Sanofi turns to AI to cut Chinese recruitment churn

As the hunt for skilled employees in China intensifies, pharmaceutical group Sanofi is using artificial intelligence to identify the most promising and loyal applicants.

“We are facing a very fierce talent war, not just from multinationals but from a lot of Chinese companies that want to expand,” says Iris Zhou, head of talent acquisition for Asia Pacific.

Sanofi has been using an algorithm for the past two years to assess interns on their potential cultural fit with the organisation. Only then are the shortlisted candidates interviewed by HR staff.

The French company’s push into using AI for recruitment has developed at a time of growing debate on how far recruitment algorithms overcome bias. For example, Amazon dropped its experimental initiative to use AI to spot the best recruits after it became clear that the system it built unwittingly favoured men.

Yet Sanofi is optimistic about its nascent AI initiative.

Roger Morawski, who oversees Sanofi’s digital and artificial intelligence recruitment initiatives from its headquarters in Paris, says China was a logical place to experiment with such technology. He points to the country’s focus on digital innovation and the company’s need to address the cost of high staff turnover. “The market is crazy there, with 30 per cent of our sales staff leaving each year,” he says.

Sanofi wanted both to increase the number of junior sales representatives, and to expand its targeting focus beyond the most talented undergraduates pursuing medical and related degrees in China’s top universities.

In 2017, it launched a “golden medal” summer internship programme to bring in students to see if they were suitable to hire after graduating. But with a target of 800 hires, and 10 times as many candidates, that was an impossible task for just four staff in its human resources department reviewing and interviewing entirely face-to-face.

Sanofi prepared an algorithm based on characteristics it identified in its existing successful employees. By seeking those who shared its corporate culture, it hoped to reduce the numbers leaving early, often for rival employers.

While online recruitment programs often test technical and cognitive skills, Sanofi focused on screening for personality-based characteristics. It developed questions framed around five qualities: teamwork, courage, respect, integrity and willingness to keep learning.

“We wanted to know their culture and values — whether they would feel comfortable with Sanofi,” says Ms Zhou. “That way, we will know who is more suitable.”

One of the first steps was to use WeChat, the social media platform most popular with students in China, to promote Sanofi as an attractive employer. It emphasised that its priority was to hire future managers and leaders rather than highlighting the shorter-term appeal of pay or working conditions.

Those who expressed interest were immediately invited to apply on their mobile phone for an internship. They could either type responses over 20-30 minutes or speak them, with voice recognition software transcribing the words. Natural language processing interpreted their answers.

Initial market research, borne out by subsequent experience, suggested that candidates liked the quick response, whether it was an invitation for interview or a rapid rejection.

Amit Joshi, professor of digital marketing and strategy at IMD, the Swiss business school, who has researched AI, says its application in recruitment, especially in entry-level jobs, is one of the fastest areas of growth. This is most notable in markets such as China and India, where there is a strong digital culture, intense competition for recruits, and limited regulation. “From the perspective of the organisation, it can offer significant efficiency gains,” he says.

The value for job-seeking students and the long-term impact on employers remain open to question, however. Prof Joshi, who has not examined Sanofi’s system, cautions in general that there is “a bit of alchemy” in the underlying algorithms of AI systems, and that some may be based on “pseudoscience”. Most applications have been adopted too recently to be able to assess their impact on hiring, retention and productivity, he says.

At Sanofi, however, Ms Zhou maintains that the company is pleased with the results of its AI initiative. More than half of those it brought in as interns were subsequently hired — an improvement on the more typical third. Despite concerns that applicants might be able to game the system to provide the “correct” answers, she says, “there is no right or wrong answer, and even students who retook the test once [they were] better prepared and knew the questions scored the same”.

She says Sanofi tested its algorithm rigorously, and found that it beat the human judgment of interviewers in identifying the characteristics it sought. “We are very confident that students chosen this way outperform those hired normally.”

Mr Morawski says Sanofi has expanded its experiment to include an English-language version in India, and it is exploring other locations and languages, including French. He is also investigating potential regulatory hurdles, such as Europe’s GDPR (General Data Protection Regulation) rules, although he does not anticipate problems.

He points out that the company may continually redefine what it looks for in recruits, especially at a time when Sanofi’s new chief executive, Paul Hudson, is scrutinising the company’s corporate culture.

Meanwhile, its innovations include an AI chatbot in China that can help employees identify promotion opportunities. The technology may still be experimental, but it is spreading fast in China and beyond.

This article has been updated to indicate the percentage of sales staff in China who leave Sanofi each year