How Companies Are Recruiting In The Age Of A.I.
Pitney Bowes received good news and bad news. The good news was that, after a lot of careful research and planning, the global tech giant — widely recognised for its postage meters — recently unveiled a new e-commerce fulfilment centre capable of processing more than 40,000 parcels in an hour!
The bad news was that the business services company didn’t have a large enough workforce in place to utilise its brand new state-of-the-art processing facility to its fullest capacity. And it needed to recruit enough able-bodied workers as quickly as possible. So it did what any modern tech company with an eye on future global dominance usually does: it went all in on artificial intelligence.
A few days later, visitors to Pitney Bowes’ careers webpage were surprised to find a chatbot welcoming them and displaying all open positions close to their geographical location. The chatbot also smoothly guided them through a pre-selection process — which covered questions about their physical fitness and shift flexibility preferences — scheduled interviews when required, and sent appropriate email invitations.
The use of this AI-powered chatbot allowed Pitney Bowes to, in some cases, close positions 9 days earlier. At a time when the average time-to-fill a position in the U.S. hovers around 30–40 days, that means significant savings, both in time and money, for companies. Instead of utilising several temporary recruiters, HR departments are increasingly relying on AI to help them out in high-volume recruitment scenarios, and to find the most suitable recruits.
Attracted by the size of an inefficient job market and to fulfil organisational needs, investors are pouring money into a whole variety of start-ups aimed at disrupting the recruitment process. As is typical of disruption, these new ventures are digging into nearly every stage of the recruitment process to identify areas for performance and efficiency leaps.
Several efforts are currently underway to create machine learning algorithms that comb through online databases in search of suitable candidates; intelligent chatbots that are capable of undertaking a simple pre-screening process by posing standardised questions; and even digital profiling programs to assess a candidate’s body language using game theory and advanced psychology.
Yet another aspect of recruitment which has grown tremendously over the past few years is candidate relationship management, consisting of campaigns and communication material which help employers maintain a positive brand image in the minds of prospects and convert potential candidates into sure-fire applicants. Tools like this have taken the world of hiring by storm, and mainstream conversation has swiftly evolved from big data to machine learning to artificial intelligence.
The single biggest claim made by HR tech firms is that algorithm-driven tools can analyse thousands of candidates and zero in on the perfect one, all the while eliminating human subconscious bias from the process. But convincing evidence of technology being as biased as humans was provided by none other Amazon, the world’s largest e-commerce marketplace.
In 2014, the behemoth built an intelligent engine that sifted through applications submitted over the past decade to look for certain markers. The expectation was that it would easily identify good candidates in future applications simply by comparing those markers, positive and negative. Ironically, the engine turned out to be more human than machine when it started recommending only male candidates (who pretty much dominated Amazon’s workforce during that time period), and discriminated against female candidates. Eventually, the company abandoned its creation.
Past datasets, which are largely tainted by some form of human bias, are critical for the development of AI engines. So any future tech offspring are pretty likely to exhibit the same vulnerabilities. By utilising known discriminatory recruitment engines, companies end up making themselves more vulnerable to legal action.
Even automation tools, used for trawling databases and grabbing candidate information from various job boards, come with their own set of drawbacks. As platforms and tools attempt to distinguish talented individuals through keywords, job seekers are increasingly resorting to ‘keyword stuffing’ in an attempt to show up in most search queries. Nowadays, every candidate has a ‘good team player’ or ‘out-of-the-box thinking’ tag in their profile. This has resulted in a homogenised talent pool and an inability to differentiate good candidates from the not-so-good ones.
Technology has definitely made recruitment more efficient. But the bottom line is that people continue to prefer a personal touch, whether it comes from a recruiter or a boss. They want speed, someone to hear them out, and convenience, preferably from a fellow human being. Hence, what might really end up happening is less of technology replacing the human element, and more of a balance between the human and algorithm elements.