A recruiter reaches a candidate, confirms the person exists, confirms the résumé is theirs and still walks away thinking, “Then why does this application feel wrong?” That question is becoming more common because authenticity is only one part of application integrity.
In 2026, applicant volume is colliding with automation on both sides of the hiring process. Greenhouse says its benchmark analysis across more than 6,000 companies and 640 million applications found applications surged while recruiting teams shrank. LinkedIn reports U.S. applicants per open role have doubled since spring 2022, while recruiters simultaneously say finding qualified talent has become harder. Those facts can coexist because more records do not automatically mean more usable demand.
The record is real. That still does not tell you what happened.
Consider four applicants. One carefully reads the job and submits directly. One authorizes a job-search tool to apply broadly on their behalf. One clicks a quick-apply flow after seeing only a short description. One already exists in the employer's CRM and is attached to a requisition by a recruiter or integration.
All four records can land in the same ATS with the same broad label: applicant. Yet the level of intent, context and job-specific action is completely different.
This is why “Is the person real?” and “Did the person meaningfully apply?” should be separate questions. A fraud tool may help with the first. It cannot answer the second by itself.
Stop making one label carry four different decisions.
Most application-quality conversations become confused because four separate questions get collapsed into one:
- Identity: Is the person who they claim to be?
- Intent: Did they knowingly pursue this specific opportunity?
- Record integrity: Does the ATS accurately reflect how the candidate entered the process?
- Fit: Do they meet the job-related standard?
A candidate can be authentic and unqualified. Authentic and low intent. Authentic and attached to the wrong requisition. Authentic and highly qualified. A candidate can also be fraudulent while appearing qualified on paper. Those are different operating problems and should trigger different responses.
Start with the pathway, not the person.
When a record feels wrong, reconstruct how it entered. What did the candidate actually see? What event triggered the application? Did a job-search agent select the role? Was the person already in the employer's database? Was the record imported? Did an agency, recruiter or integration create it? Did the job title or location change after the submission?
The same symptom can have very different explanations. A candidate who cannot remember applying may have authorized broad automation. A candidate who says they never applied may be looking at a title they do not recognize because the requisition was renamed. A duplicate may have been created by an integration rather than by the candidate. An irrelevant application may be low intent, not identity fraud.
That is why the first step should be reconstruction, not accusation.
Give recruiters three lanes instead of one giant bucket called “fake.”
A practical workflow can separate ordinary recruiting from ambiguity and genuine risk:
- Routine qualification: the person and application appear normal enough to evaluate against the role.
- Intent clarification: the person is likely real, but the employer needs to confirm awareness, interest or how the application entered the system.
- Integrity escalation: evidence suggests identity misuse, proxy participation, fabricated credentials or a security concern that requires a different owner.
Those lanes stop every unusual application from becoming a security incident while still giving serious cases a path out of normal recruiting operations.
The hidden cost is the investigation work nobody planned for.
Questionable records rarely disappear on their own. Someone has to open them, compare data, search prior applications, message the candidate, review source fields, disposition the record, document the reason and sometimes escalate the case. That labor is usually absorbed by recruiting and never appears in the media report that celebrated the low cost per application.
If application ambiguity is rising, measure it. How many records require manual clarification? How many candidates deny intent? How many are duplicates? Which application methods create the most exceptions? How many fraud flags are eventually cleared? How long does resolution take?
Those numbers will tell you much more than debating whether “fake applicants” are everywhere.
Do not make the candidate prove a system assumption.
When the employer needs clarification, the language matters. “Our system says you may be fraudulent” is both inflammatory and often unsupported. A neutral request works better: confirm that you applied to the named role, verify the contact method, or clarify the specific discrepancy the employer needs to resolve.
That protects the integrity of the review and the dignity of a legitimate candidate.
What should change this week?
Take a sample of 25 questionable applications. Do not label them first. Reconstruct the pathway and classify each one by identity, intent, record integrity and fit. Then look for concentration. If 18 of the 25 came through the same workflow, you have learned something about the workflow. If most are authentic but low intent, you have learned something about application design. If several show material identity conflicts, you have learned something that belongs with security.
The practical goal is not a perfectly pure application funnel. It is a funnel where the team can tell the difference between person, intent, fit and risk without collapsing them into one score.
Next step: Read When 500 Applications Create Less Hiring Capacity Than 100, then use The New Hiring Math to rethink what application volume is actually worth.