Major tech companies have announced plans to invest in the latest iteration of workplace AI. Google, IBM and Salesforce are among the firms planning to develop agentic AI, which is part of the so-called third wave of AI.
Agentic AI uses highly sophisticated ‘AI agents’. These are workplace assistants that differ from AI chatbots, such as ChatGPT, in that they can perform tasks with minimal human oversight. Last month, Microsoft updated its Copilot Studio to enable businesses to create ‘AI employees’, which are capable of completing more complex jobs, such as qualifying sales leads and handling client queries.
According to Josh Bersin, the HR technology analyst, AI agents will soon be able to assist with talent acquisition, leadership coaching, employee onboarding and learning and development. Dayforce, Workday and Servicenow are among the HR software providers that have recently unveiled agentic AI products.
The most advanced AI agents are currently used in recruitment. Last week, LinkedIn launched its Hiring Assistant, which is designed to automate large portions of the recruitment process. Meanwhile, fast-food chain Chipotle claims to have cut the time to hire in its restaurants by 75% using an AI agent developed by Paradox, a recruitment-software company.
It is now possible for a candidate to apply for a job, be accepted and turn up on the first day without ever speaking to a human. Bersin quips: “People are coming to their first day of work asking to talk to Amelia. Well, there is no Amelia; Amelia is a bot.”
The human advantage
But taking humans out of the equation entirely can be detrimental to businesses, according to Agata Mirowska, assistant professor of HR management and organisational behaviour at Neoma Business School.
“If a company uses AI in the recruitment process, it’s already sending signals about how people are treated in the organisation. It may cause people to think twice about applying,” she warns.
HR is always worried about AI devaluing people
For Mirowska, the thought of an employee being hired without speaking to a person at the organisation is concerning. She explains: “An interview should be a two-way exchange. By not putting a human in that role, employers are potentially cutting down on the candidate’s ability to ask questions and get a feel for the culture.”
ServiceNow, for instance, has announced AI solutions for IT, customer service, procurement and HR. But its senior people director, Katie Whitehouse, insists that AI will always “work alongside human employees” rather than replace them.
“Agentic AI works like a multiplier for what human employees can accomplish, but humans must always be in the loop to provide oversight and set guardrails,” she says.
AI employees
Because AI agents can operate independently of human manipulation, Bersin speculates that employees will interact with them like colleagues. But the idea of AI being treated like a human worker in any sense has not been welcomed by all HR professionals.
Earlier this year, Lattice, an HR software company, introduced a feature that would allow ‘virtual workers’ to be inserted into organisational flow charts, onboarded and assigned managers. The announcement was met with considerable pushback.
“HR people have a very emotional reaction to AI agents,” Bersin explains. “What about the humans? HR is always worried about culture and ensuring employees feel valued. They’re worried about AI devaluing people.”
This is a concern for Mirowska, who warns: “Once you go from defining AI as a tool, to defining it as an employee, teammate or collaborator, you’re implying it is capable of taking human jobs. This can lead to stress and insecurity in your workforce.”
I worry whether companies are trying to go too far, too fast
Constanze Backhaus, global CHRO at Teamviewer, a software company, agrees that AI should be viewed as a tool to complement human abilities, but must not be equated with human employees. “Part of a successful AI implementation is helping employees move through fears and concerns about AI replacing their jobs,” she says.
Teamviewer is currently exploring the use of AI in multiple functions. Backhaus believes there are opportunities for the technology to be used in recruiting, workforce planning and employee engagement. But she cautions: “All these use cases come with their own benefits, risks and mitigating actions.”
It is important that HR leaders weigh in on this technology as it develops. They will be responsible for helping to define best practices in the implementation of AI assistants, determine which roles they can best serve and ensure the content they’re trained on is accurate and up to date. According to Bersin: “Organisations are going to need a little army of people to take care of these things.”
He adds that AI agents will ultimately enable HR to support more people and gain greater insight into employees’ concerns. While there are concerns about AI’s impact on jobs, HR staff should feel confident that the function can’t yet be entirely replaced by AI agents.
“The main reason HR exists is to help the business grow, improve the employee experience, help people to develop their careers, train managers and deal with critical issues, such as harassment and poor behaviour. AI agents will help but they can’t do all those things,” Bersin explains. “It’s an amazing tool – but it’s just a tool.”
A cautious approach
Although AI agents can greatly benefit the HR function, Mirowska urges companies to implement the technology cautiously. “I worry whether companies are trying to go too far, too fast, without really thinking through the consequences for employees and the workplace experience,” she says.
For example, many business leaders hope AI agents will be able to handle low-skilled work that is typically completed by more junior staff members. But Mirowska points out that these tasks help to prepare workers for more complicated work in the future. “We’re risking creating a gap in the workforce,” she says.
Completing menial or mindless tasks also provides employees some respite from the more taxing elements of their roles. “People can’t constantly work at 110%,” Mirowska says. “While an AI-enabled workplace might sound good, it could contribute to the stress and burnout people face if people are with only the most challenging work.”
She advises HR chiefs to assess the longer-term benefits of agentic AI, rather than rush to adopt the new technology. “HR leaders should approach AI agents in the same way they would any other HR initiative. Take a step back, look at it strategically and decide whether it fits into your strategy.”
