Key Takeaways
Candidates check AI. 58 percent of job seekers have caught an AI tool giving them inaccurate information about an employer, only 5 percent take an answer at face value, and two thirds then go to the company's own website (PerceptionX, 2026).
Surprise costs you candidates; AI does not. 70 percent of US candidates who faced an AI interview say it was not clearly disclosed beforehand, and 38 percent have walked away from a process because of one, yet only 7 percent would refuse an AI interview outright (Greenhouse, April 2026; October 2026).
In staffing, a good AI experience brings people back. 81 percent of candidates who met AI through their staffing firm rated it positively, and 93 percent of those with a positive experience would work with that firm again (Bullhorn, September 2026).
The premium goes to the firm whose claims survive the check. Recruiters now verify by hand too, so a number with its source and its consequence attached is worth more than a promise.
The AI Pre-Screen is what happens before your first conversation with a candidate. A candidate opens an AI tool and asks about your firm. The answer comes back confident, well organized, and sometimes wrong. What the candidate does next decides whether they ever reply to you, and most firms never see it happen.
The Verification Loop
PerceptionX surveyed 306 job seekers across seven countries in May, 295 of them AI users. Fifty-eight percent had already caught an AI tool giving them inaccurate information about an employer, including outdated remote work policies, wrong salary data, and company details the tool invented. Only 5 percent take an AI answer at face value. Forty-one percent actively verify the claims that matter to them, and just 16 percent said the AI response was sufficient on its own.
So where do they go to check? Two-thirds visited the company website, and two-thirds ran a Google search. Forty-eight percent opened LinkedIn. We call this The Verification Loop: the candidate asks the AI, gets an answer, then tests it against the places a firm actually controls. Your website, your LinkedIn page, and whatever a search returns under your name are the last stop. Our read is that the candidate believes whichever source gives them something specific to check.
The loop cuts both ways. For 44 percent of respondents, the combination of AI and their own fact-checking made them more interested in some companies and less interested in others. For executive candidates the exposure is sharper. PerceptionX notes that leadership quality is among the topics AI is most likely to get wrong, and leadership is the question a C-suite candidate cares about most.
Candidates Are Rejecting Surprises, Not AI
Greenhouse's spring survey of 2,950 job seekers across five countries found that, among US candidates who had been through an AI evaluation, 70 percent said AI use was not clearly disclosed before their most recent AI interview. Thirty-eight percent of US candidates have already withdrawn from a process because it included an AI interview, and a company failing to explain how AI would be used ranked among the top reasons. The stakes show up in reputation: when an AI interview went well, 38 percent of US candidates came away with a better impression of the employer. When it went badly, 34 percent came away with a worse one.
Opposition to AI is not what drives this. In Greenhouse's October follow-up, only 7 percent of US candidates said they would refuse an AI-led interview outright. What they asked for was information: 49 percent wanted to know a human would review their responses, and 40 percent wanted to know how they would be evaluated.
The staffing data points the same direction. Bullhorn surveyed about 2,300 professionals who had worked with a staffing firm in the past three years. Sixty-nine percent had met AI through their firm, and 81 percent rated the experience positively. Among those with a positive experience, 93 percent said they would work with that firm again. The technology is a given by now. Disclosure is the variable.
The Trust Gap Runs Both Ways
Recruiters carry their own version of the problem. In an October TestGorilla survey of more than 1,400 US talent professionals, 72 percent said AI made finding candidates faster. A third of them gained no confidence in the candidates they found, and 18 percent said their confidence actually fell. More than two thirds now spend an hour or more checking a candidate's credentials by hand. LinkedIn CEO Dan Shapero has said it is getting harder for employers to know who can actually do the job, with job seekers sending about 30 percent more applications than before the pandemic.
When neither side can take anything at face value, whoever shortens the checking gets the conversation. The American Staffing Association's June report with Prodoscore suggests where firms are spending the hours AI frees up: recruiter call time reached a record 286 minutes a week in the first quarter, double the level of early 2024. That is activity data, not trust data. It does show firms betting that a human voice closes the gap.
What a Checkable Claim Looks Like
Start with what candidates actually catch. The PerceptionX respondents who found AI errors named outdated remote work policies, wrong salary data, and company details the tool invented. Those are plain facts, and most firms leave them for a search engine or an AI tool to guess at. The candidate then arrives at your website looking for the real answer.
A checkable claim has three parts: a specific fact, the method or source behind it, and the date it was last true. “Flexible work” fails all three. “Hybrid for corporate roles, fully remote for recruiters, last updated September 2026” passes. Staffing leaders would never accept “competitive” on a rate card, yet it fills most careers pages.
Three places to test your own firm:
Pay and work model. State the range and the policy for each role.
Leadership. Give names, tenure, and what each leader has run. PerceptionX found leadership quality among the topics AI is most likely to get wrong, and a C-suite candidate will test that one first.
AI in your process. Say where it is used, where it is not, and who reviews the result. In Greenhouse’s October survey, 49 percent of US candidates wanted to know a human would review their responses, and 40 percent wanted to know how they would be evaluated.
Then run the test. Pull up your careers page and leadership bios, ask an AI tool to describe your firm, and compare the two line by line. Every place the answer is wrong, or the page gives nothing to check, is a gap a candidate will find before you do.
What the Premium Actually Is
No study has put a price on trust in executive search, and we will not pretend otherwise. What the research does show is where candidates lose it and where they keep it. They lose it to silence: 53 percent of job seekers were ghosted in the past year, up from 38 percent in 2024, according to Criteria's candidate experience report as reported by Fortune. They lose it to surprise, as in a process that turns out to involve AI nobody mentioned. They keep it through regular contact: Bullhorn found that 91 percent of candidates who hear from their recruiter at least weekly would continue working with that recruiter. The premium is what a candidate extends to the firm whose answer matches what they find when they check.
One caution on the evidence. Most of these studies come from vendors with a stake in the answer, including PerceptionX, Greenhouse, Bullhorn, and TestGorilla, and samples range from 306 to roughly 2,950. Read the figures as direction, not precision. Inspect what you expect. Candidates have started doing exactly that to you.
Morgan Taylor Executive Search runs healthcare staffing leadership searches on claims that can be checked. If you are about to recruit your own executive team, hold every firm you speak with to that standard, including us.
