NTRVSTA / Compliance / Bullhorn AI compliance
Compliance · Bullhorn
How to run NTRVSTA AI interviews alongside Bullhorn for staffing agencies under NYC Local Law 144, Illinois, Colorado, California and the EU AI Act: the notices, the records, the data flow and the ceilings.
Last updated November 21, 2026.
Using NTRVSTA with Bullhorn does not move the legal duties: you still give candidates notice before the AI interview, publish a bias audit summary for NYC, explain adverse outcomes in Colorado, and keep records for California and the EU. What changes is the evidence: every candidate gets the same structured interview, and the score and report sit with the candidate in Bullhorn, next to the human decision.
The integration connects Bullhorn candidates and pipelines to an NTRVSTA AI first-round interview, so a recruiter can screen the whole applicant list on a job order. Agencies can send clients a structured report with scores, highlights and short video clips. Bullhorn remains the record for the placement, and the recruiter still decides who gets submitted.
The NTRVSTA team connects Bullhorn for your agency, usually within a few days; your Bullhorn admin approves access. Most agencies start with one desk or one client account. Candidates take the interview from a link at any hour and never need a Bullhorn account. Details: NTRVSTA and Bullhorn.
| Law | What you must do | What NTRVSTA documents | Statutory ceiling |
|---|---|---|---|
| NYC Local Law 144 | Bias audit within one year before use and a public summary; notice to NYC candidates at least 10 business days before use, with how to request an alternative process or accommodation. | Independent bias audit of NTRVSTA AI interviews by Asenion, finalized October 20, 2026. You still publish your own summary and send the notices. | $500 first violation; $500 to $1,500 each later one; each day and each missed notice counts. |
| Illinois (AI Video Interview Act and HB 3773) | Tell applicants AI may analyze the video interview, explain how it works, get consent first, and delete videos within 30 days of a request. Tell employees and applicants when AI is used in employment decisions. | Candidates can withdraw consent for biometric processing by contacting NTRVSTA (Terms 5.1). Interview recordings are kept as the customer specifies, or 2 years (privacy policy 11.1). | Human Rights Act: up to $16,000 per violation and person, more with prior violations, plus damages. |
| Colorado SB 26-189 | Notice before use; within 30 days of an adverse outcome, a plain-language explanation of the decision and the technology's role; correction and human review on request; three years of records. Applies from January 1, 2027. | Carolina scores every candidate against your scorecard and writes down why, which is the raw material for the 30-day explanation. | Up to $20,000 per consumer or transaction ($50,000 if 60+), Attorney General only. |
| California (CCPA ADMT rules and Civil Rights Council regulations) | Pre-use notice, a right to opt out or appeal to a human reviewer, and access to information about the ADMT; anti-bias testing is relevant evidence; keep automated-decision system data for four years. | Each action can be set to suggest, ask first or run on its own, and offers and compensation changes always wait for a person. | Not calculated here. |
| EU AI Act | Article 50 transparency since August 2, 2026; for recruitment, high-risk duties from December 2, 2027: human oversight by trained staff, logs kept at least six months, informing workers and candidates. | Where required, NTRVSTA informs people they are interacting with an AI system (Terms 4.2). Every step is logged so you can replay it. | Up to EUR 15 million or 3% for deployer duties; EUR 35 million or 7% for banned practices. |
Ceilings are maximum statutory exposure, not legal advice. Work out yours in the penalty calculator.
NYC Local Law 144 applies to an "employer or an employment agency" that uses an automated employment decision tool, so an agency screening NYC candidates owes the bias audit summary and the notices itself, not only its client. Colorado and California rules also reach the deployer of the tool.
For systems you connect yourself, NTRVSTA's External Integrations API (beta) sends one signed webhook event, result.completed, when a final interview result is ready. It is signed with HMAC SHA-256 and carries the candidate's identity, the interview template, completion time, a result URL, a candidate thumbnail and the AI scoring summary with highlights. It deliberately excludes transcripts and other internal analysis. Results can also be pulled from GET /v1/interview-results for reconciliation. See the webhook docs and API reference. The Bullhorn connection itself is set up by the NTRVSTA team.
Carolina is NTRVSTA's AI recruiting agent. Every pass comes with a reason: she scores every candidate against your scorecard and writes down why, so "the AI decided" is never the answer. You set each action to suggest, ask first or run on its own, and offers and compensation changes always wait for a person.
Yes. The integration is live: NTRVSTA runs the AI first-round interview and results sit with the candidate in Bullhorn. The NTRVSTA team sets it up, usually within a few days, with a Bullhorn admin approving access.
No. Candidates take the interview from a link at any hour, by video or voice. Bullhorn access is only for your team.
If it scores or ranks candidates for hiring or promotion in New York City, it is likely an automated employment decision tool: you need a bias audit within a year before use, a public summary, and notice 10 business days before use.
No. The result.completed webhook carries the scoring summary, highlights, a result URL and a candidate thumbnail, and intentionally excludes transcripts and other internal analysis.
We use cookies to improve site performance, analyze traffic, and enhance your experience. Learn more in our Privacy Policy.