Human Resources

Three key applications for machine learning in the recruitment industry

By Peter Linas, EVP Corporate Development & International at Bullhorn
Human Resources
Published: 25 June 2018

For all of the talk that artificial intelligence (AI) will one day render many different jobs obsolete, there will always be certain professions that can only be performed by a human.

Recruiters, as an example, don’t need to worry: their jobs require relationship building, consulting, and strategy development – and algorithms cannot do these things on their behalf. These elements of the job make recruitment so worthwhile and rewarding, and AI, in fact, helps recruiters to do more of it.

Machine learning is just one significant subset of AI. It involves programming machines or software to learn from data, identify patterns, and use information to make intelligent decisions. As a result, it is something that has significant potential to change recruiters’ jobs for the better.

Here are three key areas in which machine learning is already helping recruiters.

Candidate relationship management
One way that recruitment agencies are already using machine learning is by employing chatbots to automatically communicate with candidates. Machine learning helps the chatbots to learn the best responses to different candidate queries, and subsequently interact with them on the recruiter’s behalf. This is particularly useful because it frees up valuable recruiter time. More sophisticated software can even automate the scheduling of meetings and calls.

Chatbots naturally raise concerns about machines replacing human employees, but they are not designed to replace recruiters. Instead, they should help them save time and resources answering common questions and addressing common problems, leaving recruiters able to focus on more important tasks.

Candidate shortlisting
Another main task that machine learning is assisting recruiters with is shortlisting candidates for specific roles. By applying machine learning algorithms to candidate data (collected from a wide variety of sources and channels, including social media, employee history, and employer information), it’s possible to quickly draw up lists of the most qualified candidates based on desirable skills and experience.

This same process can also be applied to CV screening; a machine can do it, if it knows what criteria to look for. The recruiter is then responsible for taking over the interview process and beyond. Recruitment will still be human-centric – but recruiters will have more time to focus on the interpersonal aspects of their jobs.

In addition to speeding up the process, machine learning can help tackle another problem in recruitment – discrimination. Algorithms simply cannot have an implicit or unconscious bias against any gender, race, or orientation, unless they are explicitly programmed to have this bias. Therefore, they will only be capable of picking candidates based on their qualifications and suitability for each job. Whilst human bias is often subconscious, it can still result in discrimination. With machines, there is less risk of that happening.

Placement probability
Thanks to advances in machine learning, there is a good chance that you’ll be hearing more about the concept of ‘placement probability’ when discussing candidate selection. By assessing the data of a prospective candidate against previously successfully placed candidates, algorithms can determine how likely it is that a certain candidate will get the job, and assign them a score that reflects that.

The specific machine learning algorithms used to determine placement probability are similar to those used when shortlisting candidates, and make use of the same data – social media, employee history, and employer information – plus information on desirable skills and experience.

Recruitment companies are already proving that machine learning isn’t going to replace recruiters any time soon. There will doubtless be some growing pains, but if technology can let recruiters focus solely on building stronger relationships, streamlining important processes and, ultimately, improving the candidates experience, there’s no real reason to resist it. Machine learning can make their jobs easier and more enjoyable – if they’re willing to embrace change.

For more information visit