If you have ever applied for a job, chances are you have encountered an applicant tracking system (ATS, aka talent management systems, FYI). After all, 90% of Fortune 500 companies use an ATS before putting a resume in the hands of a person. What I’m sure started as a well-intentioned tool for managing job seekers applying for open positions back in the 90s, these ubiquitous sentries have matured into somewhat of a Frankenstein’s monster.
At the most basic level, ATSes are CRM databases with candidates’ basic information, skills, and experiences. The information in these databases is sourced by candidates either directly filling out a form in some provider’s user interface, or by automated process. The data subscriber can then run queries to find matches among the records just like any other database. As ATSes became the norm in the HR space over the past quarter century, they’ve been expanded to include document parsers, custom fields, internal notes, pre-screening tests, and quality scores, to name a few, all with the intention of making the matching service more efficient for the subscriber.
Instead of achieving their goal, however, ATS creators made something much, much worse… an overly restrictive impediment to the candidate being seen and the subscriber finding the right fit that’s mutually beneficial. Either a poor understanding or inadequate user training (likely a mixture of both) has resulted in requirements including way too many keywords to the point that extremely qualified candidates are missed. And to exacerbate the problem, NONE of the major players include related keyword matching. So when Patti, the temp in HR, enters the list of requirements handed to her, none of the synonymous job titles or skills are considered to be matched. COME ON PATTI!
Disruptv solutions to the rescue!
My time as the combined principal product and technical architect was focused on engineering solutions for all these problems for the millennial/gen-z job seeker.
Mission one was learning the ins and outs of how applicant tracking systems work (or don’t work depending on your purview). From there, I invented the piece-de-resistance the “hybrid resume” which would tackle a lot of the issues ATSes pose, handling everything a candidate needs to get their resume into the hands of the decision makers. Without divulging too many trade secrets, the hybrid resume creation process is part software solution, part downloadable document.
The web-based SaaS product uses a combination of natural language processing (NLP) and deep learning to understand what’s really being asked of a candidate in a job description, verify your profile contains the correct set of skills by passing the requirements through an ontology library, and makes use of the correct keywords the HR department has put in place all in a beautifully designed document that can easily and correctly be parsed by ATSes. No tricks, no treats, just an honest representation of you that will get you seen.
“What is this magic,” you say? The AI connects the dots between everything in your profile—hard skills, soft skills, experience, aspirations, and core values—and makes sure the right keywords show up on your resume. How? When a job description is read by the system, it is read for comprehension of what’s actually being asked of a candidate. Those requirements are sent into our ontology where job titles have their standard success criteria identified, macroskills and personality traits are broken down into their subcomponents, and synonyms are calculated. Now that there’s a standard and complete array of keywords to look for, we can move into looking for matches.