AI does some amazing things for recruitment departments, like help manage high-volume applicant pools, predict recruitment outcomes and reduce bias in hiring — here’s how.
How automation and machine learning are revolutionising recruitment
Thanks to technology, such as LinkedIn and other career-building networks, recruiters can discover qualified candidates much easier than ever before. However, these sites have brought with them a daunting logistical challenge: with such a diverse pool of talent available, how can recruiters separate the wheat from the chaff and find that perfect hire quickly?
Automation and machine learning may just be the element that simplifies modern hiring. Like many industries, recruiters are increasingly turning to automation and machine learning to streamline processes and make their jobs easier. The technology is already reducing administrative work,(for example, video interviewing helped one of our clients, Hasting’s Direct, cut down their admin time from telephone interviews from 55 mins to 10 mins,) identify ideal candidates, eliminate hiring biases and collect relevant data for future recruitment.
So what do we mean by automation?
Simply put, automated recruiting takes a process that was once highly individualised, time-intensive and difficult, and uses new technology to quickly and effectively screen applicants whilst allowing recruiters to gain information about a much broader range of candidates.
At its heart, automation is about saving time and improving the candidate experience helping to fast forward right-fit candidates through the recruitment process,
It helps solve the key challenges posed in a Business2Community post: “Candidates are always on the web somewhere but how can I reach the best candidates and engage with them most effectively? How can I eliminate the candidates that are not good enough
How can I avoid the situations where you spend your time on face to face meetings with candidates who don’t match their resume?”
In certain areas of recruitment it’s not unusual for a recruiter to sift through hundreds of applications when hiring for a specific role. With the help of machine learning technology, recruitment is brought slap bang into the 21st century. Recruiters are able to screen candidates quickly and rule out any that clearly don’t meet the position’s pre-defined criteria.
LaunchPad’s Predict software is a prime example of this type of data-driven recruitment solution. Using both recruitment best practices and machine learning, Predict allows recruiters to quickly identify candidates from applicant pools that are most likely to be high-performing, and will even pick out hidden gems that might have otherwise been filtered out of your initial screening process. This not only improves the quality of hire, but saves recruiters time that can be better spent on later rounds of interviews and assessments.
We foresee that in the near future recruitment automation will get event more intelligent. Recruiters will be able to harness data and use tools to automatically send relevant messages to candidates based on their behaviours. Perhaps a candidate is responding to all your video communications. Then why not send most of your communications to that applicant as videos?
Predicting right fits - so how does it work?
With so much time and effort expended in the hiring process, recruiters can’t afford to place a poor fit. Machine learning tools are preventing this problem by helping recruiters predict a given candidate’s fit with an organisation.
In LaunchPad’s case, machine learning works in conjunction with video interviewing, another recent technology that has transformed the recruiting industry. Algorithms quickly perform extensive analysis of a candidate’s speaking patterns, micro gestures, eye movements, expressions and word choice, to correctly identify right-fit hires for the organisation, based on previous data that has been collected and analysed.
As Business2Community reports of machine learning technologies, “instead of relying on your gut feeling or candidate input only, you can predict performance by analysing past data to measure crucial skills.
But don't recruitment machine learning algorithms pick up on previous human bias?
Yes. If a machine algorithm is given data that is biased, it will base its decisions on these biases. Which doesn’t help anyone.
That’s why we strongly believe that machine learning should work in tandem with human decision-making to identify right fit hires, and predictive models should be designed using traditional practices such as job analysis, validity studies and tests for adverse impact.. Organisations need to identify and flag any inconsistencies and biases in the recruitment process so these can be addressed and rectified before being used in automated recruitment. Fortunately, machine learning and data-driven assessment are helping employers recognise their unconscious preferences.
Realistically, a face a face-to-face or telephone interview may not actually help employers “get to know” a candidate. In fact, research shows that interviews do nothing to predict a candidate’s success, and may even make matters worse by reinforcing the organisation’s hiring biases.
Changes need to be made to make recruitment processes more consistent and efficient, and automation and machine learning are integral to this.
As recruiters start to rely more on automated technologies, it's easy to assume that humans will play an increasingly minor role in the process. In fact, machine learning enhances human decision making in recruitment, but doesn’t have to replace it.
With so much time and effort expended in the hiring process, recruiters can’t afford to miss out on the perfect candidate or make a bad hire. Machine learning tools are preventing this problem. Surely this is a win-win for recruiters, a smoother more efficient process, combined with the confidence that the decisions are consistent and unbiased.
If you’re interested in how automation can be applied to improve the recruitment process, you can download our new guide for recruiters here.
(Main image credit: Steve Jurvetson/flickr)