Two recent pieces of research suggest that assumption is worth re-examining, though only one of them touches recruitment or workplace pay at all.
The first is a whitepaper from call analytics platform Infinity, based on 41 million phone calls tracked over the past year across what the paper describes as "every major sector." It contains no recruitment data. No candidates, no job seekers, no hiring calls. What it shows is how AI search is changing the way buyers approach a purchase before they ever speak to a human, and what that does to the call once it happens.
The second is a survey of over 2,100 US adults reported by HR Dive, conducted by Ruth AI and The Harris Poll. This one is closer to home: it's workplace research, specifically about pay, published by an HR trade outlet. It surveys general workers thinking about their own pay, not candidates in a hiring process, and finds that a third of them have already asked an AI tool for help with it.
Put those two together and there's a reasonable case, not proof, that the same shift is starting to reach hiring conversations too. Here's how each piece fits.
What buyers are already doing: Infinity's data
Infinity's research shows the shift most clearly in why sales calls fail. A year on, price disputes sit well ahead of any other objection, a category that's grown sharply since 2025. Meanwhile, the objection that used to dominate, buyers simply not understanding what was on offer, has shrunk to a fraction of what it was. Read together, the two trends point the same way: fewer people are calling confused, and more are calling to argue over a figure they already believe is correct.
As Daniel Wilkinson, Chief Customer Officer at Infinity, put it: "Buyers are arriving at the call already researched and already price checked. When the number the agent quotes doesn't match the number already in the buyer's head, the call fails. That isn't a lead quality problem, it's an expectation problem, and the expectation is set long before anyone picks up the phone."
That's the shape of it: research happens silently, before contact, and the human conversation becomes a negotiation rather than an education. None of it is about recruitment, but it's a clear description of what to look for if the same shift is happening somewhere closer to home.
The same behaviour, already showing up in pay: Harris Poll's data
This is where the second source matters. Infinity's whitepaper establishes the pattern exists somewhere. The Ruth AI and Harris Poll survey establishes that it already exists specifically around pay, something almost every hiring conversation touches on at some point.
One in three respondents has already asked an AI tool about their salary, a raise, or how to negotiate one. Nearly half said they'd let AI handle a pay negotiation entirely. That's not a survey of job candidates, and it isn't evidence about hiring specifically. But it confirms the exact behaviour Infinity describes for buyers, benchmarking a number with AI before a human conversation, is already active in pay and compensation.
Why that adds up to a real question for recruiters
Put the two together and the case looks like this: Infinity shows what happens to a conversation once buyers arrive pre-priced. Harris Poll shows people are already doing pre-pricing specifically around pay. Recruitment sits at the intersection of both, a conversation about a number, conducted by phone or call, that used to rely on the recruiter supplying information the candidate didn't have.
If a modest share of candidates are already doing what that Ruth AI and Harris Poll data shows a third of adults doing, arriving with a benchmarked figure already in mind, then a recruiter's screening call could plausibly fail for the same reason Infinity's sales calls do: not a poor match, but a number that doesn't match what the candidate was already told, by an AI system, to expect. No data confirms this is happening in hiring yet. It's a prediction the two sources support, not a documented outcome.
Worth testing directly
Neither source proves this is happening in your pipeline specifically. That's easy to check without waiting for someone else to publish the recruitment-sector version of this research:
- Ask directly. Add a question to first-stage screens: "Have you looked into what this type of role typically pays before this call?" It costs nothing and tells you immediately whether this is happening in your pipeline.
- Track why candidates go cold. If someone disengages after a call or offer, log the stated reason. A pattern of "the pay didn't match what I expected" is the signal both sources here would predict.
- Lead with the range, not around it. If candidates are arriving with a number in mind, vague talk of "a competitive package" costs more in wasted process than an early, honest figure would.
There's a separate point in Infinity's research worth borrowing, regardless of how the pay question plays out. Wilkinson argues businesses are sitting on a data source they routinely ignore: "If you're a marketer still treating phone calls as an offline black box, you're leaving your best first-party data on the table."
Recruitment teams have their own versions. Screening calls, offer conversations, and "why did you drop out" calls are exactly where a mismatch between candidate expectation and real offer would first show up. Whether that reason is being captured and fed back into how roles are priced is worth checking, since it isn't something either source here can confirm one way or the other.
Neither piece of research set out to study hiring. One shows how the pattern plays out once it takes hold; the other shows it's already active in how people handle pay. Together, that's a case worth testing against your own data rather than dismissing.
