Let The Right One In: Recruitment Made Easy With AI

“Instead of viewing machines as competitive interlopers, we must see them as partners and collaborators in creative problem solving as we move into the next era. The choice is ours.”
― T. H. Davenport,‎ J. Kirby. “Only Humans Need Apply: Winners and Losers in the Age of Smart Machines”

Let’s get one thing straight, all jobs are either: a) already semi-automated with software or b) will be in the next decade – even those that we consider to require a high level of emotional intelligence (EQ). Organizations living in denial of this verifiable fact are going to be left far behind.

Recruitment is an example of a very ‘human’ job. The average recruiter has to sift through dozens of CVs daily, interview several candidates on-site, and cold-call people they have sourced through numerous channels, such as social media and referrals. The reasoning goes that avoiding unqualified candidates, who only marginally fit the job description and the company culture, requires a certain level of EQ and profound personal touch. But, that line of logic is changing rapidly, with tools like data science leading the way. Parts of the HR workflow, like the tedious process of sifting through resumes, is just an analog form of data analysis. This, and many other pieces of the HR puzzle, are ripe for automating with machine learning technologies.

Machine learning is a discipline of computer science that gives software the ability to learn without being explicitly programmed. It can train itself over time to perform calculations with greater levels of precision and success. This can be applied to any stage of the recruitment process: requirement posting, resume search and review, interviewing, making offers, and onboarding — even though some are seemingly impervious to automation.

Considering the volume of data and the repetitive nature of the process, the DataRobot automated machine learning platform makes the recruiter’s job much easier. DataRobot captures the knowledge, experience, and best practices of the world’s leading data scientists, delivering unmatched levels of automation and ease-of-use for machine learning initiatives in HR every step of the way.

Stage 1 – Requirement posting

“The easiest way to obtain an updated version of a candidate’s CV is via social media.”
― Bernard Kelvin Clive, Brand Strategist @ BKC Consulting

In an increasingly digital world, the majority of candidates own at least two social media accounts. The notion of a digital community is particularly strong with software developers: LinkedIn, GitHub, Stackoverflow – the list goes on. Sourcing candidates from these platforms is essential for any respectable recruiter, who must then answer the following questions:

• How do you know which platform is the most effective for recruiting purposes?
• How do you choose between posting your job description or promoting it across social media on a limited ad budget?

DataRobot found the solution by helping a customer build and deploy an array of machine learning models, increasing the volume of qualified candidate applications by over 100 percent. Automated machine learning played a major part in helping recruiters choose which platform was more suitable for a given position, and which listings should be avoided altogether.

Stage 2 – CV review

“Sourcing and finding people is the most important. You can’t recruit, message, or network with someone you haven’t found.”
― Glen Cathey, SVP Talent Strategy and Innovation @ Kforce

Once your job description is live, you’ve reached the tedious and repetitive CV review phase. Handling a deluge of seemingly identical resumes is every recruiter’s living nightmare, but there is a way out: machine learning technology. For example, Recruit, a major Japanese group and the parent company of the job search platform Indeed.com, used DataRobot to develop a predictive model to screen new graduates from data they had gathered during the application process. The model turned out to be fantastic for screening low-potential candidates for specific requirements and reduced the need for human intervention in CV review by 67 percent, saving everyone time and money.

Stage 3 – Interviewing

“Interviewing someone is a very proactive process and requires taking a lot of agency into your own hands to get past people’s general normal self-preservation mode.”
― Brandon Stanton, Author, ‘Humans of New York’

Ah, the fun part. This is the seemingly un-automatable stage in the recruitment workflow that requires high amounts of EQ –  but an interview technology company, HireVue, disagrees. HireVue’s artificial intelligence (AI) identifies and analyzes the tone, word choice, body language, question context, and answers of candidates who have recorded video interviews to determine whether or not they’re a good fit — both culturally and professionally. This results in a quantifiable score they use to compare all the candidates, ensuring HireVue makes the optimal unbiased hiring decisions – all thanks to AI and machine learning.

Stage 4 – Offer & On-Boarding

“Loyalty is at the heart of what makes firms work.”
― Steven Levitt, Author, ‘Freakonomics’

After all that, you still have to consider the final process of closing the candidate. If you’re not quick enough, your candidate will walk, especially in highly competitive fields like software engineering and data science. But, what if you knew the probability of that happening?

For one customer, DataRobot does just that by using consolidated online data that indicates their candidate’s propensity to churn and use competitor services. This enables them to dedicate minimum marketing resources toward candidates who are less likely to bolt, while focusing career agent efforts on those who are.

Stage 5 – The Human Touch

“Some people call this artificial intelligence, but the reality is this technology will enhance us. So instead of artificial intelligence, I think we’ll augment our intelligence.”
― Ginni Rometti, CEO @ IBM

As you can see, machine learning and AI can optimize the candidate selection and placement process, helping HR find the right candidates for the job. The question now is, “Will AI replace the human in the Human Resources business?”

The answer is a resounding no. HR technology is built to supplement and augment HR processes, not eradicate them. With the cost of adoption plummeting rapidly and the data science community growing by the minute, the only thing a business leader needs to do is reach out and take the opportunity. Enterprises that turn a blind eye and refuse to adopt machine learning risk falling far behind. Let the smart machines take care of the routine, and automate as much as you can. HR are best at what they do when focused on the more important side of the job: human interaction.

About the Author:
Alex Pirlya is a Marketing Content Manager at DataRobot, where he focuses on business development activities and sales-related content. Until recently, he was a Senior Recruiter with over 5 years of experience at several multinational software companies. Alex enjoys market research, free-writing, and interviewing people for positions in software development. Out of work, he is an avid NBA fan and enjoys photography.