Vero AI for HR & Hiring

Fair-Hiring Evidence For Every Market You Hire In — One Platform.

Vero AI runs statistically defensible bias audits and validity monitoring on your hiring tests and selection procedures, then maps the same evidence to UGESP, NYC Local Law 144, Illinois, California, Colorado, the EU AI Act, UK, Canadian hiring rules, and others at once — so a growing, confusing patchwork of global hiring law becomes one standing control, not a dozen separate projects.

Selection funnel · adverse-impact ratio
live
Resume screen12,480 applicants
92%
Assessment5,742 tested
71% flag
Interview1,780 interviewed
88%
Offer427 offers
96%
Adverse impact flagged at Assessment — 71% ratio, below the four-fifths threshold. Mapped to UGESP, NYC LL144, Illinois, Colorado, EU AI Act.
The unaddressed gap

Hiring tests can quietly create legal exposure — in every market you operate

UGESP and a fast-growing list of city, state, national, regional, and global laws require employers to test for adverse impact across protected classes and to validate any selection procedure that shows it — but most teams never run the numbers, almost none monitor them continuously, and each new jurisdiction adds its own one-off project.

Before Vero AI — manual, one-off, jurisdiction by jurisdiction
✕Adverse-impact analysis run once, if ever, and rarely revisited as hiring data changes.
✕NYC bias audits handled as a standalone annual project, disconnected from other compliance work.
✕Every new law — Illinois, California, Colorado, the EU AI Act — triggers its own spreadsheet and its own consultant.
✕Public disclosure summaries drafted by hand for each jurisdiction's specific format and deadline.
With Vero AI — structured, continuous, jurisdiction-agnostic
✓Four-fifths ratio, significance testing, and effect size calculated automatically, every hiring cycle.
✓One evidence layer maps simultaneously to UGESP, NYC LL144, Illinois, California, Colorado, the EU AI Act, UK, and Canadian rules.
✓Adding a new jurisdiction means adding a control set, not starting a new project.
✓Audit-ready workpapers and public disclosure pages generated straight from the same underlying results.
Three ways to engage

Start with an audit, or go straight to continuous monitoring

A one-time statistical answer, an always-on standing control, or a public-facing trust signal — all built on the same underlying evidence engine.

Tier 1
Bias Audit
One-time statistical analysis — fixed-fee engagement
  • •Four-fifths (80%) adverse-impact ratio calculated for every stage — screen, assessment, interview, offer
  • •Statistical significance testing (Fisher's exact / z-test) on selection-rate differences
  • •Standardized mean difference (Cohen's d) on raw test scores by protected class
  • •Run across every job family, level, and location in one pass
  • •Audit-ready report with sourced findings and methodology, tied back to raw data
Best for: A defensible point-in-time answer to "do our tests show adverse impact?"
Scope a bias audit
Recommended
Tier 2
Continuous Monitoring
Standing control — re-evaluated every hiring cycle
  • •Everything in Tier 1, re-run automatically each hiring cycle as new data lands
  • •Standing thresholds — deviations generate findings automatically
  • •Trend tracking catches ratios drifting toward violation before they cross it
  • •Validity-study lifecycle tracking for tests carrying adverse impact
  • •Multi-framework overlay: UGESP + state/city law + EU AI Act + internal policy
  • •Living dashboard for HR / Legal / Board — no more one-off studies
Best for: Making fair-hiring testing a standing control, not a project.
Start continuous monitoring
Tier 3
Public Trust Center
Public-facing proof, pulled from monitored results
  • •Public page showing verified fair-hiring metrics for your specific hiring process
  • •Real-time or periodically refreshed status pulled directly from monitored Vero AI results
  • •Builds candidate and regulator trust without exposing underlying raw data
  • •Positions you ahead of emerging automated-hiring disclosure expectations (EEOC, state AI-hiring laws)
  • •Turns compliance into a recruiting and brand differentiator
Best for: Employers who want fair-hiring credibility to be visible, not just internal.
Explore a trust center
Built on the Vero AI universal compliance platform
Four-Fifths Rule
EEOC/UGESP adverse-impact ratio, per stage & group
Significance Testing
Fisher's exact / z-test on selection-rate gaps
Effect Size (Cohen's d)
Practical significance on raw test scores
Validity Tracking
Content / criterion / construct evidence lifecycle
Legal basis: EEOC Uniform Guidelines on Employee Selection Procedures, 29 CFR §1607.4(D).
One platform, every jurisdiction

