Finding candidates is getting faster while trusting them is getting harder
On September 29, LinkedIn announced Hiring Assistant 2, the next generation of its AI recruiting agent, with stronger reasoning, memory and personalization for candidate discovery, context and talent-pool management.
According to LinkedIn product data, recruiters using Hiring Assistant review 83% fewer profiles to find qualified matches and find interview-quality applicants 33% faster. LinkedIn says prescreening is completed in a median of six minutes.
Yet the same announcement says 39% of US talent-acquisition professionals cite not knowing whether a candidate or their claimed skills are real as their top challenge. As Search accelerates, Verification is moving up the bottleneck list.
The funnel is moving from a Scarcity Problem toward a Volume Problem
LinkedIn says global hiring is about 30% below pre-pandemic levels while applications per applicant are up about 30%. Qualified talent can remain scarce even while recruiting teams face much more application volume.
Greenhouse benchmark data from more than 6,000 companies and more than 640 million applications show annual applications handled per recruiter rising from 1,610 in 2022 to 4,890 in 2025, an increase of 203.7%.
When candidate volume grows this quickly, the problem is no longer only discovering people. It is separating reliable signals from a much larger pool.
AI improves candidate presentation as well as recruiter productivity
Candidates can use AI to improve resumes, tailor experience descriptions to a job description and prepare for interviews. None of that is automatically deceptive.
But the quality of presentation can become less tightly linked to actual capability. A polished narrative and real experience are not the same thing.
Recruiting teams may therefore need to inspect the evidence chain underneath important claims rather than using document quality itself as a strong signal.
Greenhouse says 91% of recruiters have encountered candidate deception
The Greenhouse 2026 AI in Hiring Report surveyed 1,200 job seekers, 219 recruiters and 446 hiring managers in the United States.
Greenhouse later reported that 91% of surveyed recruiters had encountered some form of candidate deception during hiring. Common forms included resume exaggeration at 63%, fake references at 48% and AI-assisted interview responses at 35%.
This does not mean 91% of candidates are deceptive. It means 91% of the recruiters surveyed said they had encountered at least one form of deception in their hiring work. Verification itself is becoming an independent operating cost.
Search → Match → Verify → Judge are different problems
Search asks where relevant candidates are. Match asks which profiles are closest to the current role.
Verify asks what evidence supports identity, employment, education, project experience and skill claims. Judge asks whether the verified facts add up to a person likely to solve a particular problem in a particular organization.
Making Search and Match cheaper does not automatically make Verify and Judge easy. More volume and smoother signals can actually increase the attention required at the back of the funnel.
Identity verification is not capability verification
Even when a person is real, their employment is real and their education is verified, hiring is not finished.
Teams still need to understand what the person actually did, the level of responsibility they held, their individual contribution to team results and whether the experience is portable into a different environment.
Identity Verification is therefore not the same as Capability Verification, and Capability Verification is not the same as Hiring Judgment.
The value of a good search partner can move from database size toward evidence and judgment
As AI searches large talent pools, ranks candidates and assists outreach and screening, simply saying that a firm can find many candidates becomes a weaker form of differentiation.
Value can shift toward defining the actual role, separating career claims from performed work, judging which adjacent experiences are transferable and validating whether a candidate is realistically movable under the role’s authority, compensation and risk.
The deliverable can rise from a list of ten names to an explanation of why these people are real candidates, what has been validated, where the risks sit and which trade-offs support a hiring decision.
Trust cannot be rebuilt by verifying candidates alone
An aggressive verification process can damage candidate experience if every applicant is treated as suspicious.
Greenhouse’s 2026 Candidate AI Interview Report found that 70% of US candidates who had experienced AI evaluation said they were not clearly told in advance that AI would evaluate them in their most recent AI interview.
Trust is therefore two-sided. Employers can validate candidate evidence while also being transparent about which AI systems are used, where they enter the process and how human oversight works.
BANSEOG VIEW | As Search Cost falls, Judgment Value can rise
LinkedIn’s 83% fewer profiles and 33% faster figures suggest the front of the funnel is becoming cheaper. The same announcement’s 39% authenticity challenge suggests the bottleneck is moving downstream.
Banseog frames the shift as Scarcity Problem → Volume Problem → Trust Problem and Search → Match → Verify → Judge.
If candidate discovery becomes a commodity, the premium product in search can become Judgment: combining Evidence, Context, Verification, Mobility and Trade-offs to help a company answer the final question — should we actually bet on this person?
BANSEOG VIEW
Banseog View — Search → Verify → Judge
As Candidate Discovery becomes cheaper, Verification Work can become relatively more valuable.
Identity Verification, Capability Verification and Hiring Judgment are separate stages.
Search differentiation can move from Candidate Lists toward Evidence + Context + Trade-off + Decision Support.
SOURCES
Primary sources and references
- LinkedIn — Hiring Assistant 2
Sep. 29, 2026. Confirms Hiring Assistant 2, 83% fewer profiles, 33% faster interview-quality applicant discovery, six-minute median prescreen, global hiring down 30%, applications per applicant up 30%, and 39% citing candidate/skill authenticity as a top challenge.
- Greenhouse — The 2026 AI in Hiring Report
Confirms the US survey sample of 1,200 job seekers, 219 recruiters and 446 hiring managers and the report’s focus on trust and validation gaps.
- Greenhouse — Hiring Benchmarks 2026
More than 6,000 companies and 640M applications; annual applications per recruiter increased from 1,610 in 2022 to 4,890 in 2025, or 203.7%.
- Greenhouse — 2026 Summer Awards / Candidate Fraud
Cites the 2026 AI in Hiring Report: 91% of recruiters encountered candidate deception, including resume exaggeration 63%, fake references 48% and AI-assisted interview responses 35%.
- Greenhouse — 2026 Candidate AI Interview Report
Confirms 70% of US candidates with AI-evaluation experience said AI evaluation was not clearly disclosed before their most recent AI interview.
LinkedIn’s 83%-fewer-profiles and 33%-faster figures are LinkedIn product data, not independent benchmarks guaranteed to reproduce across employers. Greenhouse’s 91% figure means 91% of surveyed recruiters reported encountering candidate deception; it does not mean 91% of candidates are deceptive. Search → Match → Verify → Judge and Search Cost ↓ → Judgment Value ↑ are Banseog analytical frames connecting the public evidence.