🚪 Your AI hiring process is filtering out high quality candidates

Hey HR folks! 👋 

Quick question before we get into it:

This week we're talking about the AI hiring wall: who built it, how candidates are already learning to work around it, and what that's doing to the quality of talent making it through to your team.

The hiring process has a new gatekeeper, and it's the algorithm. It's great at managing volume, but not so great at finding the people you'd actually want to hire. 😬 

PRESENTED BY SELECTSOFTWARE REVIEWS

The fastest way I've seen HR teams pick software

A few years back, a peer HR director asked me how other teams had shortened the HRIS hunt. Her team had been at it for months.

That's when I pointed her to SSR's free HR software matching service. Their HR software advisors put together a shortlist based on your needs, and they can help you with not only HRIS, but also ATS, payroll, LMS, performance management, EOR, and PEO.

15,000+ HR teams have used it.

4.7/5 average rating.

No fees, no commitments! (yes, really)

IN TODAY’S EDITION

🧠 AI screening tools are filtering out the wrong candidates

🛋️ The HR Break Room: How much do you trust your ATS to surface the right people?

📚 Additional Reading: the DOJ and AI hiring regulation, the women's starting pay gap, and a Chinese court ruling on AI-motivated layoffs

OPENING THOUGHTS

🤖 AI screening tools are filtering out the wrong candidates

Robert Half's 2026 Job Search Guide opens with a line that should stop every hiring manager cold: "Five hundred people just applied for that job you want." That's the current reality for most entry-level and mid-level roles. AI tools have made it trivially easy to generate a polished resume and apply to dozens of jobs in an afternoon, which means recruiting teams are drowning in volume and very logically, reaching for automation to handle it.

ATS tools, keyword screeners, and AI interview platforms now decide who makes the cut before a human ever sees a name. But the same logic that made bulk applying easy also made bulk screening flawed: when the bot looks for pattern matches, it finds them, and misses everything that doesn't fit the pattern.

ZipRecruiter's 2026 Grad Report confirms the entry-level squeeze is real: fewer of the available jobs are entry-level, while more candidates are competing for each one. The grads who are getting through, 77.2% of recent graduates landed roles within three months, got there by learning to work the system.

• Tailoring keywords.

• Reformatting layouts.

• Practicing with AI interview simulators.

They're getting through because they're good at navigating your hiring process, not necessarily because they're the best fit for the role.

LET’S UNPICK

AI tools made applying easier ➡️ 

so more people applied ➡️ 

so companies needed AI to manage volume ➡️ 

so candidates learned to optimize for AI ➡️ 

so applications started looking more alike ➡️ 

so tools became more keyword-dependent ➡️ 

you're now running a process that selects for people who are good at gaming your filter.

Robert Half describes ATS and AI tools as "the career club's bouncer", deciding who gets recommended and who gets a first interview, sometimes even analyzing tone, language, and facial expressions in early-round AI interviews. The guide is written for candidates, but the implication for HR leaders is uncomfortable: 50% of hiring managers say AI has already changed the skills they look for in candidates, yet many of the tools deciding who gets seen were built to find the old signals.

What's getting filtered out:

• Candidates with non-traditional paths: people who moved between industries, took contract work, or built their experience through freelance and gig roles rather than a clean linear track

• People whose best work shows in context: apprenticeships, project contributions, community roles, rather than in standard bullet points

• Candidates from less-resourced backgrounds who didn't have access to career centers or coaches that teach you to beat the bot

ZipRecruiter's data makes this concrete.

🔴 Working during college more than doubles a grad's odds of being employed. 81.6% with college work experience are currently employed, versus 40.7% of those without it. But access to that experience is not evenly distributed.

🔴 66.8% of recent grads cite lack of experience as their biggest obstacle to landing their ideal role. But many of them do have experience, just not the kind that reads cleanly to an algorithm.

🔴 18.3% of employed recent grads say they intentionally applied below their qualification level just to get a foothold. Your tools may be screening out overqualified candidates who would have been excellent.

🔴 The bigger irony: the single most predictive of actual job success is exactly the one your ATS cannot see. 87.8% of employed recent grads say networking was important in securing their first role. Referrals, warm intros, campus connections, none of that shows up in a keyword scan.

Meanwhile, employers are already feeling the downstream effects , 53% say their primary challenge is finding graduates with the right skills, but tightening the AI filter further is not the solution.

The fix here requires being honest about what you’re optimizing for, and adding the human judgment layer back in for the candidates your tools are most likely to undervalue.

Sources:

🔗 ZipRecruiter 2026 Graduate Report 

🔗 Robert Half Job Search Strategies Guide 2026 

🔗 HR Digest: Is There a Shortage of AI-Ready Graduates?

TAKEAWAY AND TRY

1. 🤖 Audit what your ATS is actually filtering for. Pull a sample of rejected applications from the last quarter and manually review 15-20. If strong candidates didn't make it past screening, your criteria likely needs recalibrating.

2. 🧑‍💼 Add a human review stage just below the cut-off threshold. The candidates most likely to have been unfairly filtered sit just under your score threshold. A recruiter spot-check at that tier costs a small amount of time and can surface real talent.

3. 📝 Rewrite job postings with the candidate's algorithm problem in mind. Vague requirements default your ATS to pattern-matching on irrelevant signals. Precise, skills-focused postings attract more relevant applications and make screening more accurate from the start.

4. 💻️ Evaluate your AI interview tools separately from your ATS. Tools that analyze tone, pacing, or facial expressions carry their own bias risks. Ask vendors directly how they tested for demographic fairness, and if they can't answer that, take it seriously.

5. 📈 Track where your best performers actually came from. If your top hires came through referrals or non-traditional paths, make sure your screening process isn't systematically filtering out the channel producing your best people.

TLDR;

AI hiring tools solved the volume problem and created a new one: a filter that rewards candidates who are good at gaming systems, not candidates who are good at the job. Non-traditional paths, non-linear experience, and real-world skills that don't translate neatly into keywords are getting screened out at scale. HR built this wall and HR is the one who can redesign it.

THE HR BREAK ROOM

PRESENTED BY SELECTSOFTWARE REVIEWS

🔴 Register and attend live for the chance to win a free online HR course.

In this practical session, we’ll explore how HR leaders can reposition their function as a strategic partner to the business. If you want a stronger seat at the table and clearer alignment with company growth, this conversation will show you how to make the shift.

ADDITIONAL READING

1. 🏢 Texas AG Claims Employers Ran "Ghost Offices" to Sponsor H-1B Visa Workers Texas AG Ken Paxton has issued Civil Investigative Demands to nearly 30 North Texas businesses accused of registering single-family homes and empty buildings as office addresses to fraudulently sponsor H-1B visas.

2. 📉 Microsoft CFO Flags Workforce Cuts as AI Spending Surges Microsoft's CFO said total headcount will continue declining into the next fiscal year, even as the company's AI business hits a $37 billion annual revenue run rate and capital expenditures are set to exceed $40 billion this quarter.

3. 🤖 Oracle Layoffs and the Rise of Leaner, AI-Driven Teams A 30-year Oracle technical writer found out she was being laid off via a call from her manager — on her way to the hospital for back surgery.

That’s It For Today!

Thanks for reading to the end and we hope today’s edition sparked some new ideas for your workplace! 🧠

We know you’re super busy and really appreciate you saving some room for us in your inbox 😀