One score. Every signal.
A single configurable score drawing on 700+ identity, device, location, behavioral, and income signals — reduced to one clear approve, decline, or review output. Run it alongside your existing credit or risk models to recalibrate decisions, rather than replacing what you already have.
Last updated:
A fraud and credit risk score built to sit on top of what you already have
Most lenders already have a scoring model — a bureau score, an internal credit model, a fraud rules engine. The problem isn't a lack of a score; it's that the score was built on limited data, and it makes two kinds of mistakes: approving applicants who default early, and declining good applicants who would have paid on time. Smile's fraud and risk scoring product is built to correct both, not replace the model you already trust.
It works by combining 700+ signals — identity verification results, device and location intelligence, behavioral patterns, and income data — into a single score, then applying two-way risk calibration against your existing decisioning: reject applications your current model would have approved but that carry hidden risk, and approve applications it would have declined but that Smile's broader signal set shows are actually low-risk. The same signals also surface early payment default (EPD) risk — patterns like digital-identity age and device-to-address distance that correlate with a borrower missing payments shortly after disbursement — so that risk is visible before money moves, not after.
Coverage varies by product. Lending-activity signals cover the Philippines, India, Mexico, Indonesia, Colombia, and Thailand, while email and mobile footprint checks are requested by country code.