Selecting promising CVs, assessing candidate’s technical and non-technical skillsets, conducting online interviews – Artificial Intelligence systems are nowadays playing an ever-greater role in the company recruitment process. L’Atelier reports on the advantages and disadvantages of the changes taking place in this field.

Why robot recruiters are not replacing HR experts

For some time now, recruitment methods have no longer been restricted to the traditional CV submission-interview-job offer procedure. While this approach may still generally be the norm, CV sifting is now at least partly automated and takes place via a range of channels. LinkedIn has made recruitment support one of its key missions and Facebook has just recently launched into the fray with a ‘jobs’ tab on its company pages to enable firms and organisations to post their job vacancies and receive users’ applications. Meanwhile another tech giant announced in September that it had developed a specialised hiring tool: with IRIS by Watson, IBM is planning to use machine learning to identify key priorities and any particular difficulties involved in filling a given vacancy. Several days ago Google unveiled a Cloud Jobs API, which it has developed with a view to ‘improving the hiring process’ by matching up job seekers with open vacancies. So there seems to be a trend towards the use of Artificial Intelligence (AI) systems in this field.

Faced with new issues and new ways of working, the Human Resources profession is now adjusting to take these new technological tools on board. HR departments are learning to review the way they work so as to adapt to the changes taking place everywhere and the impact of the new technologies on recruitment is now undeniable. First and foremost, these new tools enable firms to directly address the Millennial generation, a group that already accounts for over a third of all working people in the United States. According to a report from the Pew Research Center, in 2030 75% of working people in the US will be these Generation Y-ers – young ultra-connected ‘digital natives’ from developing countries  who are into the on-demand economy and expect to receive information rather than having to go out and find it. In addition to technology, they have some extra expectations  about the world of work, which digital tools can go some way towards meeting.

Moreover, using an AI system as part of recruitment procedure can also make the process more efficient. There is some evidence that algorithms, the basis of AI, are more likely to take the right decision than human staff when it comes to new hirings. This is in fact the conclusion of a working paper published in late 2015 by US private non-profit research organisation the National Bureau of Economic Research, which points out that those polled for the survey who had been recruited by a process using an algorithm stayed 15% longer with their companies than those hired without any recourse to algorithms. Still, this observation should probably be imbibed with some care.

So, what exactly are these technologies, and in particular these AI systems, which help improve the candidate experience and make it easier for the HR department to do its job? What added value do they bring? A number of US startups are now offering online platforms, software programmes, chatbots and recruitment questionnaires to help both parties save time and improve the process.

AI systems helping to overcome prejudice

One of the positive attributes of Artificial Intelligence systems is the fact that they are not prey to emotions or in thrall to received ideas. So, when it comes to recruitment, AI can help to make the system fairer. Unfair discrimination during recruitment drives still exists, both in Europe and the United States and it is frequently blamed on what is often unconscious prejudice on the part of recruiters, who tend to take on people with whom they have something in common.

According to the 8th ‘Barometer’ survey (2015) carried out under the auspices of the International Labour Organisation, looking into unfair discrimination at work, 85% of those looking for work believe that unfair discrimination in the recruitment process is a frequent occurrence in France. In the United States too, both racial and sexist prejudices persist, as Harvard researcher Sendhil Mullainathan, author of a famous report on the subject, has pointed out. Tech companies are especially well-known for not being representative of population diversity, even though the advantages of background diversity in companies have been well documented.

Robot Mya from millennial job search company FirstJob is one of the Artificial Intelligence systems designed to help recruiters avoid unwarranted bias. This chatbot talks to the job applicant, answers his/her questions and then puts a quantitative value on his/her reasons for wanting the job and on his/her skillsets. Mya helps the applicant to understand the CV selection phase better and also calculates how likely s/he is to accept the post if offered, based on information received during the conversation. The aim is for Mya to automate 75% of the recruitment process while at the same time guarding against any unconscious unfair discrimination.

No-one doubts the usefulness of such systems and they are certainly likely to save HR departments a great deal of time. However, while AI systems are reputed to be objective, nothing prevents them being designed to pick up on certain key words or to exclude a certain type of job applicant, unless the systems themselves learn and use the company’s best practice in this field. With this proviso in mind, the US National Science and Technology Council encourages transparency in an October 2016 report entitled ‘Preparing for the Future of Artificial Intelligence’.

Les AI accélèrent le processus de recrutement

Artificial Intelligence systems speed up the recruitment process. Copyright: alphaspirit

Technologies speeding up the recruitment process

Some tools on the market are intended to modernise the recruitment process without disrupting it. Job applicants are assessed on the same criteria and go through the same steps as they did twenty years ago, but with the support of new technology. Utah, US-based startup HireVue has for instance developed a service for conducting interviews online. Applicants are filmed and answer questions on the HireVue platform. Then HireVue will analyse the responses, focusing on the applicant’s choice of words, level of stress in his/her voice, micro-expressions, how fast s/he speaks, and so on, in order to determine the best candidates, who will then go on to take some additional tests relevant to the specific business field.

