The questions a US recruiting operation should be able to answer

Demographic data collected for equal-opportunity reporting exists to measure decisions, which means it must not be capable of influencing them. The structural question is whether that data is stored separately from the hiring record and invisible to the people making decisions, or whether it sits on the same screen as the candidate's name.

The second question is about disposition. When a candidate is rejected, is the reason recorded from a fixed list that can be counted and reconstructed later, or as free text that means whatever its author meant that afternoon? Free-text rejection notes are the weakest possible position if a decision is ever examined, because they cannot be aggregated and they frequently contain reasoning nobody would choose to write down deliberately.

Firms holding federal contracts sit under a further set of obligations administered by the Office of Federal Contract Compliance Programs, including record-keeping and outreach requirements that do not apply to other employers. Whether you are covered is a contractual question to establish before you need the answer.

Several jurisdictions now regulate tools that automate or substantially assist an employment decision. New York City requires that an automated employment decision tool used for hiring or promotion has been subject to an independent bias audit, and that candidates receive notice. Other states and cities have adopted or proposed rules with different definitions and different triggers.

The question for a recruiting operation is first whether anything you use falls inside the definition, which is narrower than people assume and broader than vendors like to admit. A ranking that a recruiter treats as advisory may or may not be in scope depending on how it is actually used, and how it is actually used is a matter of fact about your desk rather than of intent.

If a tool is in scope, the practical requirements are records and disclosure. You need to know which candidates were scored, on what basis, and to be able to hand an auditor the data set an independent audit needs. Commissioning that audit and issuing candidate notice remain the employer's responsibility, not the software vendor's.

Pay transparency rules are one of the fastest-moving areas of US employment law and they operate at state and city level with genuinely different requirements. Some require a good-faith range in the job posting itself. Some require it only on request or at the offer stage. Some apply to remote roles that could be performed from the state, which catches employers who never intended to post there.

Separately, many states and cities restrict asking a candidate about their salary history, and some restrict relying on it even if volunteered. These two families of rules are frequently confused. Disclosure rules are about what you must publish; salary-history rules are about what you must not ask.

The operational questions are about process rather than knowledge. Where in your posting workflow is the range checked, and against which jurisdictions? Who decides what a good-faith range is? Is the salary-history question removed from your application forms and your intake scripts, and does anyone check that it stays removed when a recruiter builds a new form?

Outbound recruiting touches three distinct regimes. Telephone rules govern automated dialling and calls to numbers on do-not-call registries, with the analysis depending on how the call is placed and what it is for. Text messaging in the United States additionally requires carrier-level brand and campaign registration, which is obtained by the sender rather than provided by a platform. Commercial email carries its own requirements around identifying the sender, providing a working opt-out and honouring it promptly.

The recurring question underneath all three is whether you can evidence, for any given contact, how you obtained their details, what they agreed to, and what happened when they asked you to stop. That is a data question your systems either answer or do not.

Whether a specific recruiting message is treated as commercial or transactional, and which regime applies, is a legal determination that depends on content and context. Get it decided once, for your actual message types, and write the answer down.

Recording consent in the United States is determined state by state. Some states require only one party to the conversation to consent, which for a recruiter's own call is often themselves. Others require every party to consent, and the analysis follows the location of the participants rather than the location of your office. A recruiter in a one-party state calling a candidate in an all-party state is the case that catches firms out.

The same analysis applies to AI notetakers and transcription, which are recording by another name even when no audio file is retained.

The safe operating question is whether your system can restrict recording by jurisdiction rather than relying on a recruiter to remember, whether a spoken disclosure is played before recording begins, and how long recordings and transcripts are retained. Handling all-party states by disabling recording is a defensible design; handling them by assuming consent is not.

Employment eligibility verification in the United States is the employer's obligation and it is procedural: specific forms, specific timing, specific document handling, and specific retention periods after employment ends. For a staffing agency placing workers on its own payroll, the agency is the employer and the obligation is yours. For permanent placements it usually sits with the client, but the boundary is worth confirming rather than assuming.

The related risk is over-collection. Asking for documents earlier than permitted, or specifying which documents a candidate must produce, creates exposure of a different kind. Document capture in a recruiting system should therefore be configurable rather than universal.

Retention runs both ways. Some records must be kept for a defined period; others should not be kept longer than necessary, particularly where state privacy laws give candidates deletion rights. Those two pressures conflict, and resolving them is a policy decision your counsel should make and your systems should then enforce.

Surhires supports several of the mechanics described above: demographic data stored separately from the hiring record and not visible to recruiters, coded disposition lists rather than free-text rejections, opt-in capture and suppression handling with unsubscribe and stop-keyword processing, configurable retention windows, recording restrictions by jurisdiction with a spoken disclosure, document capture with expiry reminders, and a check of a posting's salary range against the disclosure rules of the places it is published to.

Where an automated scoring tool is in scope of a rule like New York City's, the product records which candidates were scored and exports the data an independent bias audit needs.

None of that is a compliance guarantee, and none of it is offered as one. Whether your operation meets its obligations depends on how you configure the software, how your recruiters actually behave, and legal determinations that only your own advisers can make.