HR directors can’t only rely on the use of data to recruit. But companies have invested algorithms to help them make the first sort. If a robot is obviously not going to do an interview or negotiate the salary, the passage through big data or artificial intelligence can be a first step for recruitment.
Sorting CVs by hand is not enough to find good candidates. In the face of massive needs or when candidates are short on a job, companies need to find other methods. This is where the use of the data can be relevant.
The algorithm extracts skills
This is what Jobijoba, a French employment aggregator, offers with the CV catcher, a product launched in early 2017. The CV catcher is an application that integrates with the career site of a company or a recruitment agency.
The candidate imports his CV which is read by the algorithm.
The latter then proceeds to a data extraction: the coordinates and the name of the candidate, his career, his diplomas, and his soft skills. This data is then cross-referenced with the available offers and the candidate is directly offered ads that may correspond to him.
To extract the skills of the candidate, the algorithm is not based on those explicitly marked on the CV (software or programming languages mastered for example), it will also detect soft skills.
A possible detection because the algorithm will “focus on concepts, not words,” says Thomas Allaire, CEO of Jobijoba. For example, if a candidate indicates that they have held several leadership positions, they will be assigned leadership skills. “The algorithm will also take into account the semantic context,” says Thomas Allaire.
A standard resume
Like on Spotify
Another asset for Thomas Allaire, the CV catcher “allows candidates to find offers they did not think”. The CEO of Jobijoba suggests a parallel with Spotify. You can listen to different artists but be more fan of rock than rap. By studying all your history, Spotify understands it and introduces you to new rock bands. Jobijoba works the same way.
“The algorithm will look at all the skills and make a synthesis of all that,” summarizes Thomas Allaire. For example, if you worked in ESN as commercial and then as an HR assistant, the algorithm will understand that the guiding thread of your career is your knowledge of ESN and will offer you positions in this sector rather than a commercial position in the food industry. This sorting makes it possible to take into account atypical CVs.
Getting out the right candidates is also the niche of Yatedo Talent launched in April 2017. Yatedo Talent presents itself as the French “Google of recruitment” according to its CEO Amyne Berrada. It is a search engine based on open sources of the web.
When AI detects passionate candidates
This option will be able to search for candidates on specific skills or qualities. For example, the brick, defined by Yatedo’s CEO as a “filter for softs skills”, will detect “passionate” candidates. To do this, the AI will study the online CV of a candidate and relate the trades, tasks, and training. “The AI will see if the candidate has superior skills to others and if these skills were acquired outside the school curriculum,” details Amyne Berrada.
For example, if a person has done web development and only has a history degree, they will be considered “passionate” because they are self-taught. With Jobijoba and Yatedo Talent, data analysis goes much further than collecting candidates’ identities. That said, these tools will not replace a physical maintenance. Thomas Allaire believes that the CV catcher is mainly used to “save time for everyone”.
To the recruiter, making a “pre-selection” of the candidates by directing them towards the most relevant offers and the candidate, by examining for him the basis of the offers “in a split second”. A decisive advantage for companies with high recruitment needs. Yatedo Talent also saves time by making it easier to spot good candidates. And if the search engine “only deals with sourcing”, its use can, however “pre-interview”, since the recruiter already knows some of the qualities of the candidate adds Amyne Berrada. The latter concludes by indicating that the manager remains ultimately the sole decision maker: “The AI does not create data, it values them to help the recruiter in his decision.”
This translation is freely adapted from this article.
More about the MSc Digital Business Analytics at EMLV