By now you know that AI is...well...everywhere. No matter your sector, wherever you turn you’re seeing headlines about AI this and machine learning that. Recruiting is no exception, in fact, we’d say it’s probably one of the most impacted sectors of the business world. AI-powered software is already being used by a majority of major corporations for at least one step in their candidate journey.
For the sake of brevity, and not overwhelming you, our poor readers, with data, we’re going to focus on the sourcing stage. This first step in the recruiting process is ripe for disruption and frankly seems like low-hanging fruit for an AI-powered solution. After all, reports cite an average of 13 person-hours per week going into sourcing for just one single role. 13 hours. Per week. One role.
First Step for AI: The Job Description
Job descriptions are often the first taste of your brand voice and company culture that a candidate ever sees. This makes it crucial to get it right the first time. Unfortunately, and this holds especially true for the tech sector, many job descriptions are written with unintentionally bias wording. Research has shown that in male-dominated fields like software development it’s common to find words like “dominate, hierarchical, ninja, competitive” and so on. These are inherently masculine words that will turn many women off from even applying, as they see the role as aimed only at men.
On top of that, it’s been shown that while men will apply to a job when they meet anywhere above 60% of the requirements, while women typically feel they must meet closer to 100% before they’ll consider applying.
These factors, among others, are propagating the gender imbalance in many fields. AI text analysis tools can help not only remove the inherently biased but can also help even out the overall tone of a description. Diverse workforces have been proven to be more productive and return much better ROI.
These same text analysis tools can also reduce jargon too often found in job descriptions, which many qualified candidates may find off-putting, especially if they’re new to the workforce and are unfamiliar with the lingo. And with Generation Z just beginning their working careers, this can be the key to attracting the young up-and-comers you need in today’s job market.
Where to Post That Job Description? AI Can Help There
Increasingly, knowing where to post openings means being able to target passive candidates. These folks make up an estimated 70% of the candidate pool today and are those who are currently working in their field, yet would be open to changing jobs if the right offer came their way.
This means that you, as the recruiter, need to truly know your audience. Where they spend their free time online, and the techniques to get your opening seen by these folks, who may not be visiting old-school job boards. AI can step in here by parsing your database of previous applicants, both those you hired and those you passed on, looking at the sourcing information. Then it generates suggestions for where to post descriptions for similar openings.
In addition, predictive analytics is a growing segment of AI for recruitment and can examine large datasets containing your current employees’ information to help determine other places where similar audience members are likely to be found. Then automation tools step in and help get the description seen by as many of these potential candidates as possible.
When the Resumes Start Coming in, AI Steps Up
On a fundamental level, artificial intelligence excels at data processing and pattern recognition. Remember that 13 hours/week spent on sourcing? Imagine what that number would be if you didn’t have that resume scanner set up to parse all those resumes? And that technology is now 5-10 years out of date, from our experience. So now think about the hours you could be saving by having a trained AI scanner in place to do that initial parsing of resumes.
These solutions can be programmed to scan for keywords, keyword phrases, and even synonyms of common keywords (think AI vs artificial intelligence, for example). It can also take advantage of recent advancements in NLU/NLP (natural language understanding/natural language processing) to learn how to gauge intent. Now, your scanner can glean more than even a human reader about cultural fit, level of empathy, and more.
And the Final Step of Sourcing, Filtering Candidates
Millions of data points are nothing for modern AI. By running the pre-scanned resumes through further advanced analytics and filters you can get a more complete picture of each candidate before ever scheduling a phone or video call. To further aid your searches, point these same filters at your database of existing candidates who were not selected for previous openings and you can greatly expand your talent pool, with minimal effort.
What you’ll end up with is a list of pre-screened and pre-filtered candidates to investigate further. Spend some of your remaining 13 hours looking at social media profiles and LinkedIn pages before passing a further culled list of only the highest-quality candidates on to the hiring manager to select from.
AI-powered hiring reduces the bias that creeps into any human-powered process. Increasing diversity of hires is a key factor for many companies this year, and with good cause. Not only is it a selling point for the inclusive nature of company culture, but it’s better for the bottom line, as discussed above.
By programming your HR AI to avoid names, ages, and dates; and to focus instead on skills assessment results and industry certifications, you can reduce if not eliminate many such biases. One final cautionary note: should you discover bias creep, go back to the programming. because AI is first and foremost a software technology solution that is created and maintained by humans, human flaws will sometimes find their way in.