On its website, the fledgling company claims to have reduced recruitment time from six weeks to five days for an international company in the hotel business. Saving HR departments time is also one of the basic goals of San Francisco-based Talentoday, Marketing Manager Allie Powers told us. The startup has built an algorithm which can analyse the results of a test, “using psychology to assess the applicant’s personality traits and what motivates him/her to work.”

Companies hiring staff can certainly save money by speeding up the recruitment process but it is of course even more vital to find the right candidate for the job. Explains Allie Powers: “One of the goals of Talentoday is to help companies save time and money over the long term, because it’s a big headache having your employees quit. When an employee leaves the company, it costs around nine months’ salary to hire someone new. Millennials change jobs much more frequently than their parents did. So, HR departments need to find ways of retaining them.”  Talentoday offers one way – “a better understanding of who your recruits and applicants actually are, plus their reasons for wanting a given job, and whether they’re motivated by money, philanthropy or the need for social recognition.”  An employer who is able to take an employee’s motivation into account will have a better chance of keeping him/her at the company. “If the employee places a higher value on free time than the salary, he or she might rather have one day of telework per week than a salary rise.” It is of course crucial for the employer to know such things. This is even more important for a startup, whose survival depends on the funds it has managed to raise. A startup has less room for error, and “needs to take on the right person first time around,” points out Allie Powers.

Using AI to assess emotional intelligence

What is special about Talentoday is that its system is able to assess non-technical skills -what are known as ‘soft skills’. Any job applicant, a student or existing company employee who wishes to find out more about his/her own profile and strengths can take an online test developed by a psychologist and a data scientist. This psychometric test takes fifteen minutes to complete: it asks a series of questions, each of which has two possible but completely different answers.  

Extract du questionnaire Talentoday

Screenshot of a Talentoday test question. Source: Talentoday

“In addition to personality and motivation, we also offer predictive analysis of the way candidates will behave in certain work situations” on the basis of their test results, reveals Allie Powers, explaining: “Recruiters can go on to a dedicated portal which enables them to obtain a detailed analysis of the results – a ten-page report – and compare the personality traits of several applicants with the characteristics required for the job.” The tool enables you to superimpose the applicants’ profiles on the standard profile. “For example, if you’re looking for a salesperson, you can base your search on the characteristics of your top-performing salesperson and see if any of the candidates match up, or you can base it on the characteristics Talentoday recommends for this type of job.” This makes for an objective way of assessing relationship skills “because during an interview, people always try to show their best side, but the test reflects what they’re really like most of the time,” she underlines.

The soft skills analysis generally takes place once an applicant’s technical skills have been assessed. And the technical field is where another US startup, Pomato, comes into its own. Pomato helps recruiters and HR departments to check up on the technical skills to which the applicants’ CVs lay claim, and which IT recruiters do not always themselves know enough about. The Pomato software carries out over 200,000 calculations on a candidate’s CV in just a few seconds and makes the results intelligible to a non-technical person. The software extracts the relevant data to create easy-to-understand graphics and visuals and can thus highlight the candidate’s strengths and weaknesses for the HR department. Based on these findings, the tool will then draw up the basis for a tailored technical interview. This type of technology is clearly a highly useful recruitment tool, but it cannot entirely replace the human factor. The last word still rests with a real person.


Le dernier mot revient au recruteur humain

A human recruiter possesses intuition which AI systems lack. Copyright: Infinityy

Human beings still provide the crucial link

Artificial Intelligence cannot do everything. “This is a highly useful technology which is in the spotlight and very much in fashion these days,” says Allie Powers, stressing: “But at Talentoday we believe that data and the knowledge it brings can help the recruiter to trust his/her own instincts. We never push candidates on to our clients, we produce information so that HR departments can be more confident about the decisions they take. The HR person is the professional.” She also believes that it could be a mistake to rely too much on data. “Our data is becoming smarter than we are but I think that as human beings we have special abilities that make us capable of feeling and judging a situation, and we mustn’t lose sight of this,” she argues.

In 2015, according to the Washington DC-based Pew Research Center, at least 28% of US citizens used their smartphone to help them find a job, and half of these ‘smartphone job seekers’ used their smartphone to fill out a job application. The recruitment process of the future is already with us today, and it makes use of the latest technology.

Meanwhile Artificial Intelligence systems might well be about to extend their reach into a field that is closely related to recruitment, namely well-being at work. Lucid, a startup whose software is currently in beta version, is going this route.

By Sophia Qadiri
Managing Editor & Journalist