The discrepancy a background check will never catch

A candidate clears background verification cleanly. The employer is real, the dates match, the title on the offer letter matches the title on the résumé. Two months in, the line about running Kubernetes in production does not survive the first real incident. Nothing on the background check would have caught that, because nothing on a background check was built to.

That gap is common enough to show up in the numbers. AuthBridge's Workforce Fraud Files — H1 FY26, covering background checks run between April and September 2025, found an overall discrepancy rate of 4.33% among white-collar hires in India, with employment history the single most-flagged category at 11.15%, ahead of address verification at 7.68% and education at 4.49%.

Those numbers say the checks are working. They also say something quieter: what they are checking is narrower than what a hiring manager actually needs to know.

What a background check actually checks

The categories in that report are identity, employment history, address, education, criminal records, and reference checks. Every one of them is a yes-or-no question about a fact. Did this person work at this company. For these dates. At this title. Does this address exist. Is this degree real. A verifier calls the employer, confirms the record, and closes the file.

Those are worth confirming, and getting them wrong is a real, measurable problem in some sectors, not a formality. But notice what is not on the list. There is no category for whether the work matched the title.

AuthBridge's own recommendation for closing that gap is to move screening earlier, before the offer goes out, and to keep checking periodically instead of once at onboarding. That's a reasonable fix for catching identity and employment mismatches sooner. It does nothing for the mismatch that was never in one of those categories to begin with, no matter when you run the check.

The gap it was never built to close

A candidate can pass every check above and still have spent those two years on the edge of what the résumé line implies. They were on the team that ran Kubernetes in production. They watched the on-call rotation more than they carried it. None of that shows up as a discrepancy, because nothing about it is factually false. The company is real, the dates are real, the title is real. The depth is the only part that was never claimed to a verifier, only to a résumé template, and a résumé template does not ask hard questions.

Skill inflation and identity fraud are different problems wearing the same word, and only one of them leaves a paper trail. A verifier can confirm someone sat in a chair for two years. Only a conversation can find out what they actually did from it.

Where it actually surfaces

This is exactly the gap a background check hands off, whether it says so or not. Once the facts are confirmed, someone still has to find out whether the depth is real, and that only happens in conversation. The résumé probe table is the method for that part: one layer deeper than the line on the page, for the claims that show up on almost every technical résumé.

A clean background check and a shallow interview answer are not in tension with each other. They are two different questions, and both are worth asking, in that order: confirm the person was there, then find out what they actually did while they were.

Recio's résumé verification step doesn't try to replace a background check, and it isn't trying to confirm the same facts a verifier already confirms. It runs on top of that, asking the one question a background check was never built to answer: does the depth match the page. More on how that step fits the rest of the interview is on the interview engine page.

A background check tells you the job was real. It was never going to tell you whether the work was.