Request Demo

* = Required Field

Is A.I. Slowly Killing Direct Sourcing?

Sep 09 2018 - By Aleister Electrum

From childhood to adulthood, people are exposed to recruitment techniques based on varying degrees of subjectivity. On the playground in elementary school, some students are always the first recruits for a given team owing to their outward physical prowess. In high school, recruitment becomes slightly less objective as students are selected by their peers for positions such as homecoming court, or by coaches for sports teams, or various part time jobs based on referrals from others. We then hit an abrupt pivot in recruitment style when we consider college admissions; the criteria for acceptance to an elite university are not based on one’s objective academic performance, but instead they are evaluated on potential future success based on their character and extracurricular repertoire. This parallels with the ideal recruitment style of a company: a group of human recruiters choose from a large pool of eager applicants based on how the recruiters feel the applicants will contribute to the success of the entity. This illustrates the concept of direct sourcing, a concept that has fallen out of favor in the corporate world.

This timeline paints a somewhat counterintuitive picture for success at the highest levels of achievement. Instead of objective accomplishments mattering more and more as one gets older, it seems that the reverse is true. Warren Buffett describes the three important characteristics of a successful recruit as “integrity, intelligence, and energy” where the most important is integrity, something that can only be evaluated by a human. It may seem obvious that a good employee possesses these characteristics, but this is where we must return to the analogy of college admissions to understand why this is not a reality for many companies. For the class of 2021, Stanford received 47,451 applicants, admitted 2,071 and 1,706 enrolled. This means that the acceptance rate was 4.4%, but the enrollment rate was 82%, suggesting that the admissions officers did their job well by not wasting resources on students who did not really want to be at Stanford. On the contrary, at a large school such as Arizona State University (ASU), where 24,764 students applied and 20,431 were accepted, only 8,230 enrolled, a measly 40% enrollment rate to the 83% acceptance rate. These data suggest that ASU’s lack of selectivity disenchants potential star recruits from committing, since these recruits do not feel the same pride of being selected over their peers as students from Stanford do. This cycle is difficult to break out of without a significant change in the policy of the respective entities.

In the business world, this same cycle is evident. The current trend in recruitment is to increase reliance on recruitment metrics. When a lead recruiter walks into a meeting to discuss their progress on a specific opening, they are expected to have a specific number of potential recruits along with myriad metrics about their efforts. This mirrors the selection perspective of ASU, where excellent recruits are undifferentiated from the masses, since the focus is heavily based on metrics to determine success. This means that many excellent recruits in the corporate world, who know that they are qualified for the job, are less likely to be interested in an offered position if they are not made to feel special or chosen over others. In a perfect world, all companies would utilize the direct sourcing methods of Stanford, where personal interviews are conducted with each potential candidate to gain insight into the character of the individual, gathering data beyond the facts listed on a resume. Unfortunately, in the current corporate paradigm, companies simply do not have the time or money to invest in this process, thus many potential stars recruits are not given the personalized treatment which has been proven to facilitate a more successful outcome.

“If we leverage A.I. in the hiring process, who will we blame for bad hires and how will we implement corrective action?” - Aleister Electrum, Futurist and Author

In the interest of saving time and money, many companies have turned to other forms of recruiting, including automation technologies such as Arya, which promise “an unbiased data-driven experience that humans are incapable of creating on their own – leading organizations to find the candidates that are the best fit.” These so-called “A.I. models” of recruiting are often just an algorithm written to match incoming candidate resumes to a few baseline resumes put in place by the company. This does not represent true A.I., because in order for something to develop intelligence autonomously, it must be able to use an autonomous “feedback loop” as a means of learning. For example, hiring software such as Hiretual does not have a way to constantly improve its accuracy without user input, thus it should not be considered A.I. At the end, the human subjectivity of recruiting still provides the final decision for determining whether or not a person should be hired. We have already established a correlation with college recruitment, which empirically shows that human subjectivity is better than an “A.I. model” at finding candidates that are the best fit. This is because the workplace is where people interact with other people to produce meaningful results, not just a group of machines relaying messages without complex decision making.

But this does not mean that technology in general has no place in the recruiting space. On the contrary, it is a necessary evil meant to save time and money for companies by aggregating and sorting lists of candidates. The common approach to this technology is to find as many people that might fit the desired characteristics as possible; however, in the modern age, this is simply not good enough. Companies desire results and metrics, and while a product may say that there were 1,000 matches for a particular position and show a glamorous preview of each person’s credentials, this is not actually helping a recruiter do his or her job effectively, since these “A.I.” products do not tell you whether or not a person has any interest in working for you. Finding a person who might be interested a new opportunity and has the required skill set should be the job of the technology, and a person’s job should to determine if these people fulfill Warren Buffett’s 3 characteristics of a successful employee, and to use his or her subjective judgment to win these star recruits to your cause.

Did you know?


According to McKinsey, companies are more profitable when teams are ethnically diverse


Of candidates are passive job seekers according to Talentnow 2018 Recruitment Trends


Of surveyed HR leaders said employee engagement was a top priority in 2018 and beyond