Skip to content

How artificial intelligence is changing the management of talent

Following in the footsteps of the film that told how a manager revolutionized baseball by applying analytics to the search for players, corporations are rethinking their methods to attract the best employees by appealing to big data.

Online games were easy until you reach challenge number six. The candidate was applying for a job at Unilever for which he had to complete different puzzles designed to test 90 cognitive and emotional features, from memory and planning speed to the power of concentration and his willingness to take risks.

More than a million job seekers have already gone through this type of test experience, developed by Pymetrics, a startup co-founded five years ago by Frida Polli. Neuroscientist trained at MIT with a Harvard MBA, Polli is at the forefront of new ways to assess talent for brands such as Burger King and Unilever, based on decades of neuroscience research that says they can predict common behaviour among high performers. “We warned that this combination of data and machine learning would be immensely powerful, leading to the recruitment of the paper-and-pencil procedure to the future,” explains Polli, sitting barefoot on a sofa in her Spartan office near the Flatiron district of New York, where about four dozen engineers, data scientists and industrial psychologists work in brilliant iMacs.

Pymetrics is part of a legion of startups that make noise by taking advantage of artificial intelligence, big data and other technological tools to shake up the hiring market. The research firm CB Insights predicts that risk investments in hiring technology startups will reach US $ 2.9 billion in 2018, which implies a 138% increase compared to the previous one.

What is driving all this investment? A January 2018 survey of more than 1,000 top executives found that attracting and retaining talent is their main concern, above the anxiety caused by the threat of a global recession, the trade war that drives Dondald Trump and even problems of competition. But Polli complains that human resources remain an “archaic system” that is based on “prejudiced” assessments of “irrelevant data such as CVs and cover letters.” Polli launches figures like a scientist with a dust cover in a film that anticipates a disaster that people ignore dangerously: recruiters study each CV on average just six seconds; Three-quarters of the candidates are eliminated in this phase, often arbitrarily, and new hires who pass the test fail in their positions between 30 and 50 percent of the time.

Companies, therefore, turn to new technologies to help make increasingly precise employment decisions, from hiring to productivity. Polli says: “It’s Moneyball for human resources”, in reference to the book by Michael Lewis – who was later taken to the movies with Brad Pitt as protagonist – about a Major League baseball team that revolutionized the sport by applying statistical analysis to Incorporate players who could compete against rivals with much more resources. The statistics-based approach to the film valued the races scored and the percentages of arrival based on traditional metrics used by the sport as batting averages and anecdotal evaluations such as how to move the bat.

The corporate world has been eager to adapt this model of team incorporation since then, but until recently it did not have the tools to fully evaluate employees. Now, as companies adopt more productivity software and business tools such as Slack and Workday, management has access to amounts of data on employee activity.

A few weeks ago around 70 programmers met for a hackathon in IBM’s Massachusetts lab. A handful of young engineers in front of a large conference room explain how they are trying to give voice to Watson, an AI-assisted talent advisor. The bot is designed to help employees navigate their careers at IBM, giving feedback via an app for everything from retraining opportunities to job improvements. Employees set specific goals for their careers in the app and Watson guides them through possible means of advancement, explains the required training and estimates the time it takes to get a promotion.

The team behind the project hopes that the speech will make the experience of chatting with Watson more accessible. His prototype is very rough and Watson is having problems with the local dialect. “Intense accents can be more difficult than a foreign language,” says cognitive software engineer Cameron MacArthur, who jokes that the system is “created to accept Spanish or Mandarin, but not the Boston accent.”

The Watson bot is currently a pilot project with 12,000 employees but will be available to the 366,000 workers in IBM’s workforce this year as part of the talent training and retraining company’s commitment, especially as the millennials – a generation that is twice as likely to leave a job as older colleagues – this year becomes the largest share of the US workforce. “Things have changed dramatically with the availability of big data and IA,” says the head of Human Resources at IBM. Diane Gherson. “There is a real shortage of capabilities in this new era.”

At IBM, they acknowledge that even a company that receives 2.5 million CV a year for tens of thousands of jobs faces a talent shortage, which Watson is helping them to recalibrate with the future in mind. At the end of the day, the company boasts more than 1500 blockchain employees, a field that did not even exist a decade ago. What better way to hire for these positions than to train talent internally? The company invests US $ 500 million annually in teaching and training employees. It has an internal development academy and is associated with Coursera and Udacity in a portfolio of online courses. Watson tracks the progress of employees in their training via a system of digital qualifications: IBM people have obtained 40,000 IA certifications for completing courses in areas such as architectural foundations and conversation services, for example. “We know what your capabilities are and we’ll tell you if it’s declining so you can make changes,” says Gherson. “If he’s a Java programmer and the demand for that ability is shrinking, Watson will say, ‘Here are these blockchain programmers, they did these courses and had these experiences and they’ve been promoted three times since then.'”

Gherson recognizes that these training efforts will make employees more attractive to competitors. It is one of the reasons why the company has also seriously invested in the analysis of the workforce. Through Watson executives and supervisors can also analyze employee performance and development – a “Fitbit (digital wristband) for bosses” as Gherson calls it – which automatically alerts them if a team member is not earning enough based on their fitness. Through algorithms, the system is reaching the point where IBM manages to anticipate the departure of employees. By comparing trends in the data of workers who have left with current employees, the company is able to identify patterns of one-way risk.

Last year, according to the company, this warning system resulted in approximately US $ 100 million in net savings based on what it would have cost to replace the lost talent.


Also published on Medium.

Published inStartups
%d bloggers like this: