By Josh Bersin
One of the most important things companies do is hire people, and it’s still a bit of a black art. Most companies look at candidates job history, they call references, they give them tests, and they bring them in for interviews. And despite all this effort, HR leaders tell me they still make mistakes as much as 25% of the time.
Why? Today the criteria for success is driven by cognitive abilities, cultural alignment, and fit between a job seeker’s ambitions and those of the company. The research we completed earlier this year at Bersin by Deloitte found that the highest-performing companies are 90% more likely to use these “non-resume” factors in selection, more proof how complex hiring has become.
And this trend is getting stronger. Most research on the future of work shows a steady increase in demand for personal communication skills, creative problem-solving skills, and what always called learning agility – a candidate’s ability and willingness to learn. These are all things not visible in your college pedigree, GPA, or even job history. (Watch my talk on the Future of Work for more on this topic.)
It turns out AI is a perfect fit for this problem. Vendors are now successfully applying intelligent algorithms to create tests, simulations, and even analyze video job interviews to make hiring more effective. And some of these assessments have the potential to upset some big forces in the market.
Here are some examples.
The fast-growing company Pymetrics, founded by Frida Polli, a neuroscientist from Harvard and MIT, has developed a series of cognitive and neurological tests that are fun and easy to take, but have direct statistical correlation in roles like engineering, sales, and customer service. The technology can assess up to 90 different traits through these tests, and virtually eliminates bias and discrimination in hiring. Companies like Unilever, Hyatt, Accenture, and Tesla swear by the system, and they and others have almost eliminated the need to look at resumes and education history to hire top candidates in these roles.
Pymetrics has even open sourced their tools for proactive bias reduction by sharing its algorithm auditing tools. They understand the enormous challenge of reducing bias in these systems, since most of the training data is based on prior hiring success. The company recently received an additional $40M in funding and includes Workday as one of its investors.
The well-funded company Imbellus just announced a $14.5 million round of funding to help it continue to develop its simulated-based assessments that are now used by McKinsey to evaluate the problem solving of new hires. I’ve tried the simulations and they’re quite mind-bending and interesting, and they clearly test complex thinking in ways that go far beyond typical SAT or other tests. The CEO Rebecca Kantar is laser-focused on replacing the aging and dated SAT test (which is nearly a billion dollar market) to change the way companies hire and thus how colleges assess high performing students.
HireVue, one of the pioneers in video interviewing, can now capture more than a million meaningful data elements about a job candidate in each minute of video, and can tell managers things about candidates’ truthfulness and confidence in answering questions. They also have clients who love the technology, especially for high volume hiring in retail, customer service, and hospitality. Hilton has seen a 16% increase in hiring diversity and a tremendous increase in efficiency using this technology. The company now has more than 600 customers and has delivered more than 5 million video interviews.
PhenomPeople, another well-funded company started by engineers from India, has completely reinvented the recruitment process with a focus on end-to-end marketing. Sourcing, recruiting, internal career mobility, and managerial assessment are all linked, so PhenomPeople decided to build a recruitment system that looks like a career portal. Now it is augmented by AI, making it easier than ever for recruiters to find the right people; candidate communications are targeted the way marketers target ads; and the behavior of candidates (internal and external) is tracked to help personalize the job search experience. They call it Talent Relationship Management (TRM), which is a great way to describe it.
LinkedIn, of course, just announced a large range of new AI-based job placement and search tools, plus its own applicant tracking system. LinkedIn’s new tools let recruiters more effectively find the right candidates, write job descriptions that are most likely to reach the right diversity of candidates, and now provides enormous amounts of data to help target the right demographic, location, experience, and other characteristics. Again all in an effort to remove this error-proned process of bringing lots of people in for interviews. (Read more about LinkedIn’s announcements here.)
A new startup by the name of Orderboard.ai is focused heavily on the most in-demand jobs (cybersecurity experts, AI engineers, etc.) and can not only assess capabilities and job fit, but can also match an individual with the actual makeup of the team being hired. Orderboard’s AI has been able to help recruiters more than double the quality of candidates they can find, and its “attractiveness algorithm” gives companies almost 50% more likelihood a hard to find candidate will accept a call or consider a position.
And then there’s the enormous opportunity for AI to improve screening, which is a huge waste of recruiters’ time. Chatbots like Mya (the pioneer in this space), Olivia, Myra, IBM Watson Recruiter, and an exciting one by the name of Yva are getting smarter by the day. I’ve looked at many of these tools and they are amassing more and more intelligence about the types of questions candidates ask, now able to help recruiters spend more time sourcing and selling candidates and less time screening.
(The Chatbot market is huge, so vendors are getting focused on application areas. Make sure you talk with vendors focused on recruitment, not general chat.)
Of all the potential areas for AI and cognitive technology to add value to HR, this may be the biggest. While the technology is still young, success stories are now common, so I think it’s time for every company to make sure they have AI-based assessment on their list of things to do.
The risks in all this, of course, are that the AI somehow introduces bias into the system, so these vendors are working hard to make sure their systems are unbiased, transparent, and safe. In most cases companies test these systems with pools of high performing candidates first, to make sure the algorithms don’t inadvertently reproduce bias from the old “human style” of interviewing.
Facebook ran into trouble in this area when its algorithm-based job advertisement system enabled recruiters to discriminate by age. So you do have to make sure the vendor is well versed on these issues.
For job seekers, I know this feels a little creepy but remember no employer wants to hire the wrong person. These tools should make your life easier too, since you won’t feel as dependent on having a good day for the interview to get the job that fits you well.
I’ll keep monitoring this space as it grows, but right now I’m very bullish. (I am tracking more than 1400 HR technology companies and more than 40 of them are focused on AI-based assessment, one of the biggest categories of growth.)
After years of studying the world of assessments as an analyst, I see a step-change in value here – and since hiring is the most important thing we do as leaders, this is an area where HR technology can really help a company outperform.
This article was first published at https://joshbersin.com/