88% of Employers Admit Their Filters Reject Qualified People

A woman applied for a tutoring job using her real date of birth and was immediately rejected. The next day, she applied again with the same details, except she listed a younger birth year. This time, she got an interview.

This wasn’t about a biased hiring manager. iTutorGroup set up its software to automatically reject women 55 and older, and men 60 and older. Over 200 qualified people in the US were rejected this way. The EEOC sued, and in 2023, the company settled for $365,000. It was the first settlement like this.

No one at the company actually read those applications. The decision was made by a rule.

I’ve explained before why the claim that “75% of resumes are rejected by robots” isn’t true, and I still believe that. The number comes from an old sales pitch, not real research. But just because that statistic is wrong doesn’t mean there aren’t real problems. Since I wrote about it, a lot has come to light.

A setting in the software

It’s easy to think iTutorGroup was just a rare bad actor. But their automation was standard. It was the age rule that broke the law.

In 2021, Harvard Business School and Accenture surveyed over 2,250 executives and 8,000 workers in the US, UK, and Germany for a study called Hidden Workers: Untapped Talent. Two findings are especially important for job seekers.

88% of employers said qualified, high-skilled candidates get rejected if they don’t match every requirement. For middle-skilled jobs, it was 94%. These are employers talking about their own systems.

About half of US employers used a filter that automatically rejects any application with a work gap longer than six months. It doesn’t flag these for review; it just rejects them. The filter doesn’t ask why the gap exists, so it removes people who had cancer, cared for a dying parent, or moved for a partner’s job, just like it does for someone who simply stopped looking.

Researchers estimated that 27 million people in the US are effectively shut out of hiring by filters like these.

Why everyone insists this does not happen

If you ask the software vendors, they’ll tell you this doesn’t happen. And they’re not lying.

Workday is currently defending a class action, Mobley v. Workday, brought by an applicant who says he applied for more than 100 jobs through employers using its platform and never once got an interview. He alleges that one rejection arrived at 1:50am, fifty-five minutes after he applied. In June 2026, a federal judge allowed the disparate-impact claims to proceed on the theory that Workday acted as an agent of the employers. The case is in discovery. Nothing has been proven. Roughly 14,000 people have opted in to the age-discrimination collective.

Workday's position is unambiguous. Its Chief Responsibility Officer has stated that Workday AI "does not make hiring decisions and is not designed to automatically reject candidates". Its legal defence is that customers configure the software, control their own hiring, and that Workday cannot make anybody's decisions for them.

The vendor doesn’t deny that applications are rejected automatically. Their point is that it’s the employer who sets the rules.

Both things are true. The employer sets a rule, and the filter removes anyone who doesn’t meet it. The recruiter then reviews the shortlist, carefully and in good faith, but never sees the people filtered out before the list was made. So when recruiters are asked if their system automatically rejects people, they say no, and that’s true from their perspective.

The real disagreement is about who is responsible for the rejection, and when it happens.

What this means when you apply

Often, you’re rejected because of a rule you didn’t meet, before anyone even reads your application. You’ll never know which rule it was.

You can’t see these rules, but you can make it less likely they’ll filter you out.

• Account for gaps in your own words. A one-line explanation on the resume leaves no hole for a filter to find, and it is the truth anyway.
• Use the employer's terminology where it means the same thing as yours. A knockout rule looks for the phrase in the posting, not your synonym for it.
• Put what they asked for where it can be found. Page one, under a standard heading, not on page two under one you invented.
• Answer the screening questions slowly. They are the one place where the software really does decide, and a mis-tick is a rejection nobody will ever review.

None of these steps are dishonest, and they’re not about tricking the software. They just help you get far enough that a real person will finally see your application.

Or let someone do the checking for you

At Shortlisted, this is what we do. Our system reviews hundreds of job listings every day, and a real person writes each application, using the employer’s own language. We explain any gaps and recheck the screening questions. If you want someone to review your resume first, you can get a free review at shortlisted.today/try-free.

This post is a sequel to Why Your CV Keeps Getting Rejected Before a Human Reads It.

---

Sources: EEOC, "iTutorGroup to Pay $365,000 to Settle EEOC Discriminatory Hiring Suit" (2023). Harvard Business School and Accenture, "Hidden Workers: Untapped Talent" (2021). Mobley v. Workday, N.D. Cal., order of 22 June 2026 (allegations, not findings). Workday's statement via ITPro. Allegations in live litigation are reported as allegations.