By Dhrumil Dhakan
HRM solutions have become quite essential in the operations of an organization and with the solutions being integrated with AI/ML, it has largely streamlined the whole process of onboarding, training, and other HR activities. With this in mind, it is important to know where a technology like this lie in today’s world and how beneficial or harmful can it be.
“Everything is now moving digital, HRM solutions used to be there in the past, if you look 4-5 years back, they were used for your basic HR processes and with organizations now moving to virtual, these HRM solutions really help you identify on how you’re doing the capacity planning are you really utilizing your employees. Earlier, we were not able to do a time and motion study, you had to apply various tools to get time and motion for any employee or any department,” Sharad Sharma, CHRO & Chief Ethics Officer, Pramerica Life Insurance said.
Talking about HRM solutions today, Sharma said in one click his team is able to provide time and motion study, their business capacity, productivity analysis, how are people contributing to the function, and since everything is digitized, one doesn’t need to have an interface on basic HR processes.
Automation is the biggest value these HRMS bring to the table, believes Vicky Jain, CEO & Co-Founder, uKnowva, who feels it is important to know what are the various processes within your company and where does AI/ML factor in.
“In a lot of areas, finally AI/ML, they’re all doing only one thing which is automation again, which software also did to some extent and then AI/ML came into the picture where a logical software or a normal program won’t work. A very simple example is resume parsing – earlier, what used to happen is screening used to be a 15-20 minutes task, ‘When I see a CV, he has done this much, he has these many skills, ok he is screened’, so I spent 15 minutes, now a resume parser can do it within microseconds, so this is AI that is induced,” Jain explained.
He added that there are certain areas which probably were never even done like capturing the happiness of employees, which is where this kind of solution has come in, stating that earlier no one even thought of implementing these features but now it is possible because we’re doing so much of virtual interaction, so much of timing spent on the screen. And there are many more areas like intelligently figuring out when someone might leave, might resign, by analyzing data – leaves, attendance, all of this.
Prashant Kaddi, Partner, Deloitte India also shared some insights on where does the AI/ML part come into the equation, stating, “Organizations that need the best talent, will do well only when they leverage the latest that technology has to offer, which is to use AI and Machine learning to help them identify the right talent with the right skills using AI, based on the ML algorithms that look at the history of employment trends and find the right requirements for successful hires in the past.”
He explained further that core HRMS leverages historic data and learns from it, to identify the risk of attrition by combing through performance history and pulse surveys to find the mood of the employee, understand and assess engagement levels and compare the leavers to alert the need for intervention and help the employees stay engaged with development, reward, and realignment strategies.
AI and Bias are terms that go hand-in-hand and it is important to state the role of the same in HRMS when AI plays such an important role. Kaddi added saying, “Using too much AI / ML in people process can have a negative impact to employee experience, some recent studies have shown that Automated chatbots are not trained to understand human emotion and the loss of a personal connect drives employees further away and makes them feel alienated, which is more relevant now, in a post-pandemic era with remote work prevalent everywhere.
He continued, “Bias that can creep into the algorithms due to the limited nature of keyword matching and selection processing that the software understands to recommend the right candidates for jobs.”
To overcome bias, Kaddi added that mechanisms such as model audits are implemented. Further AI ML decisions are augmented by human oversight and based on manual annotations of deviations, model recalibrations are factored in on a periodic basis.
Jain gave an example to break down the concept of bias in simple terms, “I’ll give you a simple example on how does bias creep in, and it is the same approach – like your GPS tracking, there is a bias there also and how is it dealt with. For example, if you are sitting over here, why does it say exactly the same location, it is actually capturing signals from three antennas and on the basis of the intensity of the signal of the each tower, it is able to figure out, so basically what I’m trying to say is if you’re dependent on just one source, the bias would always have been there, but there were three areas from where data is being collected and that is the key.”
On what his experience has been with bias, Sharma said, “Most of the time, the HR team, or the HR functions, throw the data but most of the time, it ends up being emotional data, it is not backed up with some intelligence on ‘What is this data telling us?’ ‘Do we hire a profile like this?’ If there is a trend of attrition that is showing profiling of people, the top management would want to really see those insights on ‘what is this data telling you?”
He added that in the past, in the absence of any AI tool or any other, decisions like these were made on your own assessment of your situation which was often presented to people, but these tools help throw light on the inferences on the data, which help make the conversation more meaningful and targeted.
“I think, HR functions have always been deep on analytics, and with all these tools & applications coming in place, it really makes the HR functions’ role look more clearer and they can possibly talk with a lot of conviction on what the data is telling them,” Sharma said.