For a decade, AI in hiring worked for one side of the table. Employers had resume parsers, keyword filters, ranking algorithms and automated rejections. Candidates had job boards and guesswork. That asymmetry has now collapsed. Job seekers research companies, tailor every application and rehearse interviews with the same class of AI that screens them. I build candidate-side AI tools for a living, so I watch this from the candidate’s seat every day. The consequence matters to every talent leader: leverage is shifting toward candidates. Hiring built on gatekeeping is quietly losing its grip, and the employers who adjust first will hire best over the next few years.
How AI in Hiring Went One-Sided
The employer side armed itself first, and it armed itself thoroughly. Jobscan’s 2025 analysis found applicant tracking systems in use at 97.8% of Fortune 500 companies. On top of that sit knockout questions, keyword matching, algorithmic ranking and one-way video interviews.
For years, the candidate on the other end had no equivalent. I remember that version of the market personally. When I first tried to break into it, I collected rejections for months. Nobody told me why. I had no data, no feedback and no way to practice. The tools doing the judging were enterprise software, and nothing comparable existed for my side. That was AI in hiring, version one: a capability only one side could buy.
Large language models closed that gap in roughly two years, at consumer prices. Capterra’s survey of nearly 3,000 job seekers across 12 countries found 58% already using AI in their search. The same survey found AI users completing 41% more applications, 157 versus 111 on average.
What Candidates Do With AI in Hiring Now
Candidate-side AI in hiring comes down to three behaviors: research, tailoring and rehearsal.
Research means a candidate arrives at the first call already knowing your pay bands, your review-site patterns and the shape of your interview process. Tailoring means every resume and cover letter can be rewritten in the language of a specific posting within minutes.
Rehearsal is where I see the biggest change. Here is what it looks like in the tools I build. A candidate runs deep research on a company, its strengths, its weaknesses and how their own skills would land there. Then they run mock interviews against different interviewer styles. One asks generic questions. One digs hard into their resume. One asks company-specific questions, one goes technical and one is deliberately aggressive. Another follows up whenever an answer sounds vague. Afterward the candidate gets feedback and runs it again. A year of that practice used to require a paid coach. Now it takes an evening.
Employers already feel the volume side of this shift. Greenhouse reported an average of 222 applications per job opening in early 2024, almost triple the level of late 2021. TalentCulture has covered how AI hiring operations are redefining recruiter workloads in response. The volume is real, and so is the logic behind it. In Greenhouse’s research, 35% of candidates said using AI felt fair because companies were probably using AI on their resumes anyway. The guess was accurate. SHRM’s 2025 Talent Trends research puts organizations using AI to support recruiting at 51%.
Why the Leverage Moved
Start with the screening layer. When both sides optimize for the same keyword filters, matching stops carrying information. A perfectly aligned application no longer tells a recruiter much about the person behind it.
Information parity came next. Salary secrecy, vague job descriptions and mystery interview processes only worked while researching a company was expensive. That research is now nearly free, and candidates filter accordingly. In one Capterra study, 38% of job seekers said they would likely turn down an offer if the recruiting process leaned too heavily on AI. Candidates run their own screens now, on criteria employers are not used to being measured against.
The cost structure flipped too. Applying used to be expensive in time, so volume stayed manageable and process friction worked as a filter. Applying is now cheap, which makes friction useless as a filter and costly as a first impression. Every extra form field and every silent week selects against the candidates with options, and those are exactly the people you wanted.
Add it up and AI in hiring has a new scarce resource. It used to be access: to openings, to pipelines, to a recruiter’s attention. Today it is the attention and trust of the specific candidates you want. Employers earn that by competing on clarity and candidate experience rather than on gatekeeping.
Five Moves for Talent Leaders in a Symmetric Market
- Publish the real job. State the salary range, the actual responsibilities, the stages of the interview process and a decision timeline. The EU’s Pay Transparency Directive makes pay disclosure to applicants mandatory across Europe from June 2026, and several U.S. states already require ranges in postings. Get ahead of the law. Researched candidates compare your posting against every rival’s, and vagueness reads as a warning.
- Answer every applicant. In Stepstone’s 2025 survey of 8,100 job seekers, 64% reported never hearing back after applying. Greenhouse found 61% ghosted even after an interview. You have AI too. Pointing it at status updates and decision dates costs almost nothing, and where legal caution allows, add a line of real feedback. Candidates talk about their treatment, and the next cohort’s research surfaces it.
- Move a human forward. Count how many automated steps a candidate clears before meeting a person, then cut that number. A second Capterra finding: 58% of job seekers are more likely to apply when the posting states that humans make the hiring decisions. Say it, make it true early in the process and brief that first human properly.
- Fix your speed to decision. Candidates with AI run wide, parallel searches, so your strongest applicant has more processes running than you assume. Set an internal service-level agreement for every stage: days to review, days to schedule, days to decide. Slow processes lose their best options first, and that loss never shows up in your funnel metrics.
- Brief interviewers for rehearsed candidates. The person across the table has studied your company and practiced against tougher interviewers than yours. An interviewer who opens the resume for the first time on the call signals that preparation only flows one way. Match the candidate’s preparation and ask follow-ups that go past the rehearsed surface. The conversation is two-way diligence whether you treat it that way or not.
The Employers Symmetry Rewards
Symmetry sounds threatening if your process depended on candidates knowing less than you. For everyone else it is good news. Clear offers, honest postings and respectful processes used to be invisible virtues. Informed candidates can finally find the employers who practice them, and they are choosing accordingly.
I expect the next couple of years to sort talent teams into two groups. One will spend its AI capacity on thicker screening walls. The other will spend it on faster answers, clearer postings and earlier human contact. Candidates have already picked up their side of AI in hiring. The open question is what employers will do with theirs.
Post Views: 1,429
