Hiring-AI compliance · built with employment counsel

Fair hiring,
on the record.

IQualify catches biased language in your job posts before they publish, turns every candidate decision into a defensible, evidence-cited record, and keeps a tamper-evident audit trail — so you can prove your hiring is fair the moment anyone asks.

Built for the rules that now govern hiring AI
NYC Local Law 144 EEOC guidance Colorado SB 205 EU AI Act Title VII · ADEA · ADA
greenhouse · Security Engineer · bias review
— 3+ years in detection engineering
— A digital native mindset; you live in Slack
— Comfortable writing postmortems
Age · ADEAAGE_DIGITAL_NATIVE · 92%

“Digital native” is an age proxy

One of the most-litigated age proxies in modern hiring — the EEOC has specifically called out the phrase as discriminatory.

Suggested fix: state the actual proficiency — “fluent with collaboration tools (Slack, Notion, Google Workspace).”
25
counsel-approved bias rules
9
protected-class categories
100%
decisions logged & replayable
0
candidate PII in the analysis store
How it works

Intelligence at every step — never a black box

IQualify drops into your existing Greenhouse workflow. Two webhooks, one shared review queue, and a record of everything. Always advisory, never auto-rejecting.

1

Scrub the job post

The moment a req publishes, IQualify reads the JD and flags biased language at the source — each flag citing the governing law, the risky phrase, and a compliant rewrite for a human to approve.

2

Review candidates defensibly

Every applicant gets a PII-stripped gap analysis against the approved requirements — qualitative tiers with cited evidence, and separate recruiter and candidate language. No scores to game.

3

Keep the receipts

Every machine suggestion and human decision lands in a tamper-evident, hash-chained log — the spine of your annual NYC Local Law 144 bias audit, replayable years later.

The platform

A compliance-grade engine, not a wordlist

The hard, defensible parts are built in from day one — the parts a legal review will actually interrogate.

⚖️

Counsel-reviewed catalog

25 bias rules across 9 protected classes, each with a legal citation, a documented false-positive guard, and a safe rewrite. Reviewed and approved by employment counsel.

🔒

PII-free by construction

The analysis store has no name, email, or phone fields — candidates are referenced only by a one-way hash. Privacy is enforced by the schema, not a policy promise.

⛓️

Tamper-evident audit

An append-only, hash-chained event log, designed to anchor daily roots to an external timestamp service. Even dismissed flags are recorded — that's what makes it defensible.

🔁

Replayable decisions

Every analysis records its model, prompt version, and seed. If a rejection is disputed two years later, replay the exact decision bit-for-bit.

🔗

Native to your ATS

Webhook-driven, no rip-and-replace. It shows up inside the Greenhouse flow your team already uses — one shared review queue.

📄

Audit report, automatic

An annual bias-audit report generated per employer — methodology transparent enough for an outside auditor to replicate, with an honest limitations section.

Coverage

It catches bias everywhere it hides

Not just the obvious phrases. Age proxies, coded gender language, credential and socioeconomic gatekeeping, citizenship and culture-fit traps — across every protected class.

Why now

Hiring AI just went from optional to audited.

NYC requires annual bias audits of automated hiring tools. The EEOC is actively enforcing. Colorado's SB 205 and the EU AI Act classify hiring AI as high-risk. Employers are being told to prove their hiring is defensible — and most have no way to do it. IQualify is that proof, built on a foundation that's hard to copy:

A frozen, append-only rule catalog that lets any historical flag be replayed against the exact rule that fired.
An architecture where candidate data is never linkable once aggregated — the privacy answer regulators want.
Detection plus defensible record plus native ATS workflow — the combination, not any one piece.
One shared engine behind the ATS integration, a candidate browser extension, and an anonymized data layer.

See it work in 60 seconds.

Click through a real requisition queue, watch the engine catch bias across six roles, review a candidate pipeline, and walk a defensible rejection from draft to sent — all on synthetic data.

Open the interactive demo
Investor or design-partner conversations: threattape@gmail.com