By Reginal Campbell · August 4, 2026 · 11-minute read

How We Broke Hiring and How We Can Rebuild It

Employers cannot find qualified people. Qualified candidates cannot get seen. Both problems come from the same system.

Employers tell me they cannot find qualified people, while candidates tell me they cannot get a conversation. I believe both of them.

I have sat on both sides of the hiring table. I have hired people, evaluated talent, managed constrained teams, and worked inside organizations where decisions had to be standardized, justified, and measured. I understand why those controls exist. When leaders are accountable for delivery but have limited time and limited hiring support, structure is not cruelty. It is survival.

I also know what it looks like when fifteen years of experience disappears into an application portal and nothing comes back.

Both realities can be true at once. Employers may be overwhelmed by applications, and qualified candidates may still be nearly impossible to see.

An analysis by hiring software company Ashby examined more than 100 million applications submitted through its platform. It found that applications per hire had tripled since 2021, reaching more than 300 for an average role. Over the same period, candidates became roughly half as likely to reach an interview.^[1]

Both sides are putting more effort into hiring and getting less human contact in return.

There may be genuine talent gaps in some industries and specialties. But we have also built a hiring system that obscures qualified people, overwhelms the people responsible for evaluating them, and rewards behavior that makes the underlying problem worse.

The people are not broken. The system is.

How Hiring Became a Volume Game

The current hiring system did not emerge from a single bad decision. It developed one reasonable response at a time.

Applying became easier. As the odds of any one application producing a response declined, candidates applied to more jobs. Employers then received more applications than recruiters or hiring managers could realistically review, so they introduced automated screening, ranking, and standardized filters. Candidates concluded that no person would see their materials unless they first satisfied those systems, so they began optimizing and automating their applications as well.

Every step in that sequence is understandable on its own. Together, they have created a process in which both sides work harder, spend more, and trust each other less.

When everyone optimizes for volume, nobody creates signal.

Mass applying is not an effective job-search strategy, and I would not advise anyone to build a search around it. But dismissing it as laziness misunderstands what is happening. Candidates apply broadly because the odds are poor, the process is opaque, and almost no one tells them why they were passed over. The system created those incentives, and candidates are responding to them.

Employers deserve the same fairness. A recruiter managing 300 applications for one position is not lazy for relying on a ranked list. They are trying to make an impossible workload manageable. Anyone who has led a team through a hiring freeze, received approval for a backfill weeks too late, and still faced an unchanged delivery date knows how quickly shortcuts become necessities.

Neither side has to be acting in bad faith for the system to fail. That is what makes the problem difficult to solve. There is no single villain to remove.

Authenticity Became Another Performance

The contradiction is especially visible in the way employers talk about authenticity.

Organizations say they want candidates to be genuine, but reward optimization. They say skills matter, yet continue to favor familiar degrees, recognizable employers, and conventional job titles. They complain that candidates sound rehearsed while asking the same predictable interview questions as their competitors.

They also say people are their greatest asset, then allow finalists to disappear into silence. Greenhouse surveyed 2,500 workers in the United States, the United Kingdom, and Germany and found that 61 percent had been ghosted after an interview.^[2]

Organizations measure recruiters on speed and volume, then act surprised when the process delivers speed and volume.

Skills-based hiring offers one of the clearest examples of the gap between public commitments and actual decisions. Harvard Business School and the Burning Glass Institute tracked 11,300 roles at large employers before and after companies publicly removed degree requirements. Fewer than one additional hire in every 700 went to a candidate without a degree.^[3]

The announcements changed. The press coverage changed. The hiring behavior barely moved.

The real requirement was not confined to the job posting. It remained in the hiring manager’s instincts, the interview panel’s comfort zone, and the shortlist that continued to favor familiar profiles. Changing the language in a posting is not the same as changing how managers make decisions.

That instinct is understandable. A hiring manager who selects a conventional candidate and gets the decision wrong has made a defensible mistake. A manager who chooses someone with an unconventional background and gets it wrong may be forced to defend the decision personally. Most performance systems do little to reward that kind of calculated risk.

As a result, many organizations retain the language of skills-based hiring while continuing to practice credential-based hiring. They then describe the outcome as a talent shortage.

The mythology surrounding applicant tracking systems adds another layer of confusion. Millions of job seekers have been told that these systems automatically reject 75 percent of resumes. I could not locate a published study behind that claim; it appears to trace back to a 2012 sales pitch from a resume-optimization company that ceased operating the following year.^[4]

Applicant tracking systems primarily store, organize, search, filter, and rank applications. In some respects, that reality is more troubling than the myth. A qualified person may not have been formally rejected by a machine. They may simply never have been seen by a decision-maker.

