Find adverse impact in your workforce data before someone else does.
Intersect™ runs the same statistical tests courts and agencies rely on — two-proportion Z-tests measured against the EEOC's four-fifths rule — across hiring, promotion, compensation, and separation. You get continuous monitoring, a defensible record of what you reviewed, and filing-ready output.
Workforce Analytics
Pay Equity
Adverse Impact
Intersectional Analysis
Pipeline & Advancement
Compliance Monitoring
Reports
Settings
From annual spreadsheets to continuous oversight.
Most employers discover a disparity when a charge, an audit, or a demand letter arrives. Intersect™ surfaces it while you still have time to investigate, correct, and document the response.
See how it works →Adverse Impact Detection
Impact ratios calculated against the four-fifths rule in the EEOC's Uniform Guidelines (29 CFR 1607), across hiring, promotion, pay, and separation.
Intersectional Analysis
Test combined categories — race and gender together — not just one at a time, with Benjamini-Hochberg correction so repeated testing doesn't manufacture false signals.
Filing & Deadline Tracking
Stay ahead of EEO-1 Component 1, California pay data reporting under SB 464, and the Illinois and Colorado equal pay requirements, on one calendar.
Defensible Documentation
A timestamped record of what was flagged, who reviewed it, what was decided, and what changed — exportable when you need to show your work.
Three steps from raw HRIS export to defensible record.
Import from your HRIS or upload a structured file. Employee-level data stays isolated to your organization.
Impact ratios and two-proportion Z-tests across every selection decision, with intersectional breakdowns.
Record what leadership reviewed and decided, then export for filings, counsel, or the board.
How the numbers are actually calculated.
Analytics that inform a legal position should be reproducible and explainable. Here is exactly what runs under the hood.
- Impact ratio against the four-fifths ruleThe selection rate for each group is compared with the highest-selected group. A ratio below 0.80 is the threshold the EEOC's Uniform Guidelines treat as evidence of adverse impact.
- Two-proportion Z-test for statistical significanceThe four-fifths rule alone is sensitive to small samples. Each comparison is also tested for significance, following the standard-deviation approach used in Hazelwood School District v. United States and Castaneda v. Partida.
- Benjamini-Hochberg correctionTesting many groups at once inflates false positives. This correction controls the false discovery rate so flagged findings hold up to scrutiny.
- External benchmarkingInternal results are compared against EEOC EEO-1 public use file data and Bureau of Labor Statistics occupational data, so you can see availability context, not just internal numbers.
Know what your data shows before you have to explain it.
Fragmented spreadsheets and point-in-time reviews leave gaps that are difficult to defend later. A continuous record closes them.
Compliance Assessment
A review of your current selection procedures, pay practices, and data readiness, with specific findings and recommended next steps.
Learn more →Filing Services
Preparation and submission support for EEO-1 Component 1 and state pay data reports, including California's SB 464 requirements and the SOC code migration.
Learn more →What Intersect™ brings into view
Leaders don't need more workforce data. They need to know what the data means, where risk is emerging, and what action to take next.
Workforce outcomes
Patterns across hiring, advancement, compensation, and retention, with the underlying numbers available.
Emerging risk
Groups approaching or falling below the four-fifths threshold, and how far the gap has moved since last quarter.
Strategic positioning
Findings translated into decisions you can explain to a board, a regulator, or opposing counsel.
Documented action
A dated record of what was identified, evaluated, and done in response.
See what your data shows.
Document your strategic positioning.
A 30-minute walkthrough using sample data, or bring a de-identified export and we'll look at yours.