A Guth Labs publication ↗
Back to News

Propel details cash-flow AI underwriting model and Plaid partnership

Six blank cream tiles connected by gold paths to a green glass lens
AI-generated editorial illustration

Propel Holdings disclosed an enhanced AI cash-flow underwriting model and a Plaid partnership on November 18, describing technology that began operating in late third-quarter 2026.

The in-house model examines six months of bank-transaction history alongside information beyond conventional credit-bureau records. Propel says it uses time-series deep learning to evaluate cash flow and repayment patterns.

Plaid supplies consumer-permissioned financial information, including income patterns, cash-flow insights and obligations. The partnership supports Propel's underwriting process rather than handing lending decisions to a general-purpose chatbot.

Propel says initial results allow additional approvals at unchanged risk thresholds. It estimates access to credit could expand by up to 15%, but that is a company projection, not an independently established improvement. A United States expansion is expected in early 2027.

The announcement does not disclose a detailed validation dataset, subgroup fairness results or a comparison of realized defaults. Those omissions limit what can be concluded about performance from the release alone.

The development matters because transaction histories can provide a different view from a static credit score. Whether that translates into broader, reliable access depends on validation and lending outcomes, not just faster analysis.

Sources and citations

Revision history

Published .