ECOA and Reg B compliance built into the decision layer.
Disparate impact scanner runs on every decision batch. Reg B top-4 adverse action reason codes are generated from decision factors and rule outcomes — not derived from score rank percentiles. Monthly fair lending report is produced automatically for your compliance file. Prism Layer does not conduct fair lending compliance review — we produce the disparate-impact reporting, reason-code coverage, and 4/5ths rule calculations your compliance team uses to conduct that review.
Request DemoDisparate impact scanner that runs on every decision batch
The scanner calculates approval rates, adverse action rates, and pricing differentials across demographic proxy groups after every decision batch. Flags appear in your compliance dashboard before your next monthly report — not when the examiner arrives.
80% rule and ratio tests
Adverse impact ratio calculated against the control group after each batch. Threshold flags appear in dashboard before the next scheduled report cycle.
HMDA proxy analysis
Where HMDA data is available, the scanner uses reported demographic data. Where it isn't, Bayesian proxy methods generate an approximation with documented confidence intervals.
Factor attribution for disparities
When a disparity is detected, the diagnostic layer traces which decision factors contribute most to the gap — giving your compliance team a starting point for the policy review, not just an alert.
Examiner-ready documentation
Monthly disparity report generated automatically in a format designed for regulator review. No manual data extraction, no spreadsheet assembly before the exam.
Adverse action generation from decision factors, not score ranks
Reg B requires that adverse action notices identify principal reasons for denial. Prism Layer generates these reasons from the actual rule outcomes and ML factor attribution — not from a ranked list of score features. The legal distinction matters when an examiner asks how you determined the principal reason.
Plain-language Reg B reason statements
Reason codes are generated from rule and factor outputs in language that maps to the CFPB's model adverse action notice forms. No translation layer needed.
LOS delivery or API output
Adverse action codes and reason statements are delivered to your LOS via the decision API response. Or exported in batch for standalone notice generation.
See your current fair lending posture.
We can run the scanner against your historical decision tape and show you what a monthly disparity report looks like before you commit.
Request Evaluation