Data in the Driver’s Seat
At Johnson & Johnson, Sjoerd Gehring and his talent-acquisition team have to say “no” far, far more often than they say “yes.” Each year, the global healthcare and pharmaceutical company receives more than 1 million resumes for approximately 28,000 open positions. Given J&J’s status as a consumer brand, it’s crucial that job seekers feel well-treated by the company regardless of whether they make the cut.
“We want to provide a consumer-grade experience for job candidates,” says Gehring, the company’s global vice president for talent acquisition and employee experience.
That’s important not only for J&J’s brand reputation, but also because qualified candidates who get a “no” the first time around may end up hearing from the company again when a position that’s a better match for them opens up.
Gehring, like a growing number of talent-acquisition leaders, is keenly focused on making much better use of the data stored within his organization’s recruiting systems to find quality candidates faster in a challenging hiring environment. Typically, the information gleaned from resumes, candidate assessments and interviews has sat there unused, thanks to resource constraints and the clunky search functions in many applicant-tracking systems.
“The search capability in the ATS is usually very limited, and it’s a time-intensive process,” he says.
More recently, however, the rise of artificial intelligence has made it much easier to unlock that data. New AI-based tools are enabling HR to identify and reconnect with so-called silver- and bronze-medalist candidates—those who impressed recruiters and hiring managers with their experience and potential but, for whatever reason, weren’t a good fit for the job to which they applied. AI—in the form of chatbots that screen job seekers and keep them updated on the status of their applications—is also making it less burdensome for HR to connect with and evaluate candidates.
Technological advancements such as these can be a godsend to recruitment functions that are often overwhelmed by data.
“Recruitment leaders are living in a very chaotic, complex stakeholder environment,” says Ian Cook, head of workforce solutions at analytics firm Visier. “They have a ton of transactional data swimming around and they’re not sure how to organize it.”
“The Rejection Business”
AI is definitely having its HR moment. According to a recent IBM survey of 6,000 executives, 66 percent of CEOs believe cognitive computing can drive significant value in HR. Half of HR executives agree that cognitive computing “has the power to transform key dimensions of HR.” Deloitte Consulting has discovered that recruiting teams it identifies as “high-maturity talent functions” are much more likely to make use of AI and data analytics than their lower-performing counterparts.
Given that only one or two out of 100 candidates typically end up getting hired, “recruiters are in the rejection business, not the hiring business,” says Jobvite CEO Dan Finnigan.
In light of that, recruiters must find fairer and more efficient ways for determining who to reject—especially now, when a bad candidate experience can easily make the rounds and cast a negative spotlight on a company in an historically tight labor market. This is where Finnigan sees chatbots playing a vital role, as they can screen candidates quickly and keep them updated on their status while decreasing time-to-hire.
“The chatbots get smarter over time—they’re fed data from the ATS on who makes it to the finalist list and who ultimately gets hired, and look at that data relative to the answers candidates give to its questions,” he says. “If the chatbot’s effective, it’s going to convert more applications into interviews, decrease the amount of time it takes to convert applications to interviews and do so at a lower cost.”
Larry Nash, U.S. director of recruiting for EY, the consulting firm formerly known as Ernst & Young, says he’s making active use of AI and recruitment-process automation at the company, which expects to make 14,000 new hires in the U.S. alone this year.
“We see AI and automation as having great potential to free up our recruiters so they can spend more time identifying great prospects and providing a better experience for our candidates and hiring managers,” says Nash.
At EY, an internal organization called Automation Central works with each department in the company to review problems and analyze what can be made more efficient. The HR department currently has more than 20 chatbots in production across a variety of processes, including recruitment. Last year, EY piloted a program in which a chatbot named “Buddy” responded to onboarding-related and general questions from 600 new university hires. Nash and his team are also testing a candidate-assistant chatbot built on IBM’s Watson cognitive-intelligence platform that will interact with candidates during their job search, recommending open positions based on their resume or their responses to certain questions.
“It helps to enhance the candidate experience and get them answers to their questions faster,” says Nash. “I think that, as candidates experience these innovations, it also helps them see our brand as more cutting-edge.”
