01
ECOA / Regulation B
Adverse action & protected-class testing
TEST
FairLend AI
The AI-powered fair lending intelligence platform that tests 100% of your lending decisions not a 300-file sample and backs every finding with an explanation you can hand to an examiner, a court, or your own board.
An Aquarient AI & Intelligent Solutions Accelerator for banks, mortgage lenders, credit unions, and consumer finance institutions.
[ECOA] [HMDA] [CRA] [SR 11-7] [Fair Housing Act]
What Is FairLend AI
FairLend AI is an AI-powered fair lending intelligence platform that ingests HMDA, German Credit, or Fannie Mae lending data, runs Four-Fifths Rule and Equalized Odds statistical testing across every application on file, and pairs every finding with a SHAP-based explanation feeding directly into case management, ECOA adverse action letters, and HMDA/CRA compliance reporting.
The 2026 Fair Lending Reality
Federal fair lending examination activity has scaled back, several prudential regulators have deferred routine fair lending exams, and disparate-impact enforcement is no longer a federal supervisory priority the way it was a few years ago. That doesn't mean the exposure went away. It moved.
- ECOA and the Fair Housing Act still give individual applicants a private right to sue, and Fair Housing Act disparate-impact theory remains valid law regardless of federal supervisory posture.
- State attorneys general and state regulators are increasingly active in the space federal examiners have stepped back from.
- HMDA and ECOA recordkeeping obligations haven’t changed for institutions above the reporting thresholds, only who’s routinely checking has.
- Plaintiffs’ attorneys have access to the same public HMDA data, and the same statistical tests examiners used to run.
A scheduled federal exam was always a forcing function, not the actual source of the underlying legal exposure. Institutions that keep testing continuously, at the population level, with a documented and explainable process are positioned to demonstrate good faith if that exposure ever surfaces through a private suit, a state inquiry, or a reinstated exam cycle.
How FairLend AI Works
STEP 01
Ingest & Standardize
Upload HMDA LAR, German Credit, or Fannie Mae data. FairLend AI maps, cleans, scores, and indexes the dataset.
STEP 02
Test at Population Scale
Run Four-Fifths Rule, Equalized Odds, and intersectional testing across every uploaded record.
STEP 03
Explain & Escalate
Flagged results automatically create tracked cases with SHAP-based explanations and adverse-action support.
STEP 04
Validate & Report
Run HMDA and CRA checks and package findings into compliance-ready reports.
STEP 05
Monitor Continuously
Compare new uploads against historical baselines to identify drift and emerging bias spikes.
Regulatory Frameworks, Mapped to Action
From fair lending requirements to testable controls, explainable findings, and exam-ready evidence.
REGULATORY COVERAGE
02
HMDA
LAR validation & submission readiness
VALIDATE
03
CRA
LMI mapping & coverage analysis
ANALYZE
04
SR 11-7
Model validation & explainability
EXPLAIN
05
Fair Housing Act
Disparate impact & treatment analysis
EXAMINE
01 DATA
02 TEST
03 EXPLAIN
04 EVIDENCE
05 EXAM-READY
Why It Matters
From population-wide testing to continuous monitoring, FairLend AI turns fair-lending compliance into an explainable, audit-ready process.
100% Population Testing
↗Tests 100% of applications on file, not a manually pulled sample of a few hundred loan files.
Explainable Findings
↗Every finding — a ratio, denial, or anomaly — comes with a SHAP explanation traceable to applicant-level features.
Automatic Case Creation
↗Automatically creates a case when a result crosses a severity threshold, with a full audit trail and assigned owner.
One-Click Exam Export
↗Packages datasets, test results, and case histories into an exam-ready evidence package for regulators, auditors, or your board.
Continuous Monitoring
↗Monitors every new upload for data drift and bias spikes, rather than waiting until the next examination cycle.
FAIRLEND AI Compliance
Intelligence
Intelligence
TEST COVERAGE 100%
EXPLAINABILITY SHAP TRACE
CASE STATUS AUTOMATED
MONITORING ACTIVE
CONTINUOUS COMPLIANCE MONITORING
Frequently Asked Questions
01
What data formats does FairLend AI accept?
HMDA Loan/Application Register (LAR) exports, German Credit format, and Fannie Mae schemas, in CSV, XLSX, or JSON, with automatic mapping to the platform’s canonical fields.
02
Federal fair lending examinations have scaled back in 2026, is FairLend AI still relevant?
Arguably more so. Federal supervisory exam activity has been reduced, but ECOA and the Fair Housing Act still give individual applicants a private right to sue, Fair Housing Act disparate-impact theory remains valid law, and state regulators are increasingly active. Continuous, population-level, explainable testing is what lets an institution show good faith if that exposure surfaces through any of those channels not only through a scheduled exam.
03
How is FairLend AI different from established fair lending compliance software?
Established platforms are built around statistical dashboards and regulatory reporting. FairLend AI adds a modern ML layer on top of that same statistical foundation SHAP-based explainability tied directly into individual ECOA adverse action letters, a natural-language assistant grounded in your own data, and automatic case creation in one closed-loop workflow rather than separate tools stitched together.
04
Does FairLend AI replace a compliance team?
No. It automates the testing, explainability, and evidence-packaging work that doesn’t require a judgment call, and routes everything that does a flagged disparity, a low-confidence result to your compliance officers as an owned, tracked case.
05
Is FairLend AI a packaged product or a custom build?
It’s an Aquarient accelerator a proven fair lending testing and case management foundation configured to your data formats, your institution’s risk thresholds, and your existing compliance workflow.
