michelle Young, Plaid Credit Product Lead, explained that while cash flow data provides a more complete picture of borrowers, lenders haven't always been able to access and act on these insights at scale. "The next generation of LendScore and specialized models close that gap at scale, and with Arc, we're giving lenders new tools to expand access to more affordable credit," Young stated2

2

Transformer-Based Model LendScore Arc Shows Significant Lift

Plaid also introduced LendScore Arc, a transformer-based model that represents the company's most advanced credit risk prediction tool. Unlike traditional models, Arc learns from the order and timing of a borrower's transactions, analyzing financial behavior as it unfolds over time1

1

In early testing, LendScore Arc delivered significant predictive lift over the core model, including 20% improvement on deep subprime borrowers and 24% lift on superprime borrowers. This transformer-based approach positions Arc as Plaid's best-performing credit model for lenders ready to adopt advanced AI-powered financial models1

1

2

Instant Link Speeds Up Credit Applications

To streamline the lending process, Plaid launched Instant Link, which enables borrowers to consent to share their cash flow insights for future credit applications by connecting their financial accounts to Plaid Consumer Reporting Agency. Lenders gain access to these insights in under two seconds, creating a faster experience for eligible consumers1

1

2

Related Stories

Plaid and Algebrik AI Partner to Revolutionize Loan Origination for Credit Unions

Financial Institutions Consolidate Hundreds of AI Models Into Single Transaction Systems

AI-Driven Synthetic Fraud: A Growing Threat to Financial Institutions

AI Revolutionizes Credit Risk Models in Fintech and Banking

AI Foundation Model Enhances Fraud Detection

Beyond credit, Plaid rolled out an AI foundation model purpose-built for fraud detection. Trained on patterns and hundreds of millions of data points from the Plaid Network, the model analyzes the full sequence of events rather than isolated snapshots. In internal evaluations, it delivered up to 40% relative improvement over previous baselines, enhancing Plaid's fraud detection solution, Protect1

1

2

Signal Improves ACH Payment Risk Assessment

Plaid's sequential foundation model now powers Signal, an ACH payment risk model that reads an account's transaction history in sequence to better predict payment risk. In testing, the model helped Signal prevent 26% more ACH returns without increasing false flags1

1

This development is particularly relevant as 94% of surveyed firms are not yet fully compliance-ready for new Nacha rules requiring every bank and business using the ACH network to actively monitor for fraud, according to PYMNTS Intelligence and Plaid collaboration research2

2

Will Robinson, CTO at Plaid, emphasized that the company's credit and fraud models bring deep financial context to every problem they solve. "They build on foundation models that already understand how financial behavior unfolds over time, across the Plaid Network, and that means better decisions, and better outcomes, for our customers and the millions of people who depend on those services to manage their own financial lives," Robinson said1

1

2