The More Perfect Your Resume, the More Useless It Becomes

Dabulis was recently laid off. He decided to look for a job seriously—and to do it “smartly.”

He read every available guide and followed their advice. He rewrote his resume with AI, aligned keywords with job descriptions, and even ran it through an ATS check to confirm the machine could parse every section correctly. Then he opened LinkedIn, used the Easy Apply button, and started submitting.

One, ten, one hundred applications. AI-powered submission tools made the process absurdly fast, almost thoughtless. He got into a rhythm and applied to over 700 positions.

Responses: zero.

Not 700 rejections—just 700 applications sent out without even receiving a single automated rejection email. He did everything “right”—his resume was optimized, his submissions were efficient, his keywords matched, and the entire process perfectly followed every guide’s checklist. The result was complete silence.

Later, he realized something: when you spread yourself so thin across hundreds of applications, you’re not a candidate—you’re just a whisper in the wind.

The irony is this: he wasn’t buried in silence because he didn’t use AI—he was buried because he used it too smoothly. At the same time, hundreds of thousands of others were using the same tools and methods, stuffing similarly optimized resumes into the same handful of recruiters’ inboxes. In 2025, over 1.2 million people used AI to job search, and LinkedIn application volume surged 45% in a year. Recruiters described themselves as “drinking through a fire hose.” Dabulis wasn’t a victim of this flood—he was just one drop in it.

So this article isn’t asking “how to make your resume better.” Beneath that question lies a more uncomfortable premise: if resume optimization truly worked, why would someone who perfected theirs and applied 700 times hear nothing but silence?

The Résumé Was Originally a Cost Signal

To understand why resumes fail today, we first need to understand why they ever worked.

A résumé was never a complete proof of ability. It’s too short, too easy to embellish, too hard to verify—and recruiters know this. But for a long time, it accomplished one key task: helping employers quickly rank dozens or hundreds of applicants to decide who’s worth a conversation first. It could do this because of an implicit assumption—creating a good résumé costs effort.

That cost was a real barrier. You had to reflect on and organize your experiences, decipher what a company truly wants from the job description, highlight the relevant parts, and tailor them. Someone willing to do this for a specific role sends a completely different signal from someone who mass-sends the same résumé to 200 companies. Recruiters can’t verify every line, but they can read your sincerity from “Was this résumé tailored for me?” Résumé screening filters not just ability, but largely attitude—and attitude is proven by the investment of time and effort.

The rise of ATS was originally to save effort on the employer side. Applicant Tracking Systems are now nearly standard, used by 98% of Fortune 500 companies. They structure received resumes, tag them, and sort them by keywords and criteria, sparing recruiters from manually reading every single one. When a job opening goes from receiving dozens to hundreds of applications, something has to help sort the pile.

But ATS also spawned a ghost story that persists to this day: “75% of resumes are automatically rejected by ATS before anyone sees them.”

That statistic is false, and its propagation history is telling. The source was a résumé optimization company called Preptel, which closed in August 2013 and never published any methodology to support the number. It survived later through media citations—Forbes cited Preptel in 2014, CIO cited Forbes in 2018, CNBC cited CIO in 2019—each assuming the previous link had verified it, with no one going back to ask how that 75% was calculated. When someone searched Google Scholar for “ATS” plus “rejection rate,” they found zero supporting studies.

The reality is far more mundane. In 2025, a survey of 25 recruiters across a dozen ATS platforms found that 92% said their systems do not automatically reject people based on résumé content—they only rank, not eliminate. What decides your fate isn’t the machine—it’s the person who only looks at the top 20 after the machine ranks them.

This truth matters because it shows that even before AI, the bottleneck of the résumé was never “can it pass the machine?” but “can it rank high enough in that pile to be seen?” A résumé carefully tailored for a role naturally matches keywords and ranks higher—cost signal and machine filtering were aligned in that era.

AI has twisted that alignment out of shape.

When Everyone Can Produce a Perfect Résumé in Ten Seconds

AI has turned what once required cost—keyword alignment, machine-readable formatting, professional phrasing—into something that can be mass-replicated at zero cost.

In 2025, over 1.2 million people used AI tools to job search. Over 580,000 used it to write or revise their résumés, and 430,000 used AI to generate entire résumés. Even more telling is another number: 770,000 people, about 64%, used AI specifically to check if their résumé could pass ATS. In other words, today’s job seekers’ biggest anxiety is no longer “Is my résumé good?” but “Will it even be seen?”

When generating a résumé that looks professional, hits all the right keywords, and passes machines becomes a ten-second task, the résumé can no longer distinguish the diligent from the mass-applicants. Everyone’s résumé looks professional. Everyone’s keywords match. When something is universal, it ceases to be a signal.

The consequences on the recruitment side have already become a disaster. Over the past year, application volume on LinkedIn has surged over 45%. One recruiter described the current state as “drinking through a fire hose.” She and her peers typically handle 15 to 20 positions simultaneously, with each résumé getting 30 seconds to 2 minutes of attention. Hundreds of applications per role is a massive workload just for the first pass.

In LinkedIn’s own survey, over one-fifth of U.S. HR professionals spend 3 to 5 hours daily reviewing applications, yet 70% of employers say fewer than half of all applications meet all the job’s requirements. The recruiter’s colleague put it more bluntly: out of hundreds of applications, truly suitable ones are usually “fewer than 5.”

Both sides are using AI, turning this into an arms race. Job seekers use AI submission tools for mass applications, which alone has boosted application volume by 25%; employers use AI for screening, with 24% of companies already using AI to run entire interview processes. One side is flooding the channel; the other is filtering frantically. In between, the thing called a résumé has had its signal-to-noise ratio crushed to nearly zero.

Moreover, AI submission tools are actively generating noise. The recruiter said these tools easily exaggerate or distort a candidate’s experience, resulting in “the wrong people being mass-submitted to the wrong jobs by AI.” The machine doesn’t know if you’re actually a fit—it only knows to click “apply” for you. So what recruiters receive isn’t just more applications—it’s more and worse.

This is the real situation facing résumé optimization today. You open a guide and follow it, optimizing keywords, passing ATS, making the language more professional—every step you take, hundreds of others are also doing with the same tools. You’re not setting yourself apart; you’re burying yourself deeper into the pile of résumés being flooded in by the fire hose.

Four Approaches, and Their Ceilings

The AI résumé optimization methods floating around essentially boil down to four. Look closely at each, and you’ll see they’re all solving a problem that no longer matters.

First, keyword optimization—feeding your résumé to the machine. This is the most mainstream approach, done by 64% of AI-using job seekers. It’s built on the decade-old fear mentioned earlier, and that 75% statistic is false. The reality is 92% of ATS systems don’t automatically reject—they only rank.

Sounds like good news, but it’s not. When 180 people apply for the same job and the recruiter only looks at the top 20, being ranked 150th is no different from being rejected. What keyword optimization can do is prevent you from ranking even lower due to formatting issues—that’s your entry ticket. 99.7% of recruiters do use filters; missing a hard skill keyword or a misaligned job title will get you filtered out. But after entry, you’re still at rank 150. Keywords solve “don’t get eliminated,” not “get selected”—and those two are worlds apart.

Second, fully automated mass submission—apply to as many as possible. The logic is simple: if visibility is a matter of probability, increase the denominator. The person who applied to over 700 jobs with zero replies followed this path.

Mass submission has two hidden costs. One, AI submission tools may exaggerate your experience to boost matches, landing you in roles you’re completely unsuited for, actually lowering your credibility as a candidate source. The other is the time window—to control application volume, many companies now post jobs for only two or three days. The recruiter said the old saying is now true: if you don’t get in during the first two days, you don’t get in at all. Mass submission sounds like casting a wide net, but much of your net lands in places that have already closed.

Third, letting AI write the entire résumé. The most convenient—and the most dangerous. Resume-Now surveyed 925 U.S. HR professionals, and 62% held negative views of AI-generated résumés, with a significant number rejecting applicants outright because of it.

The reason isn’t complicated. When hundreds of résumés are generated by the same models with similar prompts, they converge—same sentence structures, same “results-oriented” verbs, same safely bland tone. Recruiters who see hundreds daily can smell it. The “AI smell” has shifted from neutral to a demerit: it signals you didn’t spend time, and not spending time is exactly what recruiters are trying to filter out. The effort you tried to save is precisely the only signal a résumé can still convey.

Fourth, tailoring each résumé to each job description. This is the only approach that still works—and it’s the most expensive. Only 3.5% of people actually do this.

It’s expensive because AI can’t save you the effort here. Place your résumé and the specific job description side by side, line by line: which of my experiences can prove I have the skill they want? The point they emphasize most—have I made it prominent? AI can help you think here, identify gaps, suggest what to highlight, but it can’t make the judgment for you, because only you know the true weight of your experience. Recruiters are blunt: comparing an AI-written, AI-submitted résumé to someone who truly spent time studying how to align their résumé with the role, the latter is far more likely to get a call.

Of these four approaches, the first three all try to use AI to bypass “spending effort,” yet effort is precisely the only part of a résumé that still holds value. The fourth doesn’t bypass it—it invests effort where it counts. But the thing it requires isn’t a résumé-editing trick; it’s how seriously you take the role.

What Recruiters Are Actually Looking At

The world in guides and the world on a recruiter’s desk are no longer the same.

Guides tell you that if you optimize your résumé and pass ATS, opportunities will come. The recruiter’s actual practice is different. To avoid drowning in applications, many companies post jobs for only two or three days. By the time you’re polishing your third draft, that door may already be closed.

Worth hearing is another remark. When asked what they value most when picking résumés, the recruiter gave an answer she herself found “boring”: a clear, concise résumé that shows how your skills match the role. It sounds like fluff, but it precisely points to the employer’s reaction after signals lost their meaning—when all résumés are AI-flattened into uniformity, employers revert to things harder to fake.

The most obvious is employee referrals. The same recruiter said referrals will be the way forward. The logic is direct: a current employee willing to stake their reputation to recommend you is something AI can’t fabricate. It reintroduces cost and risk, and cost and risk are exactly the sources of signals. When the old signal of the résumé drowns in noise, the old method of referrals becomes more valuable.

There’s a colder reality hidden here. Machine screening doesn’t get gentler when résumés flood in—it gets cruder. Research by Harvard and Accenture long ago pointed out that mechanical criteria like degree requirements, employment gaps, and age directly bar millions who could otherwise do the job. Internal LinkedIn data shows that people over 55 are about 60% less likely to be contacted by recruiters than equally qualified younger candidates. These dimensions have little to do with your ability to do the job, but when everyone gets 30 seconds, employers rely even more on such crude filters.

Thus, an awkward misalignment emerges. Job seekers pour all their energy into résumé wording and keywords, believing that’s the battlefield; employers have long been deciding on another battlefield—they’re finding people in their networks, on referral lists, among those fewer than 5 applications that “obviously show effort.” Harvard also found that companies willing to hire from overlooked talent pools are 36% less likely to face talent shortages, and the hires perform better on six key metrics. Bypassing mechanical screening to find people is itself a better deal for employers.

The marginal return on résumé optimization is rapidly approaching zero, while the less-traveled path alongside it—being remembered by a real person, getting referred, speaking through your work—is yielding higher returns. The step guides obsess over is precisely the one employers rely on less and less.

Invest Effort Beyond the Résumé

All this isn’t to say you shouldn’t refine your résumé. You still need one—clean, clear, letting someone understand your abilities in thirty seconds. The ATS filters you must pass should be passed; getting filtered out for missing keywords is the most unfair way to die. These are the floor, not the ceiling.

What I’m really saying is: stop placing all leverage on wording. The degree you can polish your résumé with AI is the same degree hundreds of competitors can reach. This no longer helps you win—it only helps you not lose too badly. Redirect that extra time to three areas where returns haven’t yet been flattened by AI.

First, treat the résumé as proof of fit, not self-promotion. Mass-sending a résumé that brags about you is different from tailoring one to prove, point by point, “I can do what you need for this specific role.” The latter is done by only 3.5% of people—which is precisely why it still works. Place your résumé and that job description side by side, item by item: my experience can prove I have the skill they value most? You can let AI help you think here, but the final judgment must be yours. One résumé crafted this way outweighs a hundred mass submissions.

Second, invest the time saved into signals beyond the résumé. When the recruiter said referrals are the future, it wasn’t politeness. A current employee willing to stake their reputation on you is something AI can’t fake. Equally unfakeable are your public work, your reputation within a community, even a tiny bit of real connection you’ve built with the employer. These things all cost effort, require exposure, take time to grow—and it’s exactly this cost that makes them signals again in a résumé-flooded market.

Third, accept an uncomfortable fact: today, being seen matters more than being well-written. Being ranked 150th is the same as being rejected; a job posted for two days closes. No matter how beautiful your résumé is, if it doesn’t enter the range of being seen, it doesn’t exist. Job searching must shift from mass-submission to finding the right people—apply less, but each application goes to a place you truly want to work and where you can get a real person to see you.

Back to Dabulis. After sending 700 applications with no return, he stopped mass-submitting. Instead, he cultivated relationships on LinkedIn, reached out to recruiters directly, singled out companies he truly wanted to join, and only then did responses start coming.

He didn’t make his résumé better. What he did was step off the road where everyone uses AI to drown each other, and onto a road that still requires real effort—and where a real person can still see you.

The counterintuitive truth of job searching in the AI era is here. When something becomes universally achievable at zero cost, it can no longer help you stand out. Those 700 perfect résumés don’t prove you weren’t hardworking—they prove you directed your hard work in the wrong place.

Appendix: Sources Cited in This Article

Recruiter “fire hose” / 700 applications 0 replies / referrals are the future / 2-day window (CNBC, 2025-10-29): https://www.cnbc.com/2025/10/29/recruiters-are-drinking-through-a-fire-hose-of-job-applications-experts-say.html
1.2 million use AI / 64% use it for ATS / only 3.5% tailor per job (Kickresume, 2025): https://www.kickresume.com/en/press/ai-job-search-data/
ATS truth / “75% auto-rejected” is false stat / 92% don’t auto-reject / 99.7% use filters / Harvard (CoverSentry, 2026): https://www.coversentry.com/ats-statistics
LinkedIn application volume up 45% in a year (NYT, 2025-06-21): https://www.nytimes.com/2025/06/21/business/dealbook/ai-job-applications.html
62% of employers dislike AI-generated résumés (Resume-Now survey, 925 HR professionals, 2025): https://www.resume-now.com/job-resources/careers/ai-applicant-report
24% of companies use AI to run entire interview processes (The Interview Guys, 2025): https://blog.theinterviewguys.com/how-many-companies-are-using-ai-to-review-resumes/