The expectation is that, over time, AI and robotic-process automation will enable recruiters to spend more time with prospects and less with administrative processes, ultimately leading to higher productivity per recruiter and decreased cost per hire, says Nash.
“The good news is that AI is going to slowly eliminate the ‘rinse and repeat’ element of recruiting and, instead, let recruiters focus on the thing they love, which is the matchmaking component of recruiting,” says Finnigan.
AI will also help HR determine where its most successful hires come from: referrals, talent pools, internal hires?
“We have a machine-learning algorithm that will predict how long it will take you to fill an open requisition based on your company’s historic performance in filling that job, on other companies’ historic performance in filling that type of job, the local market and how many applicants you have in your funnel,” says Finnigan. “You’ll start to see AI becoming more prescriptive in what you should be doing to get more qualified candidates into the funnel.”
Increasing Speed and Diversity
At Johnson & Johnson, Gehring and his team have a number of AI tools at their disposal to help them work through the tsunami of resumes they receive each year to find qualified and diverse hires. They’ve worked closely with Google on its Cloud Jobs API service, which uses machine learning for the purpose of creating a better match between job seekers and open positions. They use Textio, an AI-based solution that helps ensure J&J’s job descriptions aren’t gender-biased. And they’ve partnered with a company called HiredScore to make better use of the data residing within J&J’s systems.
Once a job requisition is created, a tool from HiredScore called Fetch sorts through data in the ATS, as well as J&J’s candidate-relationship management system and other databases, using a “two-sided algorithm” that understands when a candidate meets the requirements for the role and whether he or she would be interested in applying for the position. Once finished, it presents recruiters with a slate of potential candidates for the role. Recruiters can give feedback to the tool based on how good of a match the candidates were for the position in question. “That’s how it learns and gets smarter, especially if it’s used at massive scale,” says Gehring.
“Fetch is doing a great job of proactively presenting those candidates who were great silver or bronze medalists to our recruiters,” he says.
Textio, meanwhile, has enabled J&J to boost its total number of qualified female hires by 13 percent and increase the number of female hires for hard-to-fill positions by 9 percent, says Gehring.
“Even a relatively simple AI tool can have a very noticeable impact on our ability to attract more diverse talent into the organization,” he says.
Gehring takes a fairly unusual approach to finding new technology solutions. Rather than sit through vendor briefings or demos, he forms partnerships with established companies (such as Google)and promising start-ups to refine and enhance their products.
One of those partners, HiredScore founder Athena Karp, previously worked as an investment banker at Merrill Lynch. She decided to start her company after analyzing large HR vendors such as ADP and Paychex and noticing the vast volume of available, yet underutilized, HR data that was out there. She also was interested in finding a fairer and more effective way for women and minority candidates to break into fields such as tech.
HiredScore scans resumes as soon as they’re entered into an ATS and grades them with an A, B or C, based on the candidate’s qualifications. The intent, says Karp, is to help companies avoid the need to rely on keyword searches, which can be subject to unconscious bias.
“How many times are people using keywords such as ‘golf’ or ‘fraternity’ just to whittle down a vast pool of candidates?” she says. “I just knew the process could be fairer and more efficient.”
Bringing “Meaningful Disruption”
Gehring says he’s very choosy about partners.
“The conversation I always have with a start-up’s CEO is: ‘Tell me about your team, culture and values,’ ” he says. “The moment I start to feel that those things are not aligned with what we stand for, then we always pass, regardless of how good the product is.”
This choosiness extends to his process for determining where and how J&J will implement AI and automation in its recruitment processes.
“One of the biggest pitfalls that recruitment leaders face is that there’s so much new tech out there, it becomes like a sport to pick the shiniest object to add to your portfolio,” says Gehring. “What you end up with is a highly fragmented ecosystem of tools that delivers a very poor user experience to candidates and hiring managers.”
What Gehring and his team have done is to first envision what they wanted J&J’s candidate and hiring-manager experiences to be, then translate those experience maps into digital strategies, and decide what part of the ecosystem they wish to buy and what part to build.