I think HR should write down what happens after a fraud flag before any vendor’s detection feature goes live. The tools being launched this autumn say a flag starts a human review. Whether that review happens, and who runs it, is a decision for the employer.
What the vendors say a flag means
Hirevue’s November 10 release on advanced fraud detection states that a flagged signal prompts human review and does not, by itself, establish fraud or determine a hiring outcome. It also says hiring teams can adjust which signals are active and how much weight each carries, based on the job’s requirements and the patterns in their applicant pool. Stand8, which launched a verification platform on September 28, says its platform informs human decisions and does not make automated hiring decisions, and that candidates receive notice and give consent before identity or biometric verification.
Those are sensible design statements. They also hand the employer three jobs the software cannot do: decide the weights, staff the review, and tell candidates what to expect. We described where each vendor places its checks in our comparison of Hirevue, iCIMS and Checkr. This piece is about what the buyer owes the process once the check is on.
Why the review is the weak point
Checkr’s survey of 3,000 U.S. managers involved in hiring, published October 15, gives a reason to doubt that the review will be done carefully. Twenty-eight percent of those managers said they had knowingly advanced a candidate despite suspecting dishonesty because they needed to fill the role quickly. Only 23% were extremely confident their hiring process would catch a fraudulent candidate.
Put those two numbers together and the risk runs in both directions. Under deadline pressure a flag can be waved through. A flag can also be treated as a verdict by a recruiter who has no procedure to follow and no time to ask a candidate for more information. Hirevue says its notifications prompt recruiters to ask for more information. Whether a recruiter can do that on a Friday with a requisition aging is a staffing question.
Ilan Frank, Checkr’s chief product officer, said in the release: “Employers can no longer assume that a polished resume, a convincing interview or a seemingly valid identity tells the whole story.” I agree with the premise. A flagged signal is an extra piece of the story and nothing more, and an employer that treats it as the ending has swapped one assumption for another.
The strongest objection
The best argument against my position is that speed matters and fraud is real. Checkr reports that 40% of managers have interviewed a candidate later found to be using a fake identity, up from 31% a year earlier. If an employer waits to draft a review standard, more of those candidates pass through. A short, simple rule, such as rejecting on any flag, would at least stop them.
The rule would also stop honest candidates who happen to use a VPN, a shared network or an unusual device, which are the kinds of location context the Hirevue release describes. Stand8’s release is careful here, listing the Illinois Biometric Information Privacy Act, the Illinois Artificial Intelligence Video Interview Act, New York City Local Law 144 and California’s automated-decision system employment regulations as requirements its program is designed around. An employer cannot borrow that design by buying the software. It has to supply its own notice language, consent step and human decision point, and a reject-on-flag rule skips the last of those.
The standard I am asking for does not need to be slow. It can be a one-page procedure completed before launch and refined after the first month of flags.
What I would write down
First, name who reviews a flag and how many working days they have. Second, list the follow-up a recruiter may ask of a candidate, such as a second identity document or a short live conversation, and apply it to every flagged candidate the same way. Third, decide the signal weights by role and record who approved them, since Hirevue describes weighting as an employer choice. Fourth, write the notice candidates see before the process begins and keep a copy. Fifth, track the outcome of every flag, including the ones that turned out to be nothing, so the weights can be corrected.
The last item connects to an argument I made earlier, that pre-apply verification needs a baseline before it earns a budget line. A flag log is that baseline. Without one, the next budget conversation will rest on the vendor’s numbers and Checkr’s survey, and HR will have no figures of its own.
Source: Hirevue, PR Newswire. Also drawn from Stand8 and Checkr.