Candidates are therefore being coached to defeat a machine that does not work exactly as they were told, while the decisive moment may be a rushed human review of a ranked list.

The Business of Hiring Fear

Artificial intelligence has intensified these concerns, but it has also created a market for fear.

Some of the loudest warnings about AI-generated applications come from organizations that sell staffing, screening, or fraud-detection services. Robert Half surveyed 2,000 hiring managers and reported that 67 percent of HR leaders said AI-generated applications had slowed their hiring processes.^[5] Robert Half also sells staffing services, and the same announcement argues that employers increasingly need that support.

Huntress reported that 23.2 percent of applicants to its own job postings were flagged as possible fraud risks through the detection product it sells.^[6]

Vendor research should not be dismissed simply because the vendor has a commercial interest. But those incentives should be made visible, and the findings should be described precisely.

A risk flag is not confirmed fraud. A mismatched phone number can trigger one, as can other inconsistencies that may have legitimate explanations. A forecast is also not the same as a measurement. Gartner has projected that one in four candidate profiles worldwide could be fake by 2028, while its survey of 3,000 candidates found that 6 percent admitted to interview fraud.^[7] Only one of those numbers came from counting reported behavior.

A less sensational finding receives far less attention.

A randomized study involving 480,948 job seekers examined what happened when applicants received writing assistance for their resumes. Hiring increased by approximately 8 percent, employer satisfaction did not decline, and some of the greatest gains went to people who were not writing in their first language.^[8]

The study examined help describing real experience more clearly. It did not test a machine inventing qualifications, titles, or accomplishments, and it does not prove that every use of AI improves hiring. It does, however, demonstrate that clearer communication of real experience can benefit both candidates and employers.

Using AI to clarify the truth is assistance. Using it to invent the truth is fraud.

The central questions are not especially complicated: Is the information true? Can the candidate defend it? Does it reflect work they actually performed?

What Employers Can Change

The most useful hiring reforms do not require another platform. They require people with authority to change what their organizations have learned to tolerate.

Shorten the application

Appcast found that applications requiring one to five minutes were completed at roughly four times the rate of applications lasting longer than fifteen minutes.^[9] Employers should remove unnecessary account creation, stop asking candidates to retype information already included in an uploaded resume, and move substantial assessments until after an initial conversation.

Leaders should also complete their own application process on a phone before approving it. That simple test would expose many of the frustrations candidates are expected to accept as normal.

A shorter application may increase volume, but inconvenience is a poor substitute for evaluation. Organizations should manage volume by assessing relevant evidence, not by making the process difficult enough that qualified people abandon it.

Publish a meaningful salary range

A 2025 National Bureau of Economic Research study found that pay-disclosure laws increased wages without reducing hiring or the number of job postings in the markets studied.^[10]

Employers should publish a range they are prepared to honor and explain what determines where a candidate falls within it. A range so broad that it could describe several different positions is not transparency. It is compliance without useful information.

Evaluate the work

Candidates should be asked the same core questions and assessed against criteria established before interviews begin. When appropriate, employers should include a short exercise that resembles the actual work.

Structure is not unnecessary bureaucracy. It is how organizations compare candidates using evidence rather than familiarity, chemistry, or whoever most resembles the person who previously held the role.

Work samples also provide a stronger signal than generic interview performance. A well-designed exercise can reveal how a person frames a problem, handles ambiguity, communicates tradeoffs, and reaches a decision. It should be limited in scope and should never be used to obtain unpaid consulting work.

Communicate like a human

Candidates should know the stages of the process, the expected timeline, and whether automated tools are being used to evaluate them. After an interview, they should receive a clear answer, even when that answer is no.

Someone should also be accountable for closing the process. When everyone assumes another person will communicate the decision, silence becomes the default.

That silence may feel operationally efficient, but it carries a reputational cost. Candidates may also be customers, professional peers, referral sources, and future applicants. The way an organization rejects people becomes part of its employment brand.

Measure hiring quality, not just hiring speed

Time to fill tells leaders how quickly the process ended. It does not tell them whether the right person was hired, whether that person remained with the organization, or how many qualified candidates abandoned the process.

Organizations should track whether new hires remain after eighteen months, whether their managers would hire them again, how quickly they become effective, and how many qualified candidates withdraw before a final decision. Those measures connect hiring to business performance instead of treating it as an administrative transaction.

What Candidates Can Control

Candidates cannot redesign an employer’s hiring process, but they can make better decisions about how they participate in it.

Use AI for clarity, not invention

Candidates can compare a job posting with their actual experience, identify claims they cannot yet support, sharpen the language used to describe their work, research the organization, and rehearse their answers aloud.