The global hiring-law patchwork, mapped to one evidence layer

Fair-hiring and algorithmic-discrimination law is no longer one federal standard — it’s a fast-moving mix of city ordinances, state statutes, and international regulations, each with its own triggers, audit cadence, and disclosure format. Vero AI maps your evidence to all of them at once instead of treating every new law as a new project.

UGESP / EEOC (Federal)
29 CFR §1607 four-fifths rule and validation requirements — the statistical baseline for every audit on this page.
NYC Local Law 144
Independent annual bias audit, public results summary, and 10-business-day candidate notice for automated employment decision tools.
Illinois HB 3773
Illinois Human Rights Act amendment (effective Jan. 1, 2026) — prohibits AI-driven discriminatory effects and zip-code proxies, requires employee notice.
California FEHA / ADS Regulations
Civil Rights Council rules (2 CCR §§11008.1–.4, effective Oct. 1, 2025) on automated-decision-system testing, recordkeeping, and proxy discrimination.
Colorado AI / ADMT Law
SB 26-189's disclosure, human-review, and data-correction requirements for covered automated decision-making technology in employment, effective Jan. 1, 2027.
Connecticut & Other State AI Laws
Emerging state statutes (e.g., Connecticut SB5) regulating automated employment decision tools, tracked and added as they take effect.
EU AI Act — Annex III Employment
High-risk classification for recruitment, candidate screening, and promotion AI, with risk assessments, bias testing, and human-oversight obligations.
UK Equality Act & ICO Guidance
Equality Act 2010 protected-characteristic discrimination rules plus ICO/DUAA safeguards for automated decision-making in recruitment.
Canada — Ontario, AIDA & Quebec Law 25
Ontario's Working for Workers Four Act AI job-posting disclosure, federal AIDA developments, and Quebec Law 25 automated-decision rights.
Custom Control Sets
Any internal policy, union agreement, or newly enacted jurisdiction — encoded once, evaluated consistently everywhere you hire.
Not sure which of these apply to you? Ask Vero AI which hiring laws are triggered by the markets you hire in — no form, no sales call.
Ask Vero AI
One evidence layer, any footprint

Built for employers hiring in one city — or fifty countries

Every employer needs the same statistical foundation. What changes is how many rulebooks that foundation has to satisfy at once — and that’s exactly what Vero AI is built to scale.

Single-Market Employers

Companies hiring primarily under one federal, state, or city framework — where UGESP and a handful of local rules define the standard.

  • ✓UGESP four-fifths testing built into every audit by default
  • ✓NYC LL144 annual bias audit, public posting, and candidate notice handled end to end
  • ✓State-specific rules — Illinois, California, Colorado — layered on automatically as they apply
  • ✓Validity-study tracking for any test that shows adverse impact
Multi-Market & Global Employers

Companies hiring across US states and international borders, where every market layers on a different definition of “high-risk” and a different disclosure deadline.

  • ✓One evidence layer mapped to the EU AI Act, UK Equality Act/ICO guidance, and Canadian disclosure rules simultaneously
  • ✓Every US state and city requirement layered onto the same underlying test results, not run separately
  • ✓New jurisdictions added as a control set in days, not a new consulting engagement
  • ✓Single living dashboard for HR, Legal, and the Board across every market you hire in

What you can achieve with Vero AI

  • Unlimited — Hiring laws mapped to one evidence layer
  • 4+ — Statistical tests run automatically per cycle
  • <1 min — To first evaluated finding
  • 1 — Platform for every market you hire in
Fair-hiring FAQs

Adverse impact, bias audits, and the laws behind them

What is adverse impact analysis?
Adverse impact analysis is the statistical test of whether a selection procedure — a resume screen, an assessment, an interview rubric, or an AI scoring model — passes candidates from one protected group at a materially lower rate than another. Under the EEOC Uniform Guidelines on Employee Selection Procedures (29 CFR §1607), the baseline test is the four-fifths rule, supported by significance testing and effect size. Vero AI runs all three on every hiring stage.
What is the four-fifths (80%) rule?
The four-fifths rule compares the selection rate of each protected group to the selection rate of the highest-selected group. If the ratio falls below 80%, the guidelines treat that as evidence of adverse impact, which then requires the employer to validate the selection procedure. Because small samples can trip the ratio by chance, Vero AI pairs the ratio with Fisher's exact / z-test significance testing and Cohen's d effect size so findings hold up under scrutiny.
How do you audit an AI hiring tool for bias?
You collect selection-rate and scoring data by stage and protected class, calculate the four-fifths ratio for each stage and group, test the selection-rate gaps for statistical significance, measure practical significance on raw scores, and then document the methodology and results as workpapers tied back to the underlying data. Where impact is found, you track validity evidence for the tool. Vero AI automates that sequence and re-runs it every hiring cycle.
What does NYC Local Law 144 require?
NYC Local Law 144 requires employers using automated employment decision tools to obtain an independent annual bias audit, publish a summary of the audit results, and give candidates at least 10 business days' notice before the tool is used. Vero AI produces the underlying statistics, the audit-ready workpapers, and the public disclosure summary from the same evaluated evidence.
Do Illinois, California, and Colorado have their own AI hiring rules?
Yes. Illinois HB 3773 amends the Illinois Human Rights Act to prohibit AI-driven discriminatory effects and zip-code proxies and requires employee notice. California's Civil Rights Council ADS regulations (2 CCR §§11008.1–.4) add automated-decision-system testing and recordkeeping duties. Colorado's SB 26-189 adds disclosure, human-review, and data-correction requirements for covered automated decision-making technology in employment. Confirm current effective dates before relying on them — this area is moving quickly.
Does the EU AI Act apply to recruiting?
Yes. Annex III of the EU AI Act classifies AI used for recruitment, candidate screening, and promotion decisions as high-risk, which brings risk assessment, bias testing, technical documentation, and human-oversight obligations. Employers hiring into the EU generally need the same statistical evidence base they already produce for UGESP, mapped to a different disclosure format.
Can one bias audit satisfy several jurisdictions at once?
The underlying statistics can be shared; the disclosures cannot. Vero AI evaluates the hiring evidence once and then maps those results to each jurisdiction's control set, so UGESP, NYC LL144, state statutes, and international rules draw on one evaluated evidence layer instead of separate consulting projects.
How often should adverse impact testing be re-run?
Point-in-time audits go stale as soon as hiring data changes. NYC requires an annual bias audit at minimum, but ratios drift between audits. Vero AI's continuous monitoring re-runs the full test battery each hiring cycle and raises a finding as soon as a ratio trends toward violation, rather than after it crosses.
Vero AI for your industry

One rigorous engine for every compliance domain your clients face — cyber, quality, ESG, financial, custom.

Professional Services, Audit & Accounting

One engine across every client engagement.

  • Cyber, quality, ESG, financial, and custom control sets in the same workflow
  • Client evidence evaluated once, credited to every framework it satisfies
  • Reviewable rationale on every conclusion, so partners can sign off
Explore Professional Services, Audit & Accounting
9 regimes, one evidence set

Ready to See Where Your Hiring Data Stands?

See how Vero AI evaluates your selection procedures against the fair-hiring laws you’re already held to — in every market you hire in.