According to the article “Artificial intelligence and the future of work,” AI can be used to optimize existing processes, allowing up to 80% of saving in costs of externalization of auxiliary business services or services for clients, and improving processes, such as reformulating sales forecasts, logistics, and the supply chain.
When it comes to people management, this news is the latest trend. The American company Deepple develops solutions in data engineering and artificial intelligence applied to human resources departments.
In other sectors, its development allows us to identify the potential of the people and teams belonging to a business, and evaluate their ability to reach its goals.
By combining AI with social psychology and neuroscience, Deepple created a tool that detects employees’ propensity to quit their jobs.
“Human behavior cannot be predicted,” warns Yoel Kluk, Chief Product Officer and AI specialist at Deepple. “However, fortunately, computers are made to imitate how people work. Through algorithms, I can find the ‘perfect storm’ that indicates a person’s propensity to leave the company,” he says.
“Businesses possess an incredible amount of information about people, yet they don’t use it,” he adds. “The input for this tool is information that already exists about people: their gender, their seniority, their shift, which building they work in, or the courses they’ve taken. It’s about distributing the importance of each variable in order to determine each person’s propensity.”
Of course, he explains, this does not replace human beings’ decisions. “I can’t replace, for instance, a supervisor’s experience. But it is the goal: it reads data and is crucial for generating predictive models.”
“After ‘The Great Resignation’ (the record number of people quitting their jobs after the end of COVID-19 restrictions), recruitment cycles, rotations and employee successions have become an issue,” states Kluk.
Ezequiel Kieczkier, founding partner at Olivia and partner at Deepple, states “The impact of GPT is directing our attention towards AI. The most important companies are asking themselves the initial questions and working on the first projects to integrate these models.”
When used for people management, AI allows “objective revisions: it takes data, processes it without cognitive judgement, and places it elsewhere,” says Patricio Martínez, Vice President and Head of SAP SuccessFactors.
For example, by using video analytics, assistance within different shifts of a plant can be analyzed, as well as whether or not employees are working where they should be and under the necessary conditions. Or, in productive operations, whether they possess the certifications and have taken the necessary courses to carry out the tasks they’re assigned.
“Today, these things can be automated, done at a more efficient cost, and flagged in case an alert comes up so that someone can make a business decision using their cognitive judgement,” adds Martínez.
A survey done by Grupo Gestión on over 60 directors and managers of Human Resources in Argentina shows that 60% of them views AI as a reliable technology for staff selection.
“The more massive the process is, the greater the benefits of incorporating AI, which serves as an initial filter,” explains Daniel Iriarte, Associate Director at Glue Executive Search. “The company is willing to risk an error if 20,000 curriculums can be analyzed in 50 minutes,” he adds.
However, in executive searches “we’re not seeing as big of an impact,” he says. “Even though we have a system that possesses keyword filters, the rest is about conversation and evaluating soft skills.” With people, he assures, “there’s always an aspect that transcends data.”
BW Communications is a consulting firm that specializes in internal communication. One of the services they offer is bot development, whose demand has increased with the pandemic.
“Bots have been working over the past couple of years to automate all kinds of processes within HR: from answering frequently asked questions, to carrying out onboarding processes; from paychecks to leaves and PTOs, and to reserving workrooms,” lists Pablo Faga, their director.
These bots are relatively “basic,” since they respond to “requests” or identified keywords that offer prefabricated answers. “The process of learning is being crafted: someone identifies the questions asked by users and answers are created based on them,” he explains.
Incorporating AI will contribute a great amount of value to this type of service. “We’re experimenting so that, when a bot runs out of answers, instead of responding that they don’t have the answer, it connects with an AI engine,” states Fraga. “And, on the other hand, that it communicates with people using the specific cultural style the business is familiar with.”
This is how AI will enable a bot to respond to any type of question in a variety of ways after “studying” a business’ introduction manual, for example.
“That’s what we’re working on,” says Fraga. “Each business communicates differently, with their own language and variations. The challenge is for AI to respond in that same variation that’s specific to the business it’s speaking to.”
“We always say that bots handle tasks that don’t add value. That creates a big opportunity for carrying out more strategic tasks, such as thinking of good ideas or speaking with people,” he concludes.