01No source found
“Companies that combine AI screening with human-led final interviews cut time-to-hire by 40% while improving first-year retention by 25%.”
Found on SelectSoftware Reviews, AI Recruiting Statistics, checked October 21, 2026.
What the trail shows
SelectSoftware Reviews footnotes the sentence to a PrideStaff guide, AI in Hiring: How to Leverage Technology Without Losing the Human Touch. The guide is a staffing firm's marketing brochure: it names no study, sample or period, and its own footnote for the figure points to a JobsPikr blog post on job-demand forecasting that contains neither number.
What holds instead
The best controlled evidence on AI-run interviews feeding human decisions is a pre-registered field experiment with 67,056 job applications: applicants interviewed by an AI voice agent, with recruiters making every offer decision, were 12% more likely to receive an offer and 18% more likely to still be in the job a month later. It did not measure time to hire.
Our own record: We had both figures in our Index and three reports, and withdrew them on October 20, 2026.
02No source found
“The average cost-per-hire reduction is 30%, with some North American companies reporting 40%.”
Found on InCruiter, AI in Recruitment 2026, checked October 21, 2026.
What the trail shows
InCruiter states the average with no source. The earlier version we could trace makes a weaker claim: a SHRM Labs summary of a talk by a recruiting-AI vendor's chief executive says AI recruitment reduces cost per hire 'by as much as 30%', with no study cited. Between the two, 'as much as' became 'average'. The '31% faster hiring' InCruiter puts beside it is credited to SHRM; we found no SHRM publication that states it.
What holds instead
No study we could find measures how much AI changes cost per hire across employers. The reliable number is your own: record cost per hire for a set of roles before a pilot and for the same kind of roles after.
Our own record: We had it in our Index, three reports and the defaults of our ROI calculator, and withdrew it on October 20, 2026; the calculator now asks for your own figure.
03Not in its cited source
“Gartner predicts that by 2028, 30% of recruitment teams will rely on AI agents for high-volume hiring and early-stage tasks.”
Found on Pin, AI Talent Acquisition Guide, checked October 21, 2026.
What the trail shows
The sentence carries no link. The Gartner release linked in the next sentence, on the top trends for talent acquisition in 2026 (October 7, 2025), contains no 2028 projection and no 30% figure: it projects that by 2027, 75% of hiring processes will include certifications and tests for workplace AI proficiency, and that by 2030 half of enterprises will face irreversible skill shortages in critical roles. We found the 30% in no Gartner release we could read; Gartner's paid research we could not check.
What holds instead
On intent rather than prediction, the measured figure is Korn Ferry's: 52% of talent leaders plan to add autonomous AI agents to their teams in 2026.
Our own record: We quoted it in our agentic AI recruiters guide, and removed it on October 20, 2026.
04Not in its cited source
“Semantic AI sourcing tools now expand candidate pools by an average of 340% over traditional Boolean strings and surface 60% more relevant profiles per query, according to 2026 Gartner benchmarks.”
Found on Skillfuel, AI Sourcing Tools vs Boolean Search, checked October 21, 2026.
What the trail shows
No Gartner document is named or linked. Gartner's own release on talent acquisition trends for 2026 contains neither figure, and we found no Gartner benchmark that does.
What holds instead
We found no independent benchmark of how many more of the right people intent-based search finds than a Boolean string. A buyer can run the test directly: the same open roles searched both ways, with a recruiter rating the two shortlists without knowing which is which.
Our own record: We quoted it in our sourcing tools guide, and removed it on October 19, 2026.
05Real figure, wider claim
“Deepfake candidates are infiltrating hiring pipelines at an alarming rate, with 41% of organizations reporting they have unknowingly hired a fraudulent candidate.”
Found on The Hire Hub, Deepfake Candidates, checked October 21, 2026.
What the trail shows
The 41% is real but measures a narrower group, and 'unknowingly' is not in it. In GetReal Security's survey of 668 IT, cybersecurity, fraud and risk leaders at organizations with at least 1,000 employees, run in September 2025, 41% said their company had hired and onboarded a fraudulent candidate.
What holds instead
Quote it as the survey measured it: 41% of security and fraud leaders at large organizations report that their company hired and onboarded a fraudulent candidate.
Our own record: We quoted it as 41% of companies in two reports, and corrected it on October 19, 2026.
06Real but old
“94% of recruiters agree that their ATS has had a positive impact on their organization's hiring processes.”
Found on SelectSoftware Reviews, ATS Statistics, checked October 21, 2026.
What the trail shows
No source is given. The figure matches Capterra's Recruiting Software Impact Report from 2015: 102 HR and talent professionals responded, about three in four of them used recruiting software, and 94% of those users said it had improved their hiring process.
What holds instead
It is an 11-year-old figure from roughly 77 software users, and it says software improved the hiring process, not that recruiters rate their applicant tracking system well.
Our own record: We quoted it in our ATS market report, and corrected it to its source on October 20, 2026.
07Superseded
“Structured interviews have a predictive validity of 0.51 for job performance.”
Widely quoted; checked October 21, 2026.
What the trail shows
0.51 is Schmidt and Hunter's 1998 meta-analytic estimate, still quoted across hiring guides. In 2022 Sackett, Zhang, Berry and Lievens re-analysed the evidence and concluded that the earlier corrections for range restriction were too generous.
What holds instead
Their revised estimates put structured interviews at about 0.42 and unstructured interviews at 0.19. Structured interviews still rank first among the selection procedures reviewed, so the advice holds; the number is lower.
Our own record: We had cited 0.51 to a later paper that does not state it, and corrected it on October 6, 2026.
08Real figure, wrong context
“Deepfake-related fraud losses exceeded $410 million in the first half of 2025, with some single incidents exceeding $680,000.”
Widely quoted; checked October 21, 2026.
What the trail shows
The $410 million is Surfshark's count of media-reported deepfake fraud losses in the first half of 2025, across every kind of fraud: of the roughly $900 million it has counted since 2019, $401 million came from celebrity deepfakes promoting fake investments. It is quoted in coverage of deepfake job candidates but is not a hiring figure. We could not trace the $680,000 to any study.
What holds instead
For hiring, the measured figure is GetReal Security's: 41% of 668 security and fraud leaders at large organizations say their company hired and onboarded a fraudulent candidate.
Our own record: We quoted both figures in our hiring-fraud report, and corrected them on October 20, 2026.
How we checked
For each figure we opened the page quoting it, followed its citation, and kept going until we reached a study, survey or filing, or ran out of citations. A figure counts as sourced only when the original states it for the group it is quoted about. Pages that block automated reading were read through archived copies, which are linked.
AIRecruiter.co is published by HeroHunt.ai, which sells AI recruiting software, and some of the pages named here belong to companies in the same market. We name a page only to show where we found a claim, we applied the same test to our own reports first, and anyone can follow the links and check the work.