Are you losing qualified candidates because of your resume screener?

Brian Lee Yung Rowe
AI Workplace
Published in
3 min readMay 20, 2019

An unintended consequence occurs when your solution for one problem produces a new (unexpected) problem. Resume screeners were introduced to help recruiters sift through massive volumes of resumes quickly. Most screening tools use keywords to select “qualified” candidates. The unintended consequence is that many qualified candidates are dropped because they didn’t know to optimize their resume.

Candidates that don’t optimize the resume are warned that they may slip through the cracks or be passed over for someone else that has optimized their resume. But are candidates being dropped because they are unqualified or because resume screeners are ineffective?

Not only do you have to revert to a more boring format you have to choose your keys words VERY carefully. It may recognize the words “web developer” and not “web develpment” and because I said I did “web maintenance” as well, when I had someone run it through an ATS, it said my experience was concentrated in Manufacturing & Equipment Repair and not Information & Technology. — Gabriella Joy

Unfortunately, resume screeners are pretty simplistic. Many just use keywords and exact phrases to match people. These keywords are dictated by the recruiter or hiring manager using the ATS. If insufficient time is spent configuring the resume screener, there’s a risk of garbage in, garbage out. But if you’re using the tool to save time, how much time will you invest to get good results? Even resume screeners that use AI suffer from problems. Amazon gave up using an in-house tool because it ended up reinforcing structural bias.

The problem is that resumes are inherently flawed, and resume screeners amplify those flaws. On one side we have recruiters using resume screeners to look for specific keywords or patterns. On the other side we have candidates stuffing keywords and devising other tricks to get to the top of the list. Instead of creating a system to find qualified candidates, all we’ve done is find candidates that know how to game the system! (To be fair, many people feel they need to game their resume just to get an interview.)

As a graphic designer i’ve learned to make 2 resumes. One resume is made to beat the ATS bots, and the other is a printed version that I can hand out at job fairs and interviews. — Roy Eric Reynolds

Over-reliance on resume screeners also creates a poor candidate experience. While some people are resigned to this reality, HR professionals should be thinking about how to improve the candidate experience to attract more candidates, not less. Why bother expanding the talent pipeline just to throw out the majority with a resume screener?

Instead of relying on resumes, we use a chatbot to conduct structured interviews. It’s hard to game a structured interview, and we get far more useful information about how somebody works and how well they can work with others. We only use the resume as supplemental information for an onsite interview. This ensures that we aren’t being subject to unconscious bias early in the hiring process. It’s so effective, we’ve cut down our time to hire from 4 weeks to 1 week.

Ultimately, we need to think about what we are trying to accomplish with a resume. A resume offers a ridiculously abridged view of a person. Is that sufficient to make a decision about whether someone is qualified for a position and a cultural fit? If your answer is yes, beware the unintended consequences that may spring up.

Knowledge is meant to be free! If you enjoyed this article, share it with your friends. Subscribe to the newsletter to stay ahead of the curve on how AI will transform the workplace.

Brian Lee Yung Rowe is founder and CEO of Pez.AI, creator of Just Interview Me, a chatbot that conducts structured interviews with candidates, saving hours of valuable time while removing sources of unconscious bias. Learn more at http://justinterview.me.

--

--

Brian Lee Yung Rowe
AI Workplace

Founder & CEO of Pez.AI // Making human interaction more meaningful with chatbots and data science