Recruiters Use AI to Rediscover Talent in Their Own Backyards

By Dave Zielinski Jun 16, 2017
LIKE SAVE PRINT
Reuse Permissions

When Tony Le has a new job to fill, the first place he often looks for candidates is his own applicant tracking system (ATS). Le, senior director of global recruiting for IAC Publishing in Oakland, Calif., uses the artificial intelligence (AI) provider Brilent to search the vast database of candidates who've previously applied at the organization for those who still might be a good fit. Brilent's machine learning algorithms sift through that database to match candidates to job descriptions and deliver to Le a short list of candidates ranked by suitability.

This AI-driven "talent rediscovery" saves Le the time and cost of posting openings to job boards or social media sites, delivering candidates with a proven interest in his brand who may have been finalists for previous openings.

Such AI tools address a long-standing weakness of the traditional ATS—the capability to conduct fast and accurate searches of existing resume databases for job matches. Le said his use of Brilent has significantly slashed the time he spends searching for and screening candidates.

"Use of artificial intelligence has helped us reduce at least a week's worth of time on the front end of the cycle," he said. "Instead of taking a week or more to get candidates in front of a hiring manager, it's now often days."

[SHRM members-only toolkit: Introduction to HR Technology]

Brilent made a splash at the 2017 SourceCon conference when it was pitted against human recruiters in a competition to test how quickly the platform could match candidates to open jobs. Recruiters from around the world competed to be the top eight humans to test themselves against Brilent's artificial intelligence. Results of the competition were definitive: Brilent took 3.2 seconds to deliver its results while the humans took from four to 25 hours, according to SourceCon.

Brilent first integrates with an ATS or candidate relationship management (CRM) tool and then updates existing candidate profiles with publicly available data on the candidates' new skills, employers, locations and more to ensure they remain a good fit for the role. The technology then ranks top candidates on a scale of zero to 100 based on their qualifications, interest level and availability, said Garry Ma, CEO and co-founder of Brilent. Clients receive a detailed summary of why candidates are considered a good fit. Recruiters can also use Brilent to rank candidates on job boards for open roles, receiving notice of anyone deemed a good fit.

Le periodically keeps the AI honest by reviewing candidates the technology deems a less-than-ideal match. "There can be a danger in trusting the algorithm too much," he said. "But the results have been consistently good for us in reducing time-to-hire and improving quality of hire, whether it's for technical roles or jobs like product manager [or] salesperson."

What's Old Is Relevant Again

Talent rediscovery is also a major initiative at comScore Inc., a Reston, Va.-based company that measures brands and consumer behavior. Principal technical recruiter Pete Radloff uses the application Restless Bandit to search his ATS databases for past applicants who might be a good match for open jobs.

Restless Bandit first scours comScore's database to eliminate any resume duplication, then updates resumes with new information, compliments of spiders that crawl the web seeking fresh data on past applicants. 

When a new job opens, Radloff's first move is to use Restless Bandit to check his ATS database for potential candidates. "I want to know what's already in my own backyard before I go out to external platforms," he said. "I don't want to overlook a great candidate we might already have in-house. I recently made three good new hires as a result, two of whom I would have had to look far and wide for to find without the internal search."

Once the algorithm matches top candidates to open jobs, an automated e-mail is sent to each of those candidates, noting that the prospect had previously applied and inquiring whether he or she would like to apply again for an open role. Interested candidates click on a link to apply. 

"We compare the concept of talent rediscovery to running out to the store for a bottle of milk and then realizing you already have a fresh gallon in the refrigerator," said Steve Goodman, CEO and co-founder of Restless Bandit.

Past applicants are also retargeted through social media sites like Facebook. Because candidates often use the same e-mail address on their resumes as they do for their Facebook logins, they can be approached with direct ads on the social site.

The ad might highlight the job opening, mention that the person has previously applied at the organization and provide an option to reapply. "That tends to get a high click-through rate because it is so surgical in nature," Goodman said.

AI for Internal Mobility and Industry-Specific Searches

Talent acquisition professionals looking to use AI-driven recruiting have a growing number of options. ATS providers, for example, are beginning to enter the game. Earlier this year, San Francisco-based SmartRecruiters unveiled a new tool, Recruiting AI, that will use artificial intelligence and machine learning to help clients discover, screen and score job candidates.

Other AI recruiting companies, such as New York City-based Untapt, specialize in recruiting for specific industry verticals and have also built internal platforms that match employees to open jobs within their own companies, guaranteeing them confidentiality until they've formally applied for a position. 

Untapt also has a broader hiring marketplace that caters to candidates for technology jobs in the financial sector—known as "fintech"—as well as to hiring managers in that industry. Candidates for jobs upload resumes, LinkedIn profiles and other relevant materials and are matched by machine learning algorithms to companies where they're most likely to land interviews. Hiring managers use the platform to post open jobs, relying on Untapt's technology to quickly link them to top candidates.

"The machine learning algorithm analyzes not just words but a complete resume or profile and the context in which words are used together," said Ed Donner, Untapt's CEO and co-founder.

How precise is the matching? Candidates might list "Python programming skill" on their resumes, but that skill can mean different things to different candidates. Some might be data scientists using Python to build models, others might use Python to build web applications and still others might be infrastructure technologists who use Python to write scripts. Untapt's technology can easily distinguish between the three candidate types and match them to the appropriate job openings, according to Donner.

Dave Zielinski is a freelance business journalist in Minneapolis.

Was this article useful? SHRM offers thousands of tools, templates and other exclusive member benefits, including compliance updates, sample policies, HR expert advice, education discounts, a growing online member community and much more. Join/Renew Now and let SHRM help you work smarter.


LIKE SAVE PRINT
Reuse Permissions

MEMBERSHIP

Become a SHRM Member

Join/Renew Today

Job Finder

Find an HR Job Near You
Post a Job

SPONSOR OFFERS

Find the Right Vendor for Your HR Needs

SHRM’s HR Vendor Directory contains over 3,200 companies

Search & Connect