They should not invent titles, metrics, responsibilities, or accomplishments. The moment an invented claim is tested, credibility becomes the issue instead of qualification.

AI should help a candidate express the truth more clearly, not manufacture a more convenient version of it.

Understand the business

Candidates should know how the organization makes money, who its competitors are, what pressures it is facing, and why the role may exist now. That preparation makes it possible to discuss the work itself rather than merely perform enthusiasm for the opportunity.

This is especially important for experienced candidates. Senior roles are rarely filled solely on the basis of functional competence. Employers are also evaluating whether someone understands the operating context in which that competence must be applied.

Bring evidence

A case study, project summary, decision memo, sanitized plan, or working framework can show how a candidate thinks and what changed because they were involved.

The evidence does not need to be elaborate. It needs to be credible, relevant, and safe to share. A concise artifact often communicates more than another paragraph of adjectives on a resume.

When polish becomes cheap, proof becomes valuable.

Create human signal

Candidates should apply efficiently to positions that genuinely fit, then select a smaller number of opportunities for deeper effort. That effort might include researching the organization, contacting a specific person with a relevant question, or sharing a tailored example of work.

A generic request for a referral is not differentiation. Useful outreach demonstrates that the candidate understands something about the organization, the role, or the problem being solved.

Human signal cannot guarantee a conversation, but it can give someone a reason to look beyond the ranked list.

Rebuilding the System

It is important to remain fair to the people working inside this system.

Recruiters are handling application volumes their organizations never staffed them to manage. Hiring managers are expected to fill positions while continuing to carry their existing responsibilities. Candidates are told to personalize every application while receiving little information in return.

Most of those people cannot redesign the process from where they sit.

Leaders can.

Leaders decide what gets measured, which systems are purchased, whether an unanswered finalist is acceptable, and whether hiring is treated as an administrative chore or a business-critical decision about who will execute the strategy.

Too many organizations still treat hiring as paperwork, then wonder why execution suffers months or years later. The connection should be obvious. Hiring determines who will exercise judgment, manage risk, build relationships, solve problems, and deliver the work.

Most of these reforms do not require a major technology investment. They require clearer accountability, better management discipline, and a willingness to change what the organization measures. That is precisely why so many of them remain unimplemented.

Hiring is not an administrative process sitting outside the work. It is how an organization decides who will carry the work forward.

We built this system. We can rebuild it.

Sources

Ashby, application and hiring analysis, May 2026. The analysis covered more than 100 million applications on Ashby’s platform. Because the platform skews toward technology employers, the results describe that population rather than the entire labor market. ↑
Greenhouse, 2024 State of Job Hunting. Survey of 2,500 workers in the United States, the United Kingdom, and Germany. ↑
Harvard Business School Project on Managing the Future of Work and the Burning Glass Institute, Skills-Based Hiring: The Long Road from Pronouncements to Practice, 2024. The study examined 11,300 roles at large U.S. employers. ↑
I could not locate a published study supporting the widely repeated claim that applicant tracking systems automatically reject 75 percent of resumes. Career consultant Christine Assaf traced the figure to a 2012 sales claim from Preptel, a resume-optimization vendor that ceased operating in 2013, in “Your Job Application Was Rejected by a Human, Not a Computer”, 2020. ↑
Robert Half, survey on AI-generated applications and hiring, March 2026. The survey included 2,000 hiring managers. ↑
Huntress, The Growing Reality of AI-Enhanced Candidate Fraud, 2026. The reported figures reflect risk flags generated through the vendor’s own detection product. A flag is not the same as confirmed fraud. ↑
Gartner, candidate trust and fraud research, July 2025. The 6 percent interview-fraud figure came from a survey of 3,000 candidates. The projection that one in four candidate profiles could be fake by 2028 is a forecast without published methodology in the cited announcement. ↑
Emma van Inwegen, Zanele Munyikwa, and John J. Horton, “Algorithmic Writing Assistance on Jobseekers’ Resumes Increases Hires”, Management Science, 2025. The randomized controlled trial involved 480,948 job seekers and examined writing assistance applied to real experience rather than generative fabrication. ↑
Appcast, 10th Annual Recruitment Marketing Benchmark Report, February 2026. Appcast is a recruitment advertising vendor. ↑
Arnold, Quach, and Taska, “The Impact of Pay Transparency in Job Postings on the Labor Market”, NBER Working Paper 34480, November 2025. The study reported wage gains of 1.3 percent to 3.6 percent with no measured reduction in employment or job postings. ↑

About the author

Reginal Campbell writes about enterprise technology, AI governance, leadership, and the systems organizations build to make consequential decisions.