Identifying and retaining the right talent is the name of the game for CHROs across the board. And it’s not just this, the HR needs to ensure that there’s absolutely no bias when it comes to internal job postings (IJPs).
Staying one step ahead of the game is the need of the hour, and emerging technologies like predictive analytics and AI help CHROs do just that.
A discussion with prominent players in the market and a company pioneering in using AI for HR, brings to light the necessity of using machine learning in a domain that has, until now, seen limited usage of cutting edge tech.
Drawing the line between tech and human intervention
While data is informative, it might be incomplete. Analytics yields great answers to questions only if the data is complete and comprehensive.
“In cases like who to promote, it could be that we must supplement the analytics with our knowledge about the person. In this case, it could be that we use analytics to identify the right candidates for promotion, but then finalized the decisions with the information that the analytics did not have,” says Niel Nickolaisen, CTO, OC Tanner.
EdGE Networks, a company specializing in HR analytics, talks about the disparities that exist in matching right skills, to the right career paths and progression.
“We found Natural Language Processing (NLP), semantic analysis of data, deep learning and predictive intelligence to be methods of examining large amounts of data to uncover hidden patterns and unknown correlations that can facilitate the decision-making process,” says Arjun Pratap, CEO & Founder, EdGE Networks.
“We were trying to loop around that problem, which is reducing the bias, create consistency, create a level of accuracy, and make the process super-fast. The idea was to read a resume like the human brain and map it to the relevant job using a percentage match score,” explains Pratap.
Also read: How analytics is slowly replacing the human in HR
The fledgling HR specialist already has a predictable forecast available. So, this helped eliminate the need for recruiting a person at three times the hiring cost. “The talent acquisition function, as much as it sounds sexy on the outside, is not as important because it is a function that brings up more cost,” reveals Pratap.
So what is it that AI can accomplish, and traditional analytics cannot?
“Sometimes an attrition model will hit 85 percent, and that’s as accurate as it can get. But what you can guarantee is that 85 percent will stick like clockwork, based on the base parameters being the same,” says Pratap.
HCL Technologies, one of the early pioneers in adopting AI for HR functions, throws some light on why leveraging AI is the way forward.
“The algorithms that we currently use to enhance the available AI. and are able to scan work samples, employee history and candidate profile information related to the job, as well as match demands to relevant profiles. This not only eliminates ‘personal bias’ but makes the entire sourcing process efficient,” says Prithvi Shergill, CHRO, HCL Technologies.
Traditional analytics had a limited ability to state what the situation is, and at best, presents a what-if scenario analysis. Harnessing AI, though, goes one step further by using mature algorithms to use the same information and convert it into insight for the user to consider as they decide what probably could be the best next step.
Also read: AI is just getting started and tech giants are racing for it
Neeraj Sanan, Chief Strategy Officer at Spire Technologies, believes that harnessing AI can help organizations stop relying on placement companies altogether.
“The worldwide placement industry is worth about USD 35 billion, but if everything’s digitized, what value will these placement companies hold? We call it ‘the great placement robbery’,” says Sanan.
Leveraging AI for predicting attrition
Niel Nickolaisen, CTO at OC Tanner, shares his insights saying there are a lot of things companies can do when it comes to leveraging AI for predicting attrition. For instance, cluster analyses of employee turnover can identify areas of high and low attrition.
In addition to this, tracking the frequency and type of interactions between managers and team members can help draw conclusive insights. “Gathering and analyzing data from social networking sites (Facebook, LinkedIn, et cetera) to get a sense of employee satisfaction also plays an important role in determining the employee’s level of engagement in the organization,” says Nickolaisen.
Arjun Pratap, of EdGE Networks, explains the modus operandi that goes behind the attrition-prediction model.
“Attrition prediction looks at various aspects, one of the things is time series – looking at 30, 60, 90, and 120 days. We look at generating a probability percentage of the person leaving within these time spans. So, on the fly, I can look at people’s profiles which are skill-based, not just role-based. So, role and skill matches are very essential for us to understand what the competition is.”
“With this information, I’ll know the people best fit for the job role, the age groups, and even the gender. I’ll also know what the mismatch is in terms of historical information. Now based on this, I can predict who will stay in the company, and why high performers leave their roles,” explains Pratap.
Now the potential here is in mapping it back to other aspects. One can now predict what the attrition is going to look like, as one has market-related information, which is gathered from constantly mining through pipelines of information.
Prithvi Shergill, CHRO at HCL Technologies, gives us his point of view as an AI adopter: “Yes, our employee attrition prediction model has been matured over the last two years – and we continue to enhance and move away from simple regression to a ‘survival model’ technique that predicts attrition within a time period and prescribes probable actions to take with employees at risk,” says Shergill.