Catch a missing salary range before the advert goes out

A remote role does not have one location. Depending on the rule in play it can be governed by where the employer is, where the work will be performed, or where the candidate reading the advert happens to live. A posting that says remote and nothing else is, in effect, published everywhere it can be seen, and several jurisdictions treat it that way.

The validator asks the question the posting is avoiding: which locations is this role actually open to? Once the requisition carries that answer as structured data rather than as a sentence in the description, the check becomes tractable. A role open across twelve states is checked against twelve rule sets, and the strictest disclosure requirement among them is the one the flag reports.

If you genuinely do not want to answer that question, the validator says so rather than passing the posting silently. An unbounded remote advert is a decision, and the flag makes it a conscious one.

A validator that runs after the fact is a report. This one runs in the publishing step, between the moment the advert is finished and the moment it leaves the system, which is the only point at which the information is still cheap to change.

The output is a list of flags, each attached to a specific destination. A missing range, a range so wide it may not satisfy a good-faith requirement, a range quoted without a currency or a period, a missing statement about other compensation where one is expected, or a salary-history question sitting on the linked application form.

Publishing is not blocked by default. A flag can be acknowledged with a reason, and the acknowledgement is recorded against the posting with the user and the date. Blocking would guarantee that somebody works around the tool inside a fortnight; recording the override keeps the decision visible without stopping the desk.

A warning that says this posting may not comply is useless. It gives a recruiter no way to act and no way to judge whether the tool is right. Every flag names the jurisdiction that triggered it, states what that jurisdiction expects in plain terms, and shows the date the entry in the rules library was last reviewed.

That last part matters more than it looks. A rule quoted without a date is a rule you cannot evaluate, because the reader has no way to know whether it reflects a change made last month. A dated entry tells a recruiter, and more importantly tells your counsel, exactly what the tool was working from when it raised the flag.

Flags are also explainable to a client. When an agency has to tell a hiring manager that the range has to appear in the advert, arriving with the jurisdiction and the substance of the requirement is a different conversation from arriving with a warning triangle.

Pay transparency is not only about the advert. A number of jurisdictions restrict asking a candidate for their salary history, and the question usually survives in an application form somebody built two years ago, or in a screening template that gets copied from requisition to requisition without anyone rereading it.

The validator inspects the application form and the screening questions attached to the requisition alongside the advert copy, and flags a salary-history question where the destinations include a jurisdiction that restricts it. Expected-salary questions are treated separately from salary-history questions, because they are usually treated differently, and the flag says which one it thinks it has found.

Templates are checked once and inherited, so a fix applied to a screening template propagates to every requisition using it rather than being applied forty times by hand and missed on the forty-first.

The library behind the validator is maintained by us and every entry carries a review date, visible in the flag. We will publish when each jurisdiction was last reviewed, because a compliance tool whose currency you cannot inspect is worse than no tool, and gives a false confidence that is harder to correct than ignorance.

What we will not do is describe the library as authoritative. Rules change, take effect on dates that are announced in advance, and are interpreted by regulators and courts in ways a lookup table does not capture. Where the library is behind, the flag will be wrong, and the review date is what lets you notice.

Your counsel is the authority on what applies to your postings. The validator is a prompt for a human decision at the moment that decision is cheapest to make: before the advert goes out, in front of the person writing it.

It does not decide what your range should be. Setting a good-faith range is a compensation decision that depends on your bands, your budget and the market you are hiring in, and no tool can take it from you. The validator checks that a range is present, structured and plausible against the requirement; it has no opinion on the number.

It does not certify a posting. There is no green tick that means this advert is lawful, because that is not a statement a vendor can make about a document it did not write for a role it does not employ into. A posting that passes every check has passed every check the library knows about on the date shown.

It also does not reach postings you publish outside the system. An advert typed straight into a board, or a role a client posts themselves under your name, is not in the publishing path and cannot be checked. Where a desk works that way, the honest answer is that the validator will not help until the posting comes through the product.