How next-gen automation is helping recruiters to identify qualified talent at scale

With the Great Resignation showing no signs of letting up, recruiters are looking for all the help they can get to replenish their headcounts with qualified talent. The human resource management (HRM) market – including talent acquisition software and services – is currently valued at nearly $20 billion.

It is expected to grow at a rate of over 12% annually until 2028 on the back of continued digitization and automation of recruiting and HR operations.

Across the world, enterprises are putting an emphasis on creating and retaining the best, brightest, and most diverse employee pool. Expectedly, advances in artificial intelligence (AI), machine learning (ML), and predictive modeling are giving enterprises – as well as small/medium-sized businesses – a never-before opportunity to automate their recruitment even as they deal with radical changes in workplace practices involving remote and hybrid work.

In fact, four out of every five recruiters surveyed in an Entelo study believe productivity would increase if they could automate candidate sourcing altogether. They were unanimously of the opinion that having more data would assist them in qualifying candidates, evaluating candidate pools, improving outreach, and perfecting hiring workflows. Despite this, 42% didn’t have the data or the time to implement or dig into analytics, let alone turn the data into insights.

Enter recruiting automation solutions.

What is recruiting automation and how can it help?

Human resource or people management as a function begins with hiring. Every day an open role remains unfulfilled costs companies profit and productivity. Intelligent tools based on AI can gather relevant data on candidates, make it available to recruiters, and then process it accurately to speed up and streamline multiple sub-processes, including candidate sourcing, screening, diversity and inclusion, interviews, and applicant tracking.

“The days of physically sorting through hundreds of resumes and posting your job descriptions on each individual board are over,” notes Ilit Raz, CEO of Joonko, a talent feed solution for surfacing candidates from underrepresented backgrounds. “Without some form of automation or HR tech, you’re always going to be a step behind your competitors, especially when it comes to recruitment.”

Recruiting automation is a category of technology – delivered as software-as-a-service (SaaS) apps and increasingly powered by AI – that an organization can use to manage all aspects of its workforce. Its central aims include:

• automating recruiting tasks and workflows
• reducing cost per hire
• increasing productivity of HR personnel and recruiters
• accelerating filling of vacant posts
• bias-free hiring
• improving the company’s overall talent profile.

How does a typical AI-based recruiting automation technology help you go about achieving these goals? Here are the different functions where it can play a key role:

• Job ads: Recruiting software can automate purchase of ads on jobs platforms as well as other websites. It leverages programmatic advertising and branded content to place job postings on industry-specific sites that your target candidates frequent. It can also help you optimize your job advertising budget and reduce cost per applicant.
• Application tracking system (ATS): An ATS is software that automates the complete hiring and recruitment cycle for an organization. It provides a centralized location to manage job postings, sort through resumes, filter applications, and identify the most suitable candidates for open positions. This way, HR managers can stay organized and get easy access to details on the stage at which a candidate is in the hiring process.
• Resume screening: Manually screening resumes is one of the most time-consuming parts of recruiting. AI-based software “learns and understands” the job requirements based on the listing and filters resumes based on keywords, terms and phrases used by candidates.
• Pre-qualifying candidates: Intelligent algorithms can determine probable candidates by evaluating their skills, experience and other characteristics with those of previous hires and the published job role. They can also rank or grade these candidates as they move them forward in the hiring process. AI-based chatbots can gather basic information by initiating conversations with candidates and “learn” more about them. The algorithms can also scan through their LinkedIn, Twitter, Facebook and other social profiles as well as industry-specific platforms on which they’re active (such as Stack Overflow for developers) for a better idea of their personality, knowledge, abilities and aptitude.

When can recruiting automation go wrong?

Despite the advances in recruitment automation software, it is not a panacea for hiring challenges. There is no technology cure for broken recruiting processes. Data overload is one critical problem. Recruiters have so much data (on candidates as well as job roles) these days that they have neither the time nor the skills to analyze it and arrive at the right decisions. Many times, the cost and complexity of accessing and verifying this data turns out to be prohibitive.

Another long-standing problem is bias. While the recruiting process itself is frequently biased (owing in no small part to companies’ propensity to rely on employee referrals), the use of AI and automation in hiring can sometimes compound the problem.

“If you don’t have a representative data set for any number of characteristics that you decide on, then of course you’re not going to be properly finding and evaluating applicants,” says Jelena Kovačević, IEEE Fellow and Dean of the NYU Tandon School of Engineering.

“For example,” she continues, “if Black people were systematically excluded in the past, or if you had no women in the pipeline, and you create an algorithm based on that, there is no way the future will be properly predicted. If you hire only from Ivy League schools, then you really don’t know how an applicant from a lesser-known school will perform, so there are several layers of bias.”

In an infamous instance, Amazon developed an AI-based recruiting tool that analyzed patterns in resumes received over a ten-year period and ended up discriminating against women. Needless to say, they scrapped it.

The biggest area where data and AI have failed is Diversity, Equity, and Inclusion (DEI). Some of the biggest diversity-related mistakes in recruiting that are amplified by automation and machine learning are:

• Insensitive, elitist or less inclusive language in job postings (drives diverse candidates away from applying)
• Limited sourcing and restricted candidate pools (leaves out candidates from another region or those who didn’t attend certain schools)
• No remote work policy (keeps out candidates with disabilities and lack of transport)
• A facetious approach to DEI aimed at meeting minimal regulatory or industry standards
• Lack of automation

The last one deserves special attention.

AI as the problem, analytics as the cure

While AI is certainly not a silver bullet for recruiting, it has come a far way since the Amazon fiasco. The Entelo study found that data-driven recruiting teams are already outperforming their peers. Further, 84% of recruiters are fairly confident in their ability to use AI and machine learning in their day-to-day workflow.

The million-dollar question is: How can recruiting automation technology use AI algorithms in the hiring process without adding (and amplifying) human bias into the mix?

The answer lies in establishing company-specific performance benchmarks, identifying key metrics to objectively measure the competency of candidates, and using talent analytics to measure the success and efficiency of your recruitment efforts.

Algorithms that fulfill the purpose they’re built for frequently do so because the largest and widest datasets are available for them. It is your responsibility to collect these data points and feed them into your talent pipeline or recruiting automation software. The process is reversed on implementation – it is always a good idea to test the algorithm on a small (but diverse) pool of candidates and manually review its output before adopting it as the de-facto hiring solution for your